Document 0g7eqpza11NqgvBZmKNkDDkrk

ORGANIC CHEMICALS DIVISION RESE , REPORT NO.- 2961 JOB NO.- kjjk - Part I FINAL REPORT ON THERMINOL FLUID VAPOR PRESSURE AND THERMAL STABILITY INSTRUMENT DATE - July 6, 1963 W. N. Trump WRITTEN BY - C. E. Vogler > ST. LOUIS, MO. HARTOLDMONOOQ5234 MONSANTO CHEMICAL COMPANY Organic Chemicals Division St. Louis Research Department St. Louis Research Report No. 2961 Final Report on Thermlnol Fluid Vapor Pressure and Thermal Stability Instrument Job No. 4374 - Part I July 6, 1963 Written by: W. N. Trump C. E. Vogler Work; done by: W. N. Trump C. E. Vogler H. W. Luebke L. Fowler W. E. Koerner A. J. Bindbeutel A. L. Manno 0692 700 HARTOLDMONOOQ5235 DISTRIBUTION FOR REPORT NO. 2961 1. File 2. K. L. McHugh 3. W. R. Richard, Jr. 4. Duplicate File 5. Technical Reports Library "6~." C. E. Vogler 7. W. N. Trump 8. L. Fowler 9. W. E. Koerner 10. R. Davis - Org. Sts. 11. W. B. Hewitt - J. F... Q. 12. Extra 13. Extra 14. Extra 15. Extra 16. Extra This report contains confidential information which is the property of the Monsanto Chemical Company which shall be disclosed only to duly authorized persons. The recipient is held accountable for the filing and safe custody of this report, which must be returned on demand. 0692701 HARTOLDMONOOQ5236 TABLE OF CONTENTS Page No. INTRODUCTION........................................................................................................... 1 SUMMARY ........................................................................................................................1 CONCLUSIONS............................................................................................................. 1 RECOMMENDATIONS................................................................................................... 2 PATENT STATUS........................................................................................................ 2 EXPERIMENTAL WORK AND DISCUSSION ...................................................... 2 A. Development of Operating Procedure 1. Early Work.......................................................................................2 2. Selection of Sample Size.....................................................5 3. Selection of Heating Rate . ....................................... 7 4. Detection of Low-Boilers ..........................................7 5. Vapor Pressure and Thermal Stability of the Therminols ...... .............................. 15 6. Vapor Pressures of Standards and Other Materials ...................................................................... 16 B. Design Considerations 1. Sample System and Furnace ................................21 2. Temperature Recording andControl System . . 29 3. Pressure Recording System ...................................... 36 4. Programming System ....... ......................... 36 5. Recorder ............ ......................... 39 ACKNOWLEDGMENT ............................................................... APPENDIX ............... ................................... 39 39 0692702 HARTOLDMONOOQ5237 1. INTRODUCTION An instrument was wanted for control use at Anniston which could be used to screen various lots of Aroclor production for designation as Thermlnol fluids. Desirable in this Instrument was the ability to detect low-boilers and material which would decompose appreciably at operating temperatures. It was antici pated that specifications for the Thermlnol fluids could be based on this Instrument. Another requirement of the instrument was that it would require a minimum of operator attention. An instru ment which was considered to best fit the above requirements was one based on vapor pressure. This report concerns Itself with the design, fabrication, operation and performance testing of such an instrument herein referred to as the Vapor Pressure and Thermal Stability Instrument. SUMMARY An instrument was designed and built which would measure the vapor pressure of materials which are liquid at ambient temper atures . This instrument was tested for functionality, reproducibility and reliability with various lots of the Thermlnols and with pure substances of known vapor pressure. The effect of possible contaminants upon the vapor pressure and thermal stability of the Thermlnols was determined. An operating procedure for using the Instrument was developed. CONCLUSIONS The Vapor Pressure Instrument adequately records the vapor pressure of the Thermlnols and other liquids as a function of temperature. Unstable impurities in the Thermlnols which decompose at 200JOO^C. can be readily detected from the vapor pressure trace at the 500 ppm chloride level. Low-boilers such as benzene are de tectable at the 0.1$ level but levels of biphenyl lower than 2$ are not detectable from the Instrument record. Degassing of the inner sample cell surface requires tempera tures in excess of 250eC. with the system under vacuum and is necessary to avoid any contribution to the system pressure in ex cess of the vapor pressure of the sample during a run. Dissolved gases in the sample also must be removed by applying vacuum to the sample prior to run. 0692703 HARTOLDMONOOQ5238 recommendations It is recommended that vapor pressure specifications for the Tlierminol fluids be established based on a large number of meas urements made at Anniston. PATENT STATUS A preliminary disclosure of invention has been filed. No ref erence to a similar device was found reported in the literature. EXPERIMENTAL.WORK AND DISCUSSION A. Development of Operating Procedure A detailed operating procedure is given in the instruction manual which is included in the appendix of this report. In this section the experimental work and observations which led to the operating procedure are discussed. 1 Early Work The Instrument with the original cell design shown in Figure 1 was tested with substances of known vapor pressure, namely, water and dodecane and also with samples of the Therminols whose vapor pressures had been determined with an isoteniscope by Dr. K. L. McHugh's group'. Fairly reproducible plots of pressure vs. temperature were obtained but the data deviated widely from the known vapor pressure, curves. The instrument records for the Therminols were somewhat "S" shaped exhibiting a hump in the curve at about 200C. This was believed to result from temperature gra dients in the cell. The considerable lag time between the furnace response and the sensing thermocouple also caused uneven program ming of temperature and prompted the redesigning of the cell and improving the heat transfer from the furnace to the cell. The equipment modifications are described in detail under Design Con siderations (Section B). Dodecane was the first material tested in the modified instru ment with the redesigned sample cell shown in Figure 9 of the In struction Manual. The pressure vs. temperature plot was in excel lent agreement with the known vapor pressures' for dodecane and the programming of temperature was uniform. In the subsequent series of measurements with the Therminol fluids, the first test gave a reasonably straight line plot and the next showed slight S-shape curvature. The vapor pressure plots from the runs which followed became progressively more S-shaped and deviated more from the known vapor pressures. A recheck with dodecane agreed well with literature values but further tests with the Therminol fluids 0692 70<t HARTOLDMONOOQ5239 MATERIAL. TV PC JjC S3 DiK'C (V.T /o^.'- ..- /ivCi-tE. : ' ft Figure 1. Sample Cell (First Design) 0692705 HARTOLDMONOOQ5240 4. again showed the hump in the curve. At this point it was thought that the hump in the vapor pressure curve might be characteristic of the Therminols possibly resulting from low-boilers in the sam ples. An attempt was made to verify this by running a sample of Thermlnol-FR-2 from which the lower boiling constituents had been removed by distillation. The vapor pressure curve still showed the characteristic hump. Another test was made to Insure that the gases in the cell were composed entirely of Thermlnol vapor. A run was interrupted after attainment of a temperature of 220C. The apparatus was turned upside down and the gases were evaluated without removing the liquid phase from the cell. The apparatus was then placed right side up and the remaining liquid phase al lowed to equilibrate with its vapor. The run was resumed at this point. The Instrument record after the outgassing followed a straight line agreeing fairly closely with the vapor pressures ob tained by isotenlscope. Further tests of this type showed that the vapor pressure curve following outgassing of the vapor space could be shifted depending upon the temperature at which the out gassing was done. These observations still pointed to low-boilers in the sample as being- suspect for the poor reproducibility and lack of agreement with the known vapor pressures. Because up to this time it had been Impossible to demonstrate the reliability of the instrument, the Instrument was tested in dependent of the sample. The temperature and pressure recording systems were tested Independently of each other by using d.hnotor driven variac to program heat to the sensing thermocouple and by controlled bleeding of nitrogen into the system to simulate vapor pressure increase with temperature. These tests revealed inade quate thermosfittlngof components In the logarithmic converters and premature failure of mercury batteries in the bridge circuits as causes for variations in the vapor pressure-temperature traces. The correction of these difficulties is described in Section B, Design Considerations. Malfunction of the strain gage and fluc tuations in line voltage were eliminated as possible causes for variation in the instrument. The mode of heating, either by in strument programming system or instrument--independent motor driven variac, was found to have little effect on the recorded vapor pressures. Another possible explanation for the observed hump in the va por pressure curve of the Aroclors was system leakage. One obser vation indicated that the .pressure relief valve leaked slowly when the system was under vacuum and thus the valve was removed from the system, (it was later replaced.) Subsequent leak tests under vac uum at room temperature indicated a tight system. However, upon heating the evacuated system to 150C., the pressure indication was similar to the development of a large-.leak (0.5 psl per minute). The cell was re-evacuated to minimum pressure at this temperature and the valves closed. The pressure increase was much slower than before. The system was then heated to 250C. under vacuum. There was a slight increase in pressure at first but after about ten 0692706 HARTOLDMONOOQ5241 minutes under vacuum at this temperature, the system appeared to hold vacuum; that is, there was no Increase in pressure in ten minutes time. The system held vacuum upon cooling to room temper ature and it was concluded that the observed pressure Increases upon heating were not due to leakage but to the release of absorbed gases on the inner cell surface. Subsequent tests have shown that the cell surface is completely degassed when heated to 250C. and evacuated for ten minutes. As a standard operating procedure de gassing at the maximum temperature encountered in a run is recom mended. A run made with Therminol-FR-2 following the degassing procedure. Figure 2, gave a vapor pressure curve of the expected shape and which agreed-well with the lsoteniscope measurements. Subsequent ^runs showed that outgassing the cell at temperatures of 250C. or higher was an essential preliminary step for obtain ing accurate vapor pressure curves. In the early work, pumping the cell down to about 5 mm Hg at room temperature was thought to suffice since the pressure remained stable at this level when the cell was closed off. The problem of completely outgassing the cell before filling it with sample was not suspectedynor evident from the vapor pressure curves and was the cause of much bewilderment in the early experimental work. The solving of this problem proved to be., the critical one in developing an operating procedure for the instrument. 2. Selection of Sample Size Theoretically the size of the sample should not affect its vapor pressure in a closed system as long as there is vapor,.space remaining and sufficient liquid phase to allow accurate measure ment of its temperature. However, the vapor pressure exerted in a closed system by a sample which is a mixture of substances of varying volatility is dependent upon the volume of sample in the sample cell. A minor component of high volatility relative to the bulk of the sample would contribute considerably more to the system pressure when the liquid to vapor volume ratio is high than when the ratio is minimum. In a practical sense, for accurate measure ment of temperature the cell would need to be filled at least onefifth full and to avoid complete filling of the cell, the sample size would not exceed 90# of the total system volume. Between the extremes of 20# to 90# filling of the cell with a mixture whose components; volatilities vary over a relatively narrow range such as the Thermlnol fluids, the effect of sample size on the observed vapor pressure would be slight. Errors arising from poor evacuation of the system before sample introduction or slight leakage in the system will be greater the. larger the size of sample. On the other hand, sensitivity to meas urement of decomposition is improved with larger sample size. Thus in choosing the sample size a balance of these factors must be sought depending on the demands made of the instrument. 0692707 HARTOLDMONOOQ5242 tAP\------------- ------------ 6 Figure 2. Effect of outgaeeing cell 0692708 HARTOLDMONOOQ5243 . 7. In some of the early experimental work, the sample volume was varied to give about 1/4 to 5/4 filling of the cell with no evi dent effect on the vapor pressure record. In later work, filling the cell about 1/5 full gave satisfactory results both from the standpoint of temperature measurement (and hence, programming of temperature) and from being able to detect low-boilers and decom position In the sample. The procedure was adopted to use 12 oc of sample (pressure relief valve in place, 10 cc with relief valve removed) In a system of 24 cc total volume. This amounts to hav ing about Y.cc of sample in a 18.7 oc sample cell, the balance of the volume being in the piping and the valve and strain gage cavities. In using the instrument for control purposes the exact amount of sample used lsn't nearly as Important as using the same amount of sample time after time. In other words, repeatability is the most essential demand of the instrument. 5. Selection of Heating Rate Heating rates of 5-1/5C./min and 10C./min gave identical vapor pressure records . (Figure 5) for a sample of Therminol-FR-2. Extremely rapid heating rates, i.e., 25C./nln, caused an appre ciable shift in the pressure-temperature record due to the lag in response between the thermocouple and the sample which was equiva lent to about 20C. At a heating rate of 10CTJ/mln or less this^ lag is not appreciable. At heating rates less than 5C./mln, cjr- cling of the temperature with time became significant. These 'Fac tors bracket the range of usable heating rates between 5 and 10C./ min which appear reasonable also from the standpoint of elapsed time per determination. Another consideration in choosing the heating rate is the rate of decomposition of sample. Prolonging the run would accentuate the effect of decomposition gases on the pressure of the system and make the vapor pressure record more sensitive ini detecting unstable material in the sample. Unstable Aroclor, however, was found to decompose rapidly enough to not be a limiting factor at a heating rate of 50C./mln. In the early experimental work 5e'C./mln was chosen as a rea sonable heating rate and most of the work was done at this heating rate. The observations made later merely confirmed that the choice had been a proper one. The operating manual specifies 5C./mln as the heating rate to be used with the Therminol fluids. .4. Detection of low-boilers ------------------------------------------------------------------------------------ The chlorinated biphenyls usually contain dissolved gases which if not removed from a sample add to the system pressure during a run giving a pressure-temperature record higher than the vapor pressure of the material. The effect of degassing a sample prior to a run is shown in Figure 4. The same effect is observed 0692709 HARTOLDMONOOQ5244 3. when a sample Is admitted Into a poorly evacuated system where the fixed gas left in the system is compressed when sample is admitted and Increases in pressure with Increase of temperature during the run. Evacuation of the sample to about 5 mm Hg at room tempera ture was found to adequately degas the samples and was adopted as a standard operating procedure. For the more viscous materials such as Therminol-FR-3, the degassing was facilitated by warming the sample to about 50C. but care must be taken to avoid distil ling out the low-boilers when degassing sample at elevated temper atures . It is essential that the contributions to the system pressure from dissolved gases in the sample and residual gases in the evac uated cell be minimal when the instrument is used to detect lowboilers in the sample. Nearly all of the Thermlnol samples' studied with the instru ment contained some low-boilers. Figure 5 shows the effect of re- moving the more volatile components from a sample of Thermlnol-FR-2, the difference in the vapor pressure curves being more evident at the low pressures because of the logarithmic scale of the lnstru- ,, ment record. The presence of low-boilers in the Thermlnol samples ' Imparts more curvature to the vapor pressure plots at the low pres sure extremity of the instrument record. The presence of low- boilers in the Thermlnol samples was confirmed by Dr. K. L. McHugh's group by making the lsotenlscoplc measurements on samples where the low-boilers were removed and on those where they were retained. This effect is shown in Figure 6. Tests to determine the sensitivity of the instrument to low- boilers in the Thermlnols were made, using benzene and biphenyl as representative of possible contaminants. In Figure 7 is shown the effect of a relatively large amount of low-boiler in Thermlnol- FR-1, l.e., 1.6# benzene. Here the slope of the Instrument record is much less steep than for the uncontaminated sample. Lower con centrations of benzene do not alter the slope as drastically but levels as low as 0.2# benzene are readily detected from the instru ment record and it is estimated that a level of 0.1# benzene would*., also be detectable from the vapor pressure curves. Biphenyl at a 5# level in Thermlnol-FR-1 did not appear to .alter the slope of the vapor pressure curve (Figure 8), but did shift the curve toward higher vapor pressures the difference being equivalent to about 2CrC. on the temperature scale. Judging from these results, a 1# level of biphenyl in FR-1 would not be discernible from the vapor pressure curve and a 2# level would be questionable because of the slight variances in Instrument calibration. Using the change in slope at the low pressures as a guide post to contamination rather than the shift in the vapor pressure curve has considerable merit. For Instance, the contamination of Thermlnol-FR-1 with biphenyl could not be distinguished from a somewhat lesser chlorinated bi phenyl (intermediate between Aroclor 1232 and Aroclor 1242) con taining no unchlorinated biphenyl but exhibiting approximately the same vapor pressure. 0692710 t . : HARTOLDMONOOQ5245 9 HARTOLDMONOOQ5246 10 r /^\ Figure 4. Effect of preliminary outgasslng of sample , 0692 712 HARTOLDMONOOQ5247 t-Vt--------- 11 Figure 5. Removal of volatile components during outgasslng of sample 0692713 HARTOLDMONOOQ5248 &sej-- i /oo\~ 12 Figure 6. Comparleon of vapor pressure recording with lsotenoscope measurements 0692 7H HARTOLDMONOOQ5249 ioc r wo\ 13 Figure 7. Detection of low-bolllng Impurity 0692715 HARTOLDMON0005250 14 Figure 8. Detection of biphenyl 0692716 HARTOLDMONOOQ5251 15. Contamination studies were considered necessary only with Thermlnol-FR-1 samples since the other members of the Thermlnols have lower vapor pressures and lower levels of contamination would be evident from thetvapor pressure curves. 5. Vapor Pressure and Thermal Stability of the Thermlnols Typical Instrument records for runs with the Thermlnol fluids, FR-1, FR-2 and FR-3, and also Aroclor 12J2 are shown In Figure 9 The vapor pressures agree well with those determined by lsoteniscope for samples treated in similar manner. The lsotenlscope measurements were made only at subatmospheric pressures and no reported vapor pressure data at higher pressures could be found for comparison. There appears to be a slight curvature in the vapor pressure-curves at the high pressures deviating to pres sures lower than the expected straight line relationship. This curvature probably Is quite real but It should be noted that the slight deviations of the temperature record from an exact recipro cal temperature scale could also cause some curvature in the in strument record. All of the Thermlnols show evidence of considerable decompo sition at temperatures beginning at about 440C. The temperature at which decomposition began was evident by the increase in slope of the vapor pressure record. The decomposition which is evident at temperatures above 400C. appears to be characteristic of the chlorinated biphenyls and is accompanied by an increase in color of the sample ranging from yellow to black depending on the degree of decomposition. This decomposition also leaves a carbonaceous layer on the walls of the cell which requires some effort for re moval. It was concluded that the high temperature decomposition was unrelated to the quality of the Thermlnol fluids and could not be used as the basis for selecting Aroclor lots for use as Thermlnol fluids. Hence it is specified in the operating instruc tions for the run to be terminated at 400cC. to avoid the high temperature decomposition and hence to facilitate the cleaning of the cell. Another reason for avoiding the high temperature decomposition is that the carbonaceous layer on the cell walls contains trace amounts of metals from corrosion of the cell. The instrument record of a sample of "unstable" Aroclor 1242 (Figure 10) in addition to the characteristic high temperature decomposition showed an appreciable pressure in excess of the vapor pressure at about 200C. This sample contained 1480 ppm chloride (determined by alcoholic KOH treatment of the sample) which is equivalent to about 1 mole percent of the chlorinated biphenyl being the chlorine addition product instead of substitu tion product. The addition product upon thermal decompostion (or treatment with alcoholic KOH) loses HC1 to produce the substituted biphenyl of'one less chlorine. The liberation of. HC1 Increased the system pressure in excess of the vapor pressure as tempera ture was increased until the unstable contaminant was depleted. 0692717 HARTOLDMONOOQ5252 16. Beyond this point, the decomposition gas contributed to the system pressure as a fixed gas in accordance with the gas law. The re sultant "hump" shape of the curve can be used to detect "unstable" Aroclor down to chloride levels of 500 ppm when 12 cc size samples are used. Larger size samples would make the Instrument more sen sitive since the vapor space in the cell would be decreased. With a sample volume of 22 cc (total system volume = 24 cc), it is esti mated that 50 ppm chloride "unstable" material would be detectable. Monsanto's internal specification is less than 7 ppm Cl" and even by using minimal free volume in the cell, it doesn't appear feasi ble that the vapor pressure curve could be used to detect unstable material at such low levels. This is not to imply that the instru ment isn't utlllzable as a control instrument since abnormal lots of chlorinated biphenyls containing high amounts of unstable or low-boiler contaminants could be readily detected with the instru ment . It is possible to use the Instrument in a manner that would detect small amounts of decomposition. The gain in system pres sure at room temperature after a run could be used as a measure of decomposition. It would be necessary to use the pressure off set, however, to get both pressure readings on the instrument re cord scale. In Figure 11 Therminol-FR-1 heated to 400C. at the rate of 5C./mln and cooled rapidly to room temperature caused a pressure gain of 0.15 psl. This is equivalent to 0.006# decompo sition during the run. The material spiked with "unstable" Aroclor to a chloride level of 250 ppm Cl" showed a pressure Increase of 0.75 psl. This difference is significant and measurable but only about one-fifth the pressure increase expected from 250 ppm chlor ide. This perhaps could be due to the solubility of hydrogen chloride in the chlorinated biphenyls. Chlorodecane was originally believed to decompose similarly to the chlorine addition product of biphenyl and thus was used in an experiment to determine detectability limits. It appeared to decompose much more slowly and gave continued decomposition as the temperature was Increased to 400cC., Figure 12. Evidently much higher temperatures are required for its decomposition rate to approach that of the chlorine addition product of biphenyl. Since the main function of the Instrument is to measure vapor pressure, it is possible that vapor pressure specifications for the Thermlnols could be based on measurements with this instrument. It is anticipated that such specifications would result from a large number of samples run at Anniston. 6. Vapor Pressures of Standards and Other Materials Even though the vapor pressures of the Thermlnol samples were known up to atmospheric pressure, the fact that they weren't single component materials but mixtures made these samples unac ceptable as standards for the Instrument. Several organic 0692718 HARTOLDMONOOQ5253 17 HARTOLDMONOOQ5254 18 o Figure 10. Vapor praaeure recording of unstable Aroclor 0692720 HARTOLDMONOOQ5255 Figure 11. Vapor pressure curves and residual pressures of normal and hlgh-chlorlde samples 0692721 HARTOLDMONOOQ5256 20 Figure 12. Effect of aliphatic chlorine compound 0692 722 HARTOLDMONOOQ5257 21. materials readily available in pure form and whose vapor pressure were accurately known were chosen as standards for the instrument. These Included n-dodecane, o-xylene and o-dlchlorobenzene. The instrument records for the runs with the" standards are shown in Figures 13-15" Included for comparison are literature vapor pres sure values. The agreement with the literature values is quite acceptable, the slight deviations In the lOO-lSO'C. temperature range undoubtedly caused by error in the temperature scale which is discussed in Section B. One of the standards should be run in the instrument any time there is reason to suspect shifts in cal ibration or unreliability of the instrument. This would also serve to establish whether difficulties originated in the sample or in the instrument. The vapor pressures of samples of Santowax OMP, 5$ biphenyl 95$ Santowax OMP, and biphenyl were measured with the Instrument and are shown in Figure 16. Since these materials are solid below BO^C., it was necessary to heat the lines and strain gage to about 100C. to avoid blockage in the lines. The heating of the strain gage caused a small shift in the pressure axis calibration. No correction was made for this and it did not appear worthwhile to recalibrate the Instrument for these samples when the instrument was to be used primarily with materials liquid at ambient tem perature . B * Design Considerations This section Includes a discussion of the factors involved in the detailed design of the Vapor Pressure and Thermal Stabil ity Apparatus, especially those concerned with selection of com ponents and circuits for the Instrument. 1.. Sample system and furnace The early plans for this Instrument were based upon the use of a bomb type of cell. This was to have been a container similar to the Farr combustion bomb, with the addition of a glass liner so that there would be no contact of sample with the metal of the bomb, and with a tubing connection to the pressure trans ducer. However, the problem of finding materials of construction which would permit repeated assembly after exposure to a tempera ture of 500C., and tight sealing at that temperature, seemed In superable. Therefore the design was changed to a one-piece cell, to be permanently connected to a system of valves through which evacuation and filling would be accomplished. The cell shown in Figure 1 was bu.llt from type Jl6 stainless steel pipe. The fur nace was constructed as described In the instruction manual, ex cept that originally no blower was incorporated. It was thought at that time that stratification of the heated air within the furnace would not produce significant, temperature gradients In the cell. 0692723 HARTOLDMONOOQ5258 Experimental runs on water and on n-dodecane In the cell and furnace unit thus assembled gave apparent vapor pressures higher than the literature values at all temperatures, and this positive error persisted when the furnace was cooling as well as when it was being heated. The fact that the pressure deviation did not reverse upon reversal of the furnace heat Input showed that Is was not due to existence of a radial temperature gradient In the cell (although such a gradient undoubtedly was present). The explana tion appeared to be that there was a vertical gradient of temper ature along the length of the cell; the liquid-vapor Interface was near the top of the cell, where the temperature was high, while the thermocouple rested in the well near the bottom of the cell in a region of low temperature. Measurements with fine wire thermocouples attached to the out side of the cell showed a vertlcalHemperature gradient of about 10 at low furnace temperatures and about 30 at the maximum tem perature. When the control thermocouple was removed from the well and clamped to the outside of the cell temperature control was im proved but the vertical temperature gradient was not much changed. In an attempt-!to Increase the thermal dlffuslvlty of the cell it was cut open, filled with 0.125 inch-diameter stainless steel, balls, and rewelded. However, there.r-esulted no significant change in the vertical temperature gradient, nor In the radial temperature gradient which was about 10 at a heating rate of 4 6 per minute. When a small air Jet was introduced through the bottom of the furnace, directed upward (but off center, so that it did not strike the cell), the vertical temperature gradient^disappeared while the radial gradient was not affected. . From these experiments it was concluded that l) the cell should be constructed of material of high thermal dlffuslvlty, in the thinnest possible layers, and so designed that the leaBt possible thickness of liquid sample would be Interposed In the direction of heat flow, and 2) mixing of the air in the furnace was essential to avoid vertical temperature gradients. A. new cell was built, of annular construction as shown in Figure 9 of the Instruction manual. The dimensions were selected on the basis of availability of commercial tubing, and to give a volume of about 20 ml. The thermal diffusivlties (ratios of ther mal conductivities to heat capacities per unit volume) of some possible materials 6f construction and of a Thermlnol fluid are given below. Material Dlffuslvlty, Silver Gold Copper Aluminum Flat Intun 1.7 1.2 1.1 0.81 0.26 0692 724 . ... l) HARTOLDMON0005260 24 HARTOLDMONOOQ5261 25 Figure 15. Dichlorobenzene standard 0692727 HARTOLDMONOOQ5262 26 0692 728 HARTOLDMONOOQ5263 Material Dlffuslvlty, cm2/sec 27. Tantalum ' Nickel Stainless steel, type 304 or 316 K-Monel . Inconel Hastelloy-C . Glass Thermlnol FR-2 0.22 0.17 O.053 0.042 0.039 0.036 0.007 O.OOO63 Considerations of strength at high temperature, resistance to corrosion and commercial availability In a selection of tubing sizes limited the choice of material to type 304 Btalnless steel. This Is an alloy of low thermal dlffuslvlty, quite Inferior (to the extent that this particular property Is a suitable Index of merit for this application) to such metals as nickel or platinum. However, comparison of the dlffuslvlty value for type 304 stain less steel with that for-Thermlnol FR-2 fluid shows that reduc tion of the thickness of liquid layer In the cell.ls a more im portant factor In heat transfer than would be an Increase of the thermal dlffuslvlty of the wall material. The pressure which a cell constructed as shown in Figure 9 of the Instruction manual could withstand was calculated from the tensile properties of the stainless steel. The tensile strength and yield point of type 304 stainless steeltare: Tensile strength (l) 25C. 8.5 x 104 psi 500C. 4.9 x 104 Yield strength (l) 25e'C. 2.8 x 104 psl 500C. 1.2 x 104 Tensile stress to rupture in 1000 hours (2) 500C. 1.8 x 104 psi (1) National Bureau of Standards, Circular C447, p. 257. (2) Allegheny Ludlum Steel Corp., "Stainless Steel Hand book," 1951, PP. 2, 58. The Barlow formula P = 2S (t/d) gives the pressure corresponding to a given tensile stress S and ratio of wall thickness to outer diameter t/d for a cylindrical tube. For the dimensions used, t/d = 0.028, and the calculated pressures are t ' Internal pressure to burst (S = tensile strength) 25C. 4800 psl 5006C. 2700 0692729 HARTOLDMONOOQ5264 ?8. Internal pressure to burst in 1000 hours (S = 1U00 hour rupture stress) ' 500C. 1000 psl Internal pressure to produce permanent set (S = yield strength) 25C. 1600 psl 500C. . 670 The cell pressure at which collapse of the inner wall would occur was estimated from curves given In "Metals Handbook," American Society for Metals, Inc., 1948, p. 154. For the inner tube, t/d = 0.047 s Cell pressure to collapse Inner tube 25C. 2600 psl 500C. 1100 By these calculations, then, the cell should withstand more than three times the anticipated working pressure of 500 psla for 1000 hours at 500C. Failure with Increasing pressure would occur by collapse of the Inner wall of the annulus at about 1100 psl at this temperature. Application of somewhat more than twlle the working pressure would produce a permanent deformation of the outer wall. The cell was tested hydrostatically at 25OO psl at room tem perature. This pressure is about one-half of the calculated bursting pressure, but well above the deforming pressure, and It was noted that the cell exhibited a marked barrel shape after testing waB completed. This was not considered to be deleterious to the safe use of the cell, which was mounted in the furnace unit. The design of this cell had been planned to include a tu bular pocket for the thermocouple, welded to the outside wall of the cell at the middle. This feature, however, was omitted be cause of the possibility that welding on the thin tubing of the cell wall might weaken the wall excessively. Instead of using a pocket for the thermocouple, It was merely pressed against the cell by a band of thin sheet stainless steel wrapped tightly around cell and thermocouple. The second desired modification of the furnace assembly, pro vision for circulation of air within the furnace, was obtained by installing a motor-driven blower. The driving motor was mounted on the underside of the furnace base, with an extension shaft to a blower wheel placed Just above the bottom of the furnace. The blower consisted of a paddle wheel of 0.015 Inch thick stainless steel, having 6 radial blades 0.88 Inch high and 0.5 Inch wide arranged Inside a 2 Inch diameter circle. Although the operating speed of 1675 rptn was relatively low for this small blower, smoke tests with a glass beaker substituted for the furnace cover showed that it produced rapid circulation and thorough mixing of the en closed air. 0692730 HARTOLDMONOnn.^fifi 29. The performance of the modified cell and furnace unit was tested by attaching fine wire thermocouples to the outside of the cell near the top and the bottom, and recording the cell tempera tures, as well as the air and control thermocouple temperatures, during heating. Two heating rates, 5C. and 20C. per minute, were used. At neither rate was any systematic difference of tem perature between the ends of the cell noted! At a heating rate of 20C. per minute the air temperature was about 40 higher than the cell temperature; at the 5 rate the difference was about 15. In both cases the control thermocouple, located at the center of the cell, was cooler than the end thermocouples. The difference was about 10 at the high rate and 5 at the low rate; it was ascribed to lag of response of the relatively massive, sheathed control thermocouple. As there would also be a lag of temperature of the liquid within the cell, behind the temperature of the outer wall, the actual error of temperature Indication would.be reduced. The Radiation shield surrounding the cell was not a part of the original design, but was added in an attempt to eliminate what appeared to be erroneously high vapor pressures in the lowtemperature region. The experiments described in Section A1 showed, however, that the actual cause of the observed high pres sures waB not radiative heating of the cell but Incomplete degas sing before introduction of the sample. ..The shield was retained because of Its possible utility in reducing radiative heat trans fer to the outer wall of the cell, with consequent higher temper ature there than at the Inner wall. 2. Temperature recording and control system In the design of a system to provide recording of temperature on a scale linear In reciprocal Kelvin temperature, the uie of a platinum resistance thermometer element was first considered. Relatively simple circuits for obtaining linear rotation of. a slidewlre in terms of Celsius temperature have been developed. They use Wheatstone bridge measuring systems, and It would seem to be an obvious adaptation of such a system to produce a linear readout In reciprocal Kelvin temperature. The Initial plans, how ever, called for the sensing element to be inserted In a well In the sample cell. For this use the most compact possible sensor was desired; In addition. It was necessary that the sensor have leads suitable protected for extension through the furnace at a temperature up to 500C. No platinum resistance thermometer meet ing these requirements was found. In addition, the rather high cost of resistance thermometers generally was also a deterrent to the use of this sensing element. for these reasons a thermocouple was selected Instead for measurement of the cell temperature, and an electronic circuit was designed to obtain the desired linear scale of reciprocal absolute temperature from the thermocouple input. The application of a thermocouple to recording of temperature on a scale linear In l/K. depends upon the observation that, for 0692731 HARTOLDMONOOQ5266 an iron-constantan thermocouple, the function log (E + k) is very nearly linear with respect to reciprocal absolute temperature - E being the emf of the thermocouple referred to 0C., and k an arbi trary constant. Figure 17 shows the fit obtained for two values of k. The curve most nearly conforms to a straight line, from room temperature to 500C., for k = J.O mv. An exact fit can be obtained at 50, 150 and 400C. With three adjustments available thermocouple offset voltage, recorder span and recorder zero - the recorder scale could be made to fit a scale of reciprocal absolute temperature at any three points. If the three temperatures listed were selected, the maximum temperature errors would be about -5 at 250"C., +10 at 500C., +2 at 100C. and -4 at 25C. It was thought that these errors, representing at the higher temperatures 2$ of the scale reading, would be acceptable, and the Instrument was first set up In this way. The error at 25C. was considered unimportant; however, a change was made to 50, 200 and 400C. as the calibration points In order to obtain a better fit in the 200C. region. When the Instrument was put Into use, the temper-.' , ature error in the region of 100C. was found to be objectionably - < large, amounting to about +6. The exact source of this error was not identified; possibly some non-linearity of the logarithmic com pressor was present. Redrawing of the recorder chart with dis tortion of the temperature scale to produce the desired calibration was considered as a method of correcting this error, but it was de cided that retention of the theoretical reciprocal absolute temper ature axis was preferable to a change here. Instead, the temperature scale was corrected by the Introduction of a non-linearity in the recorder function of voltage versus pen travel. In Figure 17, the curve for an offset of 2.0 mv shows a linear relation of log (E + k) to 1,/"K at temperatures above 150C.; at lower temper atures the thermocouple voltage drops more rapidly than is desired. This Is also the shape of the curve of output voltage versus tap position of a resistive voltage divider the lower branch of which is shunted by a fixed resistance. Then by the introduction of such a shunt in the recorder potentiometer circuit it should be possible to match the log (E + k) curve. This was done experimentally: resistance values in the recorder circuit were manipulated, while varying the calibration adjustments already present, until a sat isfactory fit of the temperature scale was secured. With this modification, the error of indicated temperature was reduced to +2' at 100C. and -3 at 25C. when the calibration was exact at 50r, 200" and 400"C. The logarithmic compressor had a full-scale Input of 100 volts. Therefore, an amplifier which would produce this voltage from the thermocouple output was required. Few of the large number^of com mercial d-c amplifiers have this output capability. Of ttiose which do, the Philbrick UPA-2 amplifier was selected. This amplifier has adequate performance specifications with regard to gain and drift, and the cost is relatively low. The amplifier is supplied with a filament transformer, but without plate power supply because it is intended for use In analog computer applications where a common 1) .) 0692732 HARTOLDMONOOQ5267 l* ' /* ' ' ~~ri ' t.e /o V ' j`o Figure 17. Thermocouple characteristic as a function of 1/K TC 0692733 HARTOLDMONOOQ5268 power supply serves several amplifiers,, In the present applica tion, however it seemed better to keep the temperature and pres sure systems entirely separate and thus avoid having to rely on the electrical insulation of the thermocouple and pressure trans ducer. To obtain Isolation, it was necessary to add a simple power supply to the amplifiers. The Phllbrlck UPA-2 amplifier comprises a chopper-type d-c amplifier, followed by a direct-coupled amplifier. The circuit is shown in Figure 11 of the instruction manual. The chopper am plifier section, of two stages, is bypassed by an alternate signal path to the direct-coupled amplifier of two amplifying stages and a''Ca5>9de. follower. The bypass provides the signal path for ampli fication of signals above a few cycles per second in frequency. Although response to rapid variations of signal is not a require ment here, the high frequency response of the amplifier may assist in making it more stable when used in a feedback loop. No diffi culty was found in applying this amplifier to feedback-stabilized amplification with the required gain of 3000. However, the rela tively small amount of amplification in the chopper amplifier made the unit sensitive to line voltage changes, which were reflected as small variations of apparent input voltage. Presumably this effect arises in variation of the grid-cathode contact potential of the third stage as the heater voltage changes, and in variation of the fourth stage plate voltage as the supply voltage changes. Damping of the amplifier by addition of a capacitive feedback path, as shown schematically if Figure 12 of the instruction manual, re duced noise In the amplifier output appreciably but there were-stlll some disturbances. The amplifiers in the temperature and pressure systems were first Installed with their power supplies connected directly to the line; later, in an attempt to Sfeduce noise, they were reconnected through a constant-voltage transformer. This transformer had only a small effect on the noise originating from random line voltage changes, but it did reduce greatly a previouslyobserved shift of temperature and pressure readings caused by line voltage drop when the furnace heater came on. To convert the amplified thermocouple voltage to the desired logarithmic function, there appeared to be two possible mechanisms: use of a logarithmic recorder, and use of a linear recorder pre ceded by a passive logarithmic converter. The use of direct logar ithmic recording would have several advantages. No separate logar ithmic converter would be required; the recorder, unlike a logar ithmic converter, would draw no power from the amplifier; and the logarithmic characteristic,; being established by the recorder slidewire, would be perfectly stable. On the other hand, there Is the disadvantage that no logarithmic x-y recorder is available, and that to convert a linear recorder to this use would Involve much modification and experimentation. To avoid such modification, it was decided to apply a logarithmic converting element at the input of a standard linear x-y recorder. 0692739 HARTOLDMONOOQ5269 The available methods of transforming from a voltage to a second voltage related to the logarithm of the first Involve the use of either thermionic or semiconductor diodes. (A logarithmic servovoltmeter which is available commercially could not be con sidered because of the high cost.) Ranges as large as 7 logar ithmic decades have been covered with thermionic diode converters. For the present problem a range of only 1 decade was required In the temperature system, and 3 decades in the pressure system. The Kane C-7A Logarithmic Compressor is a commercial semiconductor diode converter with a 3-decade range, which was selected for use here. The conververters were supplied In small cabinets with self- contained batteries. Except for two adjusting potentiometers, the circuit elements all were incorporated on a component board In each unit. The component boards were remounted in the recorder console of the vapor pressure instrument, with connections to a remote battery assembly. The original adjustments were replaced with fixed resistances, the need for adjustability having been eliminated by the use of stable mercury batteries and by the ad justments In the recorder. Originally the logarithmic compressor component boards were simply mounted on the panel of the Instru ment, It soon was apparent that large shifts of zero on both axes of the recorder were being produced by variations of temperature of the logarithmic compressors. The two transistors which comprise the output cathode follower of these units (see Figure 13a of the instruction manual) seemed to be extremely sensitive to variations of temperature. Although these transistors are of opposite type and therefore have opposite effects on the sign of output voltage variation with temperature, the cancellation of temperature effects was not complete. It was necessary to remount the component boards of the logarithmic compressors lr. a thermostatted box to suppress output voltage variations attributable to temperature changes. The box was made of aluminum with heaters dissipating 3*3 watts each, on each of the two larger opposite sides. A thermostat attached to the box near one of the heaters allowed a cycle of about1''2'C. at that location, which produced a cycle of about 0.2c,C. amplitude in the interior of the box. Instability of the thermostat was a pro blem in this application. The unit used, a Fenwal 32400-0, had an adjusting range of about IOO'C. per revolution of the adjusting screw. With such a wide span, slight shifts of position produced by pressure on the thermostat body or terminals had a significant effect on the set temperature. It was found that to secure regu lation at 40cC. with a stability of +0.5cC. the insulation aroundj) the box had to be cut away to prevent exertion of any force on the thermostat. Use of a unit having a narrower span of adjustment, with correspondingly lower sensitivity to extraneous influences, would have been desirable. ^ The temperature control circuit was designed to utilize the amplified thermocouple voltage already present In the temperature recording system. Linear time programming of this voltafes was considered sufficient. This does not quite correspond to a linear time Increase of temperature, but the difference should be of no 0692735 HARTOLDMON0005270 34. concern In this application.. With the specified amplifier gain of 3000, a temperature change of 1C. at the cell corresponds to a voltage change of about 0.16 volt at the amplifier output. This change is not great enough to operate any simple relay directly, but with moderate further amplification a relay could be made to operate on temperature changes of l^C. or less. Possible relays which were considered were semiconductor switches (silicon con trolled rectifiers), gas tube switches (thyratrons), and mechani cal relays. The power requirement- of the furnace heater was approximately 4.5 amp at 115 volts. This current is too large to be handled.by a small thyratron but would not be excessive for either a mechan ical relay or a silicon controlled rectifier. Use of the latter type of switching element was considered, but was rejected partly because of lack of familiarity with silicon controlled rectifiers and partly because such a circuit seemed more elaborate than was necessary for the degree of temperature control required. There fore the mechanical relay was chosen: specifically, a mercury plunger relay in order to obtain durability and quiet operation. The coll dissipation of a suitable mercury relay is in the range of 5 watts. Although hard-vacuum tubes are available which would readily supply this power, a simpler choice seemed to be the use of a thyratron operating.directly from an a-c plate supply. The thyratron requires a control voltage swing of about 1 volt, which could be obtained from a single vacuum tube amplification stage. (The amplifier must invert, so that the thyratron is non conducting when the signal voltage is high, if a normally-open relay is to be used.) The circuit shown in Figure 10 of the in struction manual was designed on the basis of these considerations. The amplifier stage was arranged for differential input, with the control signal coming to one grid of a twin trlode and a reference signal to the other grid. By making the reference signal adjust able, provision for adjusting the set point of the controller was supplied. The active half of the difference amplifier was a sim ple direct-coupled etage, with a voltage divider in the output to reduce the voltage level to the cathode level of the thyratron. The time programming was to be controlled by a voltage from the timer-driven potentiometer. Because the amplifier output vol tage would vary over a range from about 10 to 90 volts, it did not appear to be feasible merely to apply the programming voltage to the reference grid of the difference amplifier. Instead, the signal and programming voltages were added, with inversion of the programming voltage, and the relay circuit was arranged to turn on when the voltage sum was positive or off when it was negative. The simple adding network used for this purpose incorporated a variable voltage ratio, so that the time rate of Increase of the signal voltage, and therefore of cell temperature, could be varied. 0692736 HARTOLDMONOOQ5271 The temperature programmer circuit was built originally without the 0.068 mfd capacitor In shunt with the amplifier load resistance. When the circuit was tested it exhibited a hysteresis of about 2C.: that is, when the heater relay turned on, it was then necessary for the temperature to rise 2 before it would turn off, and conversely for riEing temperature. The cause of this effect seemed to be some interaction between the thyratron plate circuit and the amplifier circuit. Whatever the cause was, the effect was eliminated by by passing the amplifier load resistance. A second later addition was the 100 mfd capacitor shunting the signal branch of the voltage adding circuit. The capacitor was incorporated to provide some anticipation of temperature changes, and thus improve the temperature control. It should have been pos sible to obtain the same result, and more effectively, by shunting with a capacitance the upper arm of the voltage divider between the difference amplifier plate and the thyratron grid; but when an attempt was made to do so, the thyratron conducted continuously at all signal voltages. This result is inexplicable from a superficial analysis of the circuit. As in the case of the plate load shunt, it probably arises by Interaction between the output and input cir cuits of the controller. The added 100 mfd capacitor had fa.small effect on the cell temperature regulation obtained, a reduction in the amplitude of temperature cycling and increase of the frequency of the cycle being observed. The amplitude of temperature variation was about AC. and the frequency about 0.5 per minute. As was mentioned at the beginning of this section on design, the original choice of a thermocouple as the temperature sensing element (and following from this, the selection of the circuit elements to obtain linear recording of reciprocal absolute temper ature) was based upon the belief that a resistance thermometer would be unsuitable in physical form for use with the original cell. The cell design, however, was changed; placing the temper ature sensor in a small well was no longer a requirement. Also, compact resistance thermometers designed to have both head and leads exposed to high temperature have become available. Thus if the design were to be done again a resistance thermometer might well be chosen as the temperature sensor. A surface unit, such as Winsco 2512, should be suitable. The higher cost of the resistance thermometer would be compensated by elimination of the amplifier and logarithmic compressor required for the thermocouple system, and the higher voltage level available from the resistance ther mometer would appear to simplify the electronic design generally. In section A1 there was mention of premature battery failures. These apparently resulted from the purchase of several batteries which had been stored too long by the distributor. Batteries from another distributor gave normal service. Use of line-operated power supplies Instead of batteries was considered, but the 7 powersupplies which would be required would have Involved excessive bulk and expense. 0692737 `^ ' j HARTOLDMONOOQ5272 The original design Incorporated Mallory type TR-233 batteries for B3 and B5, which supply the output voltages of the logarithmic converters. These batteries exhibited a continuing decrease of voltage when In use. Replacement of the TR-233 batteries with the equivalent series arrangement of RM-3R cells gave a stable voltage. Although the TR-233 battery is physically composed of three RM-3 cells in series, there is apparently some difference In the man ufacturing process for the cells which results in lower stability of voltage than that characteristic of individual RM-3R cells. 3. Pressure recording system In order to detect both the effects of volatile impurities in a sample (producing an increase of pressure at low pressure) and the effects of unstable but non-volatile impurities (causing an increase of pressure at high-temperature and therefore at high pressure) it was considered necessary for the pressure recording system to cover the widest possible span of pressure. However, hysteresis and zero shift of the pressure transducer, and zero drift of the amplifier, limit the minimum pressure which may be measured with a transducer of given upper pressure limit. The range of logarithmic conversion available with the Kane C-7A Logarithmic Compressor was 3 decades. Since the commercial pres sure transducers usually are rated at 1% of full scale hysteresis error, it was thought that a 3-decade recording span would more than exhaust the ability of the transducer to measure low pres sures. Experience however gave the opposite result. It was nec essary to reduce the range of recording to 2.8 decades because of failure of the logarithmic compressor at the lowest voltages, a failure probably caused by operating the compressor at 40C. in the thermostat Instead of the nominal 25'JC. Over this range the pressure transducer showed no significant lack of repeatability. Reproducibility of the sero point of the gauge was within 0.1 psl. Preliminary calculations had indicated that an upper pressure recording limit of 500 psla would be required. A 500 psla range" pressure transducer was obtained and installed, but the first tests of the instrument indicated that the transducer was defec tive. To save time, the 500 psla unit was exchanged for a 300 psla transducer which the manufacturer happened to have. Eventual replacement of the 500 psla transducer was planned, but the lowerrange unit proved to be entirely satisfactory and was retained for permanent use. 4. Programming system The features desired in the temperature programming system of the vapor pressure apparatus were production of an approximately linear temperature ri^p over the selected span, with a minimum of attention from the operator. Accurate linearity and exact repro ducibility were not sought, since these factors should have only secondary effects on the vapor pressure versus temperature func tion. Operation of the heater from a variable transformer 0692738 HARTOLDMONOOQ5273 37. provided with a time drive was first considered. In its most slmpie form, such a device would require resetting by the operator af ter each run, but automatic resetting could be provided. For exam ple, the variable transformer might be driven by two motors through a differential, one motor driving forward during programming and the other driving backward rapidly for resetting when the program was completed. Variation of the rate of temperature rise could be obtained by feeding the motor-driven variable transformer from (or into) a second manually-set variable transformer which then would serve as the rate of rise control. Such a system is fundamentally simple, and has the advantage that the voltage applied to the fur nace Is continuous Instead of intermittent as in on-off types of control. Disadvantages are the considerable weight and bulk of the variable transformers and the rather high cost of the mechan ical components required if automatic resetting is to be incorpor ated. An additional disadvantage Is that there would be no control of temperature as such, but only of voltage applied to the heater. Because a temperature recording system was to be incorporated in the instrument, whatever type of temperature control was used, it seemed to be feasible and desirable to include an electronic temperature controller utilizing the components already present in the recording system. The programming element of Buch a con troller would be a variable resistance, which would simplify the problems of time drive and of automatic resetting. Although schemes for electronic proportional control have been simplified by the Introduction of controlled semiconductor rectifiers. It was thought that not even this degree of complication would be needed and that on-off control would suffice. A control circuit was de signed as described in section 2C, to be actuated by differences of voltage between the amplified thermocouple output and a timerelated reference voltage. The simplest method of obtaining the reference voltage appeared to be, to attach a potentiometer (sup plied with a constant voltage from a regulated power supply) to a reset timer of which several are commercially available. The me chanical design of the different makes of timers apparently is similar: a motor drives the timer mechanism through a solenoid clutch, release of which at the end of the timing period permits the mechanism to be reset by a spring. An Eagle Signal Co. model HP2 timer was chosed for use here. The setting knob of the timer was removed, and a bushing was added for attaching the reference voltage potentiometer. The return spring of the timer provided only a small torque, so a potentiometer of low torque requirements was necessary; a Beckman model G unit was chosen. The potentio meter shaft was coupled to the timer Indicator by a slotted tongue soldered to the potentiometer shaft and engaging a slot in the timer shaft. Initial trials of the assembly indicated that the torque of the timer spring was sufficient to rotate the potentio meter, but when the unit was Installed In the vapor pressure recor der several Instances of sticking were observed. The original timer spring was replaced with a heavier spring, which gave reli able resetting action. Probably the original spring would have 0692739 . ) HARTOLDMONOOQ5274 38. been adequate If a potentiometer of lower torque requirement, such as the Beckman model T, had been used. The timer Incorporated contacts operated by the clutch solenoid, which were utilized to obtain the desired self-holding function and to control the line power applied to the heater control relay. De sign of the self-holding circuit, to be energized by the START but ton and reset by operation of the STOP button or by opening of any part of the limit switch circuit, was entirely conventional. The limit switches on the two axes of the recorder, and the limit ther mostat in the furnace, were introduced as safety features. They are not expected to operate during a normal run, in which comple tion of the programming would be marked by completion of the selec ted time period with subsequent removal of power from the heater and automatic resetting of the timer. In case of some failure of the programmer or temperature controller, however, the temperatureaxis limit switch would prevent a dangerous Increase of furnace temperature. The occurrence of a pressure increase above 300 psla, at a temperature below the 500C. limit, would trip the pressureaxis limit switch and thus avoid further heating. Finally, should heating Bomehow continue even after opening of either recorder limit switch, the furnace limit thermostat would open at about 600C. to cut off the heater. Programming of the temperature by the temperature control and programming sections of the apparatus seemed to be adequate for the Intended use. A typical record of temperature versus time is shown in Figure 18. At the start of the program there was a rapid rise to about 65C. After the initial rise the average tempera ture Increased linearly to the maximum. The program had been set for 75 minutes at 5C. per minute. Actually, the temperature in creased 390C. in 75 minutes, or 5.2 per minute. Both the ini tial rapid rise of temperature and the cyclic variation superim posed on the temperature curve are probably to be ascribed to the characteristics of the furnace, rather than to any property of the control and programming system. Presumably the lag of response of the thermocouple to heat input at the heater is the cause of both the initial overshoot and the cyclic fluctuation. For most of the materials for which this apparatus was designed, no measurable vapor pressure is found at temperatures up to about 70C., and deviations of heating rate in this region are unimportant. The cyclic variation of temperature at the higher temperatures did produce a slight fluctuation of vapor pressure. The pressure var iation was easily seen on a linear recording of the pressure on a separate recorder, but did not have any great effect on the pres sure versus temperature recording. If the pressure and tempera ture variations were exactly in phase there should, of course, be no effect at all on the recording, but they might be expected to be out of phase in practice because of the temperature lag between thermocouple and sample. 0692 740 HARTOLDMONOOQ5275 6. Recorder The use of an x-y recorder for plotting vapor pressure as a function of temperature seemed to be an obvious choice. Several makes of x-y recorders are available; the Houston Instrument Corp, model HR-92-1 was selected because of its low cost. Although this recorder was supplied in a case, It was considered preferable to remount It in the electronics console of the vapor pressure instru ment in order to obtain a compact construction. Extensive circuit modifications would have been required in any event, so the orig inal calibration controls were discarded, to be replaoed by cor responding components in the console. The original measuring circuits of the x-y recorder consisted of potentiometer bridges having output voltages equal to 6 times the maximum voltages to be measured on the two axes. The potenti ometer voltages were then reduced by 6:1 voltage dividers in the recorder amplifiers for comparison with the voltages to be meas ured. This arrangement is satisfactory for the maximum measured voltage of 70 or ; `0 mv for which the recorder was designed, and has the advantage feat spurious emfs in the potentiometer cir cuits are Insignificant in comparison to the output voltage of 420 or 600 mv. In the present application, however, voltages up to nearly 400 mv were to be measured. With measured voltages of this magnitude, division of the potentiometer voltage is neither necessary to reduce interferences nor possible with a single cell power supply. Therefore the voltage-dividing networks were re moved from the recorder amplifiers, converting them to direct com parison of potentiometer and measured voltages. The recorder was intended to accept standard graph sheets hav ing a 7" x 10" grid printed on 8.5" x 11" paper. Apparently there is no commercial stock of graph papers having reciprocal absolute temperature and logarithm of pressure as the scale arguments. For temporary use a master graph was drawn, to be copied by "Xerox" or similar process. It was thought that if use of the instrument con tinued for a long period, it would be desirable to have a supply of graph sheets produced by some more accurate printing process. ACKNOWLEDGEMENT The authors are appreciative of R. H. Munchrs interest and helpful consultation In testing the instrument and to 0. Hicks and G. M. Gasser for analytical work on the project. smm. Instruction Manual. -to.-vi.-c:--< Vf. N. Trump C. E. Vogler 06927A1 ) ) HARTOLDMONOOQ5276 Figure 18. Temperature record 0692742 HARTOLDMONOOQ5277 INSTRUCTIONS FOR VAPOR PRESSURE AND THERMAL STABILITY APPARATUS 0692743 HARTOLDMONOOQ5278 ' CONTENTS 1. INTRODUCTION Page No. A. General description and theory .......... 1 B. Operating controls ................ 2 2. PRINCIPLES OF OPERATION A. Introduction ................... 5 B. Sample System and Furnace ............ 4 C. Temperature Control and Recording System ..... 5 D. Pressure Recording System ...... .................... . 8 E. Programming System ................ 9 F. Recorder .............. ............................. 11 OPERATING PROCEDURES A. Detailed Procedure .......... ......................... 1$ B. Discussion .................... 16 C. Vapor Pressure of Thermino.l Fluids ........ 18 4. MAINTENANCE A. Introduction ........... ................................... 18 B. Routine Maintenance ......... .................... 18 C. Repairs and Replacements ............. 19 D. Adjustment and Calibration ............ 22 E. Trouble-shooting ............ .................... 25 F. Parts List ............. ............................. ?1 0692744 HARTOLDMONOOQ5279 1. Section 1. Introduction A. General description and theory If a solid or liquid substance is introduced into an evacu ated, confined space, evaporation proceeds until the pressure in the space has risen to the equilibrium vapor pressure of the sub stance at the prevailing temperature. A famous equation, derived by Clapeyron, gives the temperature derivative of the resulting pressure in terms of the heat of vaporization of the substance and the volume expansion accompanying vaporization: dP/dT *= AH/T&V If it is assumed that AH is Independent of temperature, that the vapor is an ideal gas, and that the volume of condensed phase evaporated may be ignored in relation to the volume of vapor formed, this equation may be transformed to: dP/dT = ah/(RTVp) which is integrated to: In P - -AH/RT + C C being a constant of Integration. Thus, for a substance meeting the specifications above, the logarithm of vapor pressure would be a linear function of reciprocal absolute temperature. Real substances do not have heats of vaporization which are invariant with temperature, their vapors do not conform to the law of ideal gases, and the expansion on vaporization is not so great that the volume of condensed phase dissipated in vaporization may be neglected. Nevertheless, most real substances show a variation of logarithm of vapor pressure which Is nearly enough linear with respect to reciprocal of absolute temperature to make the plotting of these quantities a useful method of recording and interpolating vapor pressure data. The vapor pressure and thermal stability apparatus, shown in Figures 1 and 2, is a laboratory Instrument for recording automatically the curve of logarithm of vapor pressure as a func tion of reciprocal of absolute temperature. The sample may be any substance liquid at room temperature: the recording range ex tends from room temperature to 500'C., or to a pressure of 300 psla if this pressure occurs at a lower temperature. Rise of tem perature is programmed automatically, at a rate adjustable from 1 to 25' per minute, for any time up to 150 minutes or to the maximum temperature of 500C. Heating is terminated without oper ator attention at the end of the selected time; or if the recorder temperature reaches 500C., or the pressure 300 psla, before that time; or if the furnace temperature reaches 600C. regardless of sample temperature or pressure. Curves are plotted on a special 0692745 HARTOLDMON0005280 grapn paper of 8.5 Inch by 11 inch format. By comparing the re- corded curve of a specific sample of Therminoi or other fluid with a master curve prepared from material of known properties, the operator can detect the presence in the test sample of objection able amounts of low-boiling Impurities or of substances producing , Increased decomposition rate at high temperature. Structurally the Instrument consists of two units: a furnace assembly and a recorder console. The furnace unit contains the sample cell and associated filling and emptying connections, and the pressure and temperature transducers. It is connected by ca bles to the recorder unit which contains the electronic equipment and the x-y recorder for plotting the desired log P versus l/T function. B. Operating controls Operation of the Vapor Pressure and Thermal Stability Apparatus requires both manipulations of the sample and operations of the electrical controls. Introduction and degassing of the sample in volve the use of two stopcocks on the glass sample funnel, and three valves on the furnace base (see Figure P). STOPCOCK A, STOP COCK B and the INLET valve admit liquid from the sample funnel to the cell. The BVPASS valve connects the sample inlet directly to the outlet, for flushing and drying of the inlet connections. The OUTLET valve connects the sample cell to the outlet (which, in use, is connected to an evacuated trap). Electrical controls which are operated routinely during recor ding of vapor pressure curves are located beside the recorder, as shown in Figure 5. The OFF-ON switch controls all power to the instrument. The LOAD-.OFERATE switch, when in the LOAD position, causes the recorder pen to move to the upper right-hand corner of the platen so that a sheet of graph paper may be inserted or re moved conveniently. In the OPERATE position, the recorder pen traces the pressure-temperature record. Pressing the START but ton initiates the automatic cycle of temperature programming; the associated pilot lamp shows that the program :s running. Pressing the STOP button terminates any program vnich ie in progress, and resets the programming timer. The pilot lamp beside the STOP but ton glows when the instrument i.e in standby condition (ready for the start of programming). Tvo controls which are operated only to change program dura tion or rate of temperature rise are located on the front panel of the console. Figure 4. The TIMER is set for the desired pro gram time. Heating begins when the START button le pressed, and continues for the selected time (unless the limit of temperature or pressure Is reached before the time expires). The RATE OF RISE control Is set for the desired rate of increase of temperature. ^ . | 0692746 HARTOLDMONOOQ5281 3. The maximum temperature attained during the run Is equal to the product of cycle time (TIMER setting) by the temperature rise per minute (RATE OF RISE control setting) unless the product exceeds 500C. As described In section 1A, heating is interrupted auto matically at this temperature. Other controls in the recorder console are used for calibra tion and adjustment. They are described In section 4D. Section ?. Principles of operation A. Introduction This section of the Manual Includes detailed descriptions of the elements of the Vapor Pressure and Thermal Stability Appar atus. The Instrument consists of four, substantially Independent systems: Sample system and furnace Temperature control and recording system Pressure recording system Programming system The sample system consists of the sample cell and the associated components for handling the sample. The other three systems com prise the electrical mechanism for measuring and recording the vapor pressure curve of the sample. The block diagram, Figure 5, shows the relationship of the systems and the locations of the major components within systems. The sample system and furnace section includes the thermocouple for measurement of sample temperature and the pressure transducer for measuring the fissure exerted by the sample. The thermo couple output, passes through a reference -Junction compensating circuit to an amplifier; the amplifier output is applied through a logarithmic conversion element to the temperature axis of the recorder. The circuitry is so arranged that the recorder indica tions are linear in reciprocal absolute temperature. The ampli fied thermocouple voltage also is, applied to a comparison circuit In the temperature controller. The reference voltage of this circuit Is controlled by the program timer. If the amplified thermocouple voltage Is less than the reference voltage (Indica ting a temperature below the programmed temperature) the heater is turned on. If the voltage is greater than the reference vol tage (indicating high temperature) the heater Is turned off. Use of a logarithmic conversion for recording temperature on a linear scale of l/K. is described in the detailed discussion of the temperature system, section 2C. The pressure exerted by the heated sample Is detected by the pressure transducer, which is a strain-gauge unit energised by the power supply through the zero-adjusting circuit. The output 0692747 HARTOLDMONOOQ5282 voltage of the pressure transducer Is amplified, converted to a function of the logarithm of the amplified voltage, and recorded on a linear scale of log pressure. Programming of the sample temperature Is controlled by the timer, which runs until the selected time has expired unless oper ation Is interrupted by action of the recorder limit switches or of the limit thermostat in the furnace. B. Sample system and furnace , Figures 6 and 7 show the details of the sample system and furnace. The sample cell Is supported centrally In the furnace, surrounded by the helical heater. Below the sample cell is a small blower wheel for circulation of air within the furnace. The hlgh-temperature limit thermostat Is mourned vertically be tween the heater and the cell, and the thermocouple, also pro truding vertically through the bottom of the furnace, is clamped to the cell. The furnace base supports all of the components associated with the furnace. The bottom of the furnace is mounted on the base on a "Translte" spacer, and is filled with glass wool insul ation. The furnace cover consists of concentric shells, separated by a ''Translte'1 ring and having the annular space filled with magnesia insulation. The blower motor, pressure transducerpressure relief valve and sample system plumbing are mounted under the furnace base. Figure 8 shows schematically the arrangement of the sample system. The cell ha? two connections, so that through flow may be obtained for flushing and drying. Sample is introduced Into a glass sample funnel, which is connected to the metal sample system through a semi-ball Joint. Connections of the sample sys tem outlet to the trap and vacuum pump also is made by such a joint. A chamber In the sample funnel permits preliminary out- gassing of the sample, before it is Introduced Into the cell. The 'INLET valve admits sample to the cell : the OUTLET valve per mits removal of the sample and cleaning of the cell; and the BYPASS valve makes It possible to remove unused sample from the inlet and sample funnel without disturbing the cell. The pressure transducer is connected to one of the cell con nections. To the other is attached a relief valve opening at 600 psig. The pressure transducer is rated for 500 psla range, with permissible overrange of twice this pressure. Construction of the cell is shown m Figure 9- It- 1b a thinwalled annular cylinder, made of types 50^ and .516 stainless steel. This design provides rapid heat transfer both through the cel] walls and through the liquid within the cell, from the air 0692 748 HARTOLDMON0005283 5- which circulates downward around and through the cell. The cell is surrounded by a radiation shield to reduce heating by direct radiation from the heater. Cell temperature Is measured by the thermocouple which is held against the outer cell wall by a band of thin stainless steel. The heater, of the sheathed type with "Inconel" sheath. Is rated for operation at 250 volts and 2 kw dissipation. Operated at 115 volts it dissipates about 0.5 kw, providing sufficient heat for the furnace with moderate heater temperature. A limit thermostat is provided in the furnace to stop heating should the temperature in the furnace reach approximately 600C. This is a safety control, operative in the event of failure of the temperature recording system. C. Temperature control and recording system. This system comprises those components of the instrument in volved in the recording of sample cell temperature and in the regulating of that temperature in conformance with an established time schedule. The relationship of this system to the complete Instrument is shown in the block diagram. Figure 5> in which the temperature recording and control system occupies approximately the lower half of the diagram. The sensing element is an ironconstantan thermocouple in a stainless steel sheath, which is clamped to the outer wall of the sample cell. The thermocouple output voltage is corrected for reference Junction temperature by the compensating circuit and then is amplified by the amplifier which has a voltage gain of 3000. The amplifier output is applied to two circuit branches. In one branch the signal passes through a logarithmic compressor which converts it to a voltage propor tional to the logarithm of the amplifier output voltage. The resultant logarithmic voltage is applied to the temperature axis of the recorder, in the reverse direction. Deflection of the horizontal recorder pen position which is linearly related to the reciprocal of the absolute temperature is thus obtained. In the other circuit branch, the amplified signal is compared with a reference voltage from a potentiometer driven by the timer. The reference voltage is proportional to the elapsed time of the re cording program, while the amplifier output voltage is propor tional to the thermocouple output and thus (approximately) to the temperature of the cell. The difference of these two voltages is used to operate a relay applying power to the furnace heater. If the reference voltage is the greater (indicating that the cell temperature is lower than the temperature required by the time program) the heater is turned on; if the amplifier voltage is the greater of the two (indicating a high cell temperature) the heater conversely is turned off. In this way the cell temperature is made to rise linearly with time. 0692749 HARTOLDMONOOQ5284 6. The detailed circuit of the temperature recording and control section is shown in Figure 10. (in this figure the circuit ele ments and connections are complete, but the internal structure of the amplifier, logarithmic compressor and recorder is merely Indi cated schematically.) The thermocouple is connected through Inte gral leads to terminals on the back of the recorder console. The Junction between thermocouple wire and the copper part of the cir cuit is formed at this point. Compensation for the temperature at this Junction Is secured by a "Sensitor" silicon resistor mounted behind the panel carrying the thermocouple terminal board. The compensation circuit, energized by battery Bl, is a Wheatstone bridge one arm of which is the THERMOCOUPLE OFFSET control. This control permits the introduction of an adjustable voltage in ser ies with the thermocouple output, and Is used in calibrating the temperature scale of the recorder to bring the low-temperature end of the scale to the correct reading. The other arm of the compen sation circuit contains the temperature-sensitive resistor, with shunt and series resistances to give the required variation of 0.055 mv per degree temperature change for compensation of the cold Junction emf of the thermocouple. The compensated thermocouple output voltage ie applied to the input of the amplifier. This is a feedback amplifier with a gain of 5000. It consists of a modified Philbrick UPA-2 amplifier, the modification being addition of an individual power supply and of the feedback circuit components. The basic amplifier is a chopper-stabilized amplifier mounted on the printed circuit board shown in Figure 11. The external connections to the circuit board are shown in Figure 12. The signal enters the amplifier at terminal 2 of the circuit board, is converted to alternating vol tage, amplified, rectified and again amplified as direct voltage. The output appears at terminal 6 of the circuit board. A voltage divider consisting of 90 kllohm and 50 ohm resistors is connected across the output and common (number 1 and 4) terminals, and the Junction of these resistors forms the return terminal for the input circuit. Therefore the fraction 50/9 x 10* of the output voltage is applied In opposition to the input (because the polar ity of the output is Inverse to that of the input;, and the over all gain of the amplifier is very nearly the reciprocal of the feedback factor, or 5000. The series combination of 0.25 nifd capacitance and 10 kllohm resistance between output and input in troduces additional feedback for varying signals, decreasing the overall gain for such signals and by this means reducing the response to line voltage transients and other disturbances. The amplifier output voltage goes to the logarithmic compres sor, the output of which is a linear function of the logarithm of the input amplitude. Basically the logarithmic compressor is a voltage divider containing a semiconductor diode In series with a resistance. With a suitable steady bias voltage applied to the diode, the voltage appearing across the diode is linearly related ') i) ' _ 0692750 HARTOLDMONOOQ5285 7. to the logarithm of the voltage applied to the series combination of diode and resistance. The fundamental circuit Is indicated in the logarithmic compressor box of the diagram. Figure 10, and the complete circuit is shown in Figure 15a. Two diodes are used, in opposition, because the unit is one which was intended to operate on alternating as well as direct current inputs. A bias voltage is established by battery B2 in series with the 150 ohm and 2-28 ohm resistors. The input signal is applied to the 200 kilohm re sistance and the output appears at the junction of the diodes with this resistance. The output voltage, which is at a high Impedance level, is transformed to a lower impedance level by the two tran sistors arranged as a two-stage, complementary emitter follower. The transistors are powered by battery BJ. The components of the logarithmic compressor are mounted on a circuit board, the arrangement of which is shown In Figure 13b. The transistors, especially, are extremely sensitive to changes of temperature. To reduce the effect of ambient temperature var iations the component boards of both the temperature system and the pressure system logarithmic compressors are contained in a thermostatted box. The signal from the logarithmic compressor is Introduced into the temperature (horizontal) axis of the recorder. The re corder is a self-balancing potentiometer, which operates to keep the output voltage of the potentiometer circuit (at terminals 1 and 3 of the recorder amplifier) equal to the voltage from the logarithmic compressor (at terminals 1 and 2). The potentiometer circuit is of the Wheatstone bridge form, powered by battery b6 in series with the HORIZONTAL SPAN variable resistance. In one arm the HORIZONTAL ZERO adjusting potentiometer provides for shifting the zero point of the recorder scale. The slldewlre, mechanically coupled to the servomotor and recorder pen, is in the ether arm. The 2 kilohm shunting resistance across the lower part of the slldewlre produces a slight non-linearity of balanc ing voltage versus pen travel. In combination with the emftemperature characteristic of the thermocouple and the effect pro duced by logarithmic conversion, the recorder non-linearity gives a resulting pen motion which is nearly linear with respect to the reciprocal of the cell absolute temperature, over the range from 20^0. to 500G. The amplifier output is connected also to the input of the temperature control circuit. A resistance network composed of the 1 megohm RATE OF RISE potentiometer in series with 220 kilohm fixed resistors at each end is connected between the amplifier output and the slider of the TIMER potentiometer. The latter has applied to it a voltage which is negative with respect to the com mon terminal of the system. Thus, as the timer runs, the poten tiometer output voltage becomes increasingly negative. The slider of the RATE OF RISE potentiometer is connected to one grid of a 0692751 HARTOLDMONOOQ5286 twin-triode- difference amplifier; the other grid Is supplied with an adjustable reference voltage from the TEMPERATURE SET control. This control is adjusted so that the heater relay (controlled by the thyratron which in turn is controlled by the output of the difference amplifier) Just operates when the RATE OF RISE voltage is zero. The action of the temperature control system, then. Is to maintain zero voltage at the RATE OF RISE potentiometer tap by varying the cell temperature so that the amplifier output voltage balances the time-dependent reference voltage, and the ratio of these two voltages will be equal to the ratio of the resistances in the two branches of the circuit, measured from the RATE OF RISE potentiometer tap. When the slider is at the center point of the potentiometer, each 1 volt- Increase of TIMER potentiometer voltage requires 1 volt increase of amplifier output voltage, corresponding to a temperature Increase of approximately 6C,, At the clockwise limit- of the control the ratio is correspondingly 1.22/0.22 = 5.5 times as great; at the counterclockwise limit it Is the Inverse of this, or 0.18 times as great. The relationship between angular rotation of the timer per minute, and voltage applied to the TIMER potentiometer, is such that the potentiometer output voltage in creases 0.72 volt per minute. At the midpoint of the RATE OF RISE potentiometer this is equivalent to a temperature change of 4.6C. per minute; the extremes, corresponding to the counterclockwise and clockwise limits of the potentiometer rotation, are 0.8 and 2`5c'C. per minute. A 100 mfd capacitance connected between the thermocouple am plifier output and the clockwise end of.the RATE OF RISE potentio meter Introduces a certain amount of anticipation into the action of the temperature control circuit, tc reduce the amplitude of cycling of the furnace temperature. The RATE OF RISE potentio meter and the resistors lrA series with It comprise a voltage divider between the amplifier output and the temperature control circuit Input. Placing the capacitance lr shunt with the upper arm of this voltage divider causes the transmission through It to Increase with increasing rate of change of voltage. Thus the re sponse of the control circuit to steady temperature is not affected, but changes of temperature are utilized to turn off the heater some what before the control point Is reached when the temperature is rising or to turn It on In advance of the control point when the temperature Is falling. The action may be regarded as a phase ad vance Of the heater control relay action, compensating in part for the lag which exists between application of voltage to the heater and sensing of Increased temperature by the thermocouple. D. Pressure recording system The pressure recording system Is shown lr. the upper part of the block diagram. Figure 5- The pressure sensing element is a pressure transducer of the strain gauge type, operating from the power supply through a zero-adjusting circuit.. The output of the ' :i . 0692752 HARTOLDMONOOQ5287 9. pressure transducer goes to an amplifier, thence to a logarithmic compressor and to the pressure (vertical) axis of the recorder. Pressure in the cell is recorded on a logarithmic scale, with a range of 0.5 to 300 psla. The detailed circuit of the pressure recording system is shown in Figure 14 , Power is supplied to the pressure transducer by an Elcor AT25-75 power supply. The ZERO ADJUST and ZERO CHECK ADJUST potentiometers are in series across the power supply. The pressure transducer is a Statham PA401TC-300-1700 unit. With the nominal 25 volts applied to its input, the output at the full scale pressure of 300 psla is 250 mv. The ZERO ADJUST control with associated 50b kilohm resistor provides a variable shunt on two adjacent arms of the pressure transducer bridge element to adjust the Output at zero pressure. The ZERO CHECK ADJUST control provides an adjustable offset of the transducer zero when the ZERO CHECK button is preseed. This serves to bring the point of zero pressure on scale at the recorder for checking recorder calibra tion or for checking evacuation of the cell. Special provision for this is necessary because zero pressure does not appear on the logarithmic pressure scale of the recorder. The output of the pressure transducer is applied to an amplifier of gain 400. The amplified signal passes to the logarithmic compressor and to the recorder. The amplifier, Philbrlck UPA-2, and the logarithmic compressor, Kane C~?A, correspond to the descriptions of these units given in sect,ton C or the temperature system. In the am plifier, the feedback voltage divider consists of 80 kilohm and 200 ohm resistances to produce the necessary gain of 400. The logarithmic compressor is powered by batteries B4 for the bias and B5 for the output transistors. The recorder also is as de scribed for the temperature system except that the potentiometer bridge network is entirely conventional, with a linear relation ship between voltage and pen position. The bridge voltage is supplied by battery B? it senes with the VERTICAL SPAN adjust ment and with the VERTICAL ZERO adjustment for shifting of the scale position. The recorder scale covers a span of 2.8 logarith mic decades, extending from 0.5 psla to he full-scale range of the pressure transducer, 300 psla, E. Programming System The duration of each automatic vapor pressure run and the ter mination of the run if excessive temperatures or pressures occur is the function of the programming system. As shown in Figtfre 5, this system consists of the timer in series with the recorder limit, switches and the high temperature limit thermostat of the furnace. A detailed diagram of the circuit is given in Figure 15. The timer. Eagle Signal HP?7a6, contains a synchronous motor driv ing a cam through a solenoid c-iutih. The TINIER potentiometer is mounted on the front of the timer. The timer. mo*/br and the clutch coll are connected in parallel, so that whenever the motor is en ergized the clutch is engaged and the potentiometer shaft is 0692753 HARTOLDMONOOQ5288 10. rotated. When the end of the selected time cycle is reached the motor-driven cam opens a switch to terminate the cycle: the motor Is de-energized, the clutch releases, and a spring on the timer shaft returns the time Indicator and TIMER potentiometer to the starting point. The clutch solenoid also operates two sets of contacts which control the locking and releasing functions of the START and STOP buttons. Power is applied to the circuit at terminal 6 of the timer. With the Instrument in the standby condition, the normally-closed solenoid contact applies power to terminal 7 to light the STOP pilot lamp. Pressing the START button makes connection from the line through the STOP button, the two recorder limit switches and the limit thermostat In series to terminal 5 of the timer, ener gizing the motor and solenoid. Action of the solenoid closes the circuit between terminals 6 and 8 of the timer, applying power to the heater relay for heating of the furnace. At the same time, the other set of solenoid contacts in series with the cam switch bypasses the START button contacts, holding the solenoid In and causing the timer to continue to run until the cam switch opens or there is some Interruption of the limit switch circuit. The furnace circulating blower also runs during this time, as it is In effect connected In parallel with the timer motor. Termination of a heating cycle normally is caused by opening of the cam switch when the preselected time has elapsed. This action removes the line voltage from terminal 1 of the timer, de energizing the motor and solenoid through the limit switch circuit. Return of the solenoid to its inactive position removes the power from the heater relay and restores the entire circuit to the standby state. The 0.M7 mfd capacitor connected in shunt with the timer motor and solenoid (across terminals LI and 5 of the timer) suppresses the voltage surge which occurs at the solenol.d coll when it is de energized. Without this capacitor the ending of the timer program is accompanied by a large transient deflection of the recorder. The TIMER potentiometer Is mounted on the front of the timer In such a way that Its shaft rotates with the remalnlng-tlme hand of the timer, while Its body moves with the cycle-time hand. Atthe start of a cycle the two indicators coincide, and the poten tiometer is adjusted for minimum output voltage In this setting. As the timer runs the cycle-time hand Is fixed while the remalnlngtlme hand moves away from It toward zero time. The relative rota tion of the two parts causes the potentiometer shaft to rotate, relative to the body, in the direction of Increasing output vol tage. This action programs the Increase of furnace temperature as described In section 2C. At the end of the cycle, release of the timer solenoid clutch permits the remaining-time Indicator to return to its starting position in coincidence with the cycle-time indication, simultaneously returning the TIMER potentiometer to Its starting point. 069275A HARTOLDMON0005289 n. Changing of the cycle-time setting of the timer does not af fect the initial output of the TIMER potentiometer, because the two hands of the timer, and therefore the shaft and body of the potentiometer, rotate together. A change of cycle time changes the span of rotation of the potentiometer which is covered during the cycle, and hence the span of reference voltage applied to the temperature control olrcult. Maximum cycle time of the timer is 150 minutes. The lowest rate of temperature rise with which the full temperature range may be covered is about 3.5C. per minute. Lower rates of rise may be used with correspondingly shorter tem perature spans,; higher rates, up to the maximum RATE OF RISE con trol setting, may be used either by setting the timer for shorter times or by permitting the recorder temperature limit switch to terminate the program. The maximum useful rate of temperature rise is limited by thermal lag In the cell to about 15 C. per minute. F. Recorder The curves of vapor pressure as a function of temperature are recorded on a Houston Instruments HR-9? recorder, adapted for mounting in the top plate of the console. The .recording pen Is positioned by motion of two rods, which are driven by servomotors and also are coupled to balancing potentiometers. The servo am plifiers for the recorder are mounted on the back panel of the console. Measuring circuit components, other than the balancing potentiometers, are on the front panel. The circuits of the two axes of the recorder are similar. Each receives the output- of the respective logarithmic compressor. The servo systems operate by voltage comparison: motion of the pen, corresponding to rotation of the balancing potentiometer for the individual axis, varies the voltage output from the potentio meter circuit. If the difference cf Input and potentiometer vol tages Is not zero, the servomotor is driven in the direction nec essary to reduce the difference to zero. The potentiometer circuits are of conventional design. They are Wheatstone bridge networks having the balancing potentiometer and zero-adJusting potentiometer in adjacent arms. The output voltage of the potentiometer circuit appears between the taps of the two arms. As was mentioned in section 2C and shown in Figure 10, the horizontal axis circuit has a shunt on the balancing po tentiometer to produce a desired non-linearity of voltage versus pen position along the temperature scale. The servo amplifier circuit is shown in Figure 16. The ampli fiers for the two axes are identical- Terminal 1 of the input plug PI is the common terminal for both the measured and balancing voltages; it is grounded to the chassis of the amplifier, but the chassis is mounted on Insulating bushings so that isolation of the temperature and pressure measuring circuits from each other and 0692 755 HARTOLDMON0005290 12. from the physical ground (except at the thermocouple) Is preserved* The voltage being recorded appears at terminal 2 of the amplifier, and Is connected (through a filter to remove hum) to the grid of the first amplifying stage. The balancing voltage, at terminal 3, passes through a phase-advance network to the chopper which alter nately connects It to and disconnects It from the input grid. If the two voltages are equal, action of the chopper causes no alter nating component to appear at the input grid. If the voltages are not equal, the alternate opening and closing of the chopper con tact produces an alternating voltage which is amplified through the 4 amplifier stages and is coupled to the servomotor control winding at terminals 2 and 4 of the output plug P3. Inductance coupling Is used to bypass the d-c plate current of the output am plifier tube around the servomotor. The power winding of the servomotor is connected to the a-c line. Direction of rotation of the motor is determined by the phase of the control voltage rel ative to the line voltage, and this In turn Is determined by the sign of the difference between the measured and balancing voltages at the amplifier Input. The LOAD - OPERATE switch substitutes line voltage for the am plifier output voltage at the servomotor control winding, to cause the pen to move to the upper right comer of the recorder platen and thus make easier the removal or Insertion of chart paper. An attenuator between the second and third amplifier stages adjusts the overall gain of the amplifier so that accurate positionlng of the pen Is obtained without oscillation. The phase ad vance network in series with the potentiometer voltage at the am plifier input also assists in this by partially compensating the overall phase lag of the recorder servosystem. The paper for recording has a grid matching the l/K. and log pressure calibrations of the axes. Overall size of the paper is 8.5" x 11". Guides at the top and right edges of the platen con trol the position of the paper, which is held down also by clips at the left edge. Because the printed grid lines may not always be in exact relation to the edges of the paper, register marks are Included on the paper for alignment with the guides. The ruling area of the chart is 7-0" wide by 9.26" high, located with the up per right comer 0.62" from the comer of the sheet in each dir ection. The register marks are 0.50" from the top and right boun daries of the ruling area, approximately aligned with the top left and right, and right top and bottom corners of the chart rulings. Table I gives the positions of the grid rulings for the tempera ture and pressure axes. '^ ' ;) 0692756 HARTOLDMONOOQ5291 TABLE I CHART RULINGS 13. A. Temperature (horizontal axis) Distances measured from left edge of ruled area. Underlined temperatures are those marked on the axis. Designation of axis, TEMPERATURE - C. The numerals under the heading Ruling show the relative line weights! 1 - light; 2 - medium;T~:r"neavy. Temperature Ruling Position C. Inches 20 SF 30 35 40 45 50 5? 60 65 70 75 80 85 90 95 100 ITS 120 130 140 150 160 170 3 0.00 1 0.19 2 0.37 1 0.55 2 0.72 1 0.89 3 1.05 1 1.20 2 1.35 1 1.50 2 1.64 1 1.78 2 1.92 1 2.05 2 2.17 1 2.29 3 2.42 1 2.65 1 2.87 1 3.07 1 3.27 3 3.46 1 3.64 1 3.81 Temperature Ruling Position C. Inches 180 1 190 1 200 3 S-TC 1 220 1 230 1 240 1 250 2 260 1 270 1 280 1 290 1 300 3 5577 1 340 1 360 1 380 1 400 3 WS6 1 440 1 460 1 480 1 500 3 RegTst'er mark 1 5.98 4.14 4.29 4.43 4.57 4.70 4.83 4.96 5.07 5.19 5.30 5.40 5.51 5.70 5.88 6.05 6.21 6.36 6.50 6.64 6.76 6.88 7.00 7.50 0692757 HARTOLDMONOOQ5292 TABLE I (continued) CHART RULINGS B. Pressure (vertical) axis Distances measured from bottom edge of ruled area. Designa tion of axis, PRESSURE - PSIA. Underlined pressures are those marked on the axis. The numerals under the heading Ruling show the relative line weights: 1 - light; 2 - medium; 3-~'Heavy. Pressure Ruling Position Psia Inches Pressure Ruling Position Psia Inches 076 0.7 0.8 0.9 1.0 rr? 1.4 1.6 1.8 2.0 07? 3.0 3.5 4 .0 4.5 5.0 F" 7 8 9 10 T5 14 1 1 1 1 3 i 1 1 1 3 1 2 1 2 1 3 1 1 1 1 3 1 1 0.00 0.26 0.49 0.68 O.85 1.00 1.26 1.49 1.68 1.85 2.01 2.33 22..589? 3.01 3.18 3.33 3.60 3.8? 4.01 4.18 4.33 4.59 4.8? 16 18 20 0? 30 35 40 45 70 80 90 100 170 140 160 180 200 070 240 ?60 ?80 300 Register mark 1 1 3 1 2 1 2 1 3 1 1 1 1 3 1 1 1 1 3 1 1 1 1 3 1 5.01 5.18 5.34 5.67 5.92 6.15 6.34 6.51 6.67 6.93 7.16 7.45 7.52 7.67 7.93 8.16 8.35 8,52 8.68 8.81 8.93 9.05 9-16 9.26 9.76 Figure 17 shows the recorder chart. The original drawing for the chart is Included with the instruction manual accompanying the instrument. The recording pen is attached to a carriage resting on the Intersecting pen drive bars of the recorder. A pivoted bracket for the pen tip permits the tip to be removed from the paper when no curve is being traced. The pen tip is a "Wrico" size 7 let tering pen point. 0692758 HARTOLDMONOOQ5293 Section 5. Operating Procedures 15. A. Detailed Procedure Preparation of Instrument for Operation 1. Turn on Instrument - pilot light next to "stop" switch Indicates that the Instrument power Is on. Allow at least onehalf hour, preferrably two hours, for Instrument warm-up. 2. With selector knob In "load" position, put recorder chart In place. Return the knob to "operate" position. 5. Degas the cell by heating to 4000. under high vacuum (<0.05 psla). The Instrument temperature program may be used at the maximum heating rate setting. Procedure for Initiating the temperature program Is given, below in section C. Remove cover and allow cell to cool with both Inlet and outlet valves closed. Measurement, 1. Refer to Figures 2 and 8 for the locations of controls. Switch vacuum to sample funnel with stopcock A open and the by pass valve closed. Allow several minutes for the lines to evacu ate and then close stopcock A. 2. Fill upper chamber to mark with sample and open stopcock B to allow the sample to drip into the lower chamber. Close B when the upper chamber is emptied and continue degassing until bubbles no longer rise to surface. Relieve vacuum to the funnel and open stopcock A. 5. Adjust level of sample by opening the bypass valve until meniscus Is even with one of the volume scale divisions. *1. Open Inlet valve and admit 12 ml of sample Into the cell. Close Inlet valve. Temperature Programming of Sample 1. Make heating rate settings with instrument in "stop" posi tion. Indicator light next to stop button will be on. Timer set ting should be such to provide the desired temperature Interval at the chosen heating rate. For a heating rate of 5/mln and a ter minal temperature of 400C. (that Is, 575 rise above room temper ature), the timer setting would be 75 min, 2. Push "start" button to begin the temperature program. Lower recorder pen to recording position. 5. When run Is terminated, remove cell cover and lift recor der pen. 0692759 HARTOLDMONOOQ5294 Cleaning of Cell Following Completion of Measurement 1. When the sample has cooled to 200"C. or lower, discharge the sample by opening the outlet valve. Apply vacuum to the trap and open the Inlet valve and also stopcock A to allow sample re maining In filling funnel to partially flush the cell. Evacuate the drained cell and close valves. 2. Put acetone In the sample funnel and open inlet valve to allow cell to fill with acetone. > 3. Change to plant vacuum (or other aspirator type vacuum source), open outlet valve and flush cell with acetone finally allowing air to flow through the cell to remove excess acetone. 4. Change back to pump vacuum and evacuate cell. Close valves. 5. Repeat steps 2 - 4 as necessary to clean cell. B. Discussion Preparation of Instrument for Operation After sufficient warm up the instrument should Indicate close to room temperature and atmospheric pressure on the recorder. The pressure scale Is corrected for the liquid sample head and when the cell Is empty, the atmospheric pressure reading will be 14.4 psl. Under complete vacuum the pressure reading will be 0.5 psia when the ZERO CHECK button Is depressed. These readings will in dicate whether the Instrument calibration has shifted appreciably and should always be made Just prior to a run. Failure to degas the Inner surface of the cell before charging sample w.111 result in degassing during the run which will Increase the pressure In excess of the vapor pressure increase with temper ature. The resultant pressure vs. temperature record will have an appearance similar to that from samples containing low-boilers or decomposable Impurities or both. The chlorinated biphenyls when heated In excess of 400e'C. produce a thin layer of carbonaceous matter on the cell surface which is only partially removed during the normal clean out. This carbonaceous material and the metal surface of the cell both could strongly adsorb low-boiling mater ials or air which would require stringent conditions for their removal. It appears that a temperature of 250"C. Is sufficient to degas the cell, but to be on the safe side, heating to 400C. under full vacuum is recommended as a standard procedure. A good medium capacity vacuum pump should be used so that system pressures at least as low as 3 mm Hg can be achieved be fore the sample Is Introduced Into the cell. 0692760 HARTOLDMONOOQ5295 1? Occasionally the system should be tested for leakage. This may be done by measuring the pressure increase with time of the evacuated system following the degassing procedure. Leakage to the extent of 0.1 psl in one-half hour is permissible. Freparat1on of Sample Samples of chlorinated biphenyls usually contain dissolved gas which, if not removed, will add to the system pressure during a run. The effect will be the same as a poorly evacuated sample cell prior to the run. Evacuating the sample to several mm Hg at room temperature adequately degases Thermlnol PR-l and FR-2 fluids. The degassing of more viscous fluids such as Thermlnol FR-5 Is facilitated by warming the sample to 80-100oC. Heating of samples suspected of containing low-boilers should be avoided . if the vapor pressure instrument is being used to detect the presence of low-boilers. Removal of the low-boilers by some heat ing during degassing however improves the ability to detect low levels of decomposable impurities such as the chlorine addition products with chlorinated biphenyls which decompose in the tem perature range of 150-500"C. Temperature Programming of Sample Heating rates of 5* to 10C./min may be used with little ef fect on the temperature-pressure record. Slower heating rates cause uneven programming of temperature and at rates' higher than 10C./min, temperature lag becomes significant. At heating rates less than 10C./min, the instrument records quite close to the equilibrium vapor pressure of the sample vs. temperature. Some difficulty has been experienced with slight changes or cycling in the Instrument calibration during a run. Usually this is minor and does not appreciably change the P vs. T trace. Any shifts can usually be spotted and normally would not be confused with true pressure increases with temperature. These calibration shifts or cycles are due to Inadequate thermostatlng of the log arithmic compressor units. The present thermostat has shifted about 1 degree from its original set print but currently appears to be maintaining fairly constant temperature with only slight cycling. If this problem worsens the thermostat should be re placed with a more reliable one. A terminal temperature of 400*0. is recommended for runs with chlorinated biphenyl since the data beyond this temperature are not useful for screening purposes and the avoidance of the high temperature decomposition facilitates the cleaning of the cell. Cleaning of Cell Upon heating chlorinated biphenyl samples above 400C., a car bonaceous deposit is formed on the cell surface. This thin film is loosened and partially flakes off during clear.-out using fresh Aroclor, benzene, acetone, etc. as solvent. Acetone appears to 0692761 HARTOLDMONOOQ5296 18. be the most satisfactory solvent, having fair solubility for chlorinated biphenyls and Is easily removed from the system fol lowing the clean-out procedure. Successive evacuations, fillings, and dischargings of solvent are necessary to clean the strain gage and relief valve cavities. Three to four washings are adequate to clean the system. C. Vapor Pressure of Thermlnol Fluids Typical Instrument records of the vapor pressure of the Thermlnols follow in Figures 17-22. Included are curves which show the effect of failure to degas the cell prior to a run, the effect of unstable material, the effect of low-boiler contamination and the relative vapor pres sures of the different Thermlnol fluids. Section 4. Maintenance A. Introduction Maintenance procedures required for the Vapor Pressure and Thermal Stability Apparatus consist of: 1. Routine replacements of electronic tubes and batteries, and lubrication of mechanical parts; 2. Miscellaneous repairs or replacements not required on a routine basis; 3. Checking and adjusting instrument calibration, and adjust ing operating controls to provide accurate recording of vapor pressure curves; and 4. Locating sources of trouble when an instrument failure occurs. These four types of maintenance procedures are described In the sections following. B. Routine Maintenance Routine, or preventive maintenance procedures for the Instru ment are quite simple. They comprise lubrication of moving parts In the recorder, and periodic replacement of tubes and batteries In the electronic circuits. The recorder should be lubricated with a light instrumentgrade oil, at 6-month Intervals. Points to be lubricated are on the underside of the recorder chassis (Figure 22). One drop of oil should be applied at each point: Motor intermediate gear shaft (2) Motor output gear shaft (2) Pen drive cable pulleys (8) 0692762 HARTOLDMONOOQ5297 19. The timer presumably requires no lubrication, as the manufac turer of the unit supplies no instructions for this. Similarly, the furnace blower motor is sealed and Is not Intended to require routine lubrication. All batteries should be replaced every 6 months, or every 1500 operating hours if this occurs sooner. The set of batteries includes 5-RM42R and 12-RM3R mercury cells. Positions of the batteries are shown in Figure 23, the batteries being held in clips on a sub-panel on the back of the console panel. The RM42R batteries are mounted individually; the RM3R batteries are held in groups of 6 in fiber sleeves. Recalibration as described in section 4D Is required after replacement of batteries. The instru ment should be run for 15 hours or longer after battery replace ment, to stabilize the new batteries, before it is calibrated. If a tube tester is available, the tubes should be tested at 6 month or 1500 hour Intervals and any defective or doubtful tubes replaced. If a tester is not available, all tubes should be replaced at 12 month or 3000 hour intervals. Replacement of the 4 type 12AX7/ECC83 tubes in the Philbrlck amplifiers with ei ther this double-branded type or with type 5751 Is recommended, instead of replacement with standard type 12AX7 tubes which may be excessively noisy for this application. Calibration of the temperature or pressure scales of the instrument should be checked ) after replacement of any tubes in the corresponding amplifiers, and adjustment of the TEMPERATURE SET control should be checked after replacement of any tubes in the temperature control circuit. Normally, however, tube replacements are expected to have negli gible effects on calibration. Tube replacements In the recorder amplifiers may make it necessary to adjust the recorder gain con trols for best recorder response. C. Repairs and replacements This section Includes Instructions for removal, repair or adjustment, and installation of various components of the insturment which are not expected to require attention routinely. 1. Recorder If any extensive work Is required on the recorder mech anism, it is convenient of obtain access to this part of the in strument by unsoldering the leads to the recorder balancing poten tiometers (see Figure 22), removing the four- screws at the cor ners of the top panel, and tilting the panel backward. Mounting of the motors and potentiometers is obvious. Replacement or ad justment of the pen drive cables is done as described below: Each axis requires 11 feet of drive cable, which is General Cement type 88-100. Figure 24 shows the arrangement of the drive cord for the vertical drive, with the vertical pen drive bar in its uppermost position. The same illustration applies to the 0692763 HARTOLDMONOOQ5298 20. horizontal drive when the horizontal pen drive bar Is In its ex treme left position, the drawing being turned 90 counterclock wise. Tie the drive cord to screw 1 and make 4 turns In a clock wise direction. Loop the cord counterclockwise around pulley 2; thence under motor pulley 3, which should be turned fully clock wise. Make 2 turns around 3, on the second turn looping through the slot and around the lug on the side of the pulley. Continue to potentiometer pulley , which should be about one-fourth turn clockwise. Make 3 turns clockwise around 4_, then complete a fourth turn with a loop through the hole In the pulley and out to the stud on the side. Bring the cord on around 4_, then clockwise around pulley to stud 6. Make 4 turns clockwise around 6, pass the cord through the eye of the stud and then make 4 more turns counterclock wise. Loop the cord around pulley 7, then draw it across to pulley 8 so that it passes under the cord already in place. Loop around pulley 8 and thence back to the starting point at lug 1^, at wjilch point make 4 turns in a clockwise direction and tie the cord. To adjust the cable tension, loosen the locknut on screw 6 and turn this screw until the cable near the center of the platen de flects l/4 or 3/8 Inch on slight pressure. To adjust the pen drive bar so that the pen moves parallel to the chart lines, rotate screw 1_. Proper operation of the recorder potentiometer circuits re quires that the balancing potentiometers have the correct relation to the scale readings on the two axes. To check or adjust this condition, disconnect the recorder amplifier input plugs from the amplifiers and remove batteries b6 and B? from their clips; then connect an ohmmeter to the CCW and S terminals of the horizontal balancing potentiometer (at the left side, when the recorder is viewed from the front). (The resistance measured should be 450 ohms when the pen' Is at the left edge (20cC. mark) of the chart. If the resistance is'hot correct, adjust it by loosening the cable pulley set screw, slipping the pulley from the potentiometer shaft, and turning the shaft to give the correct resistance. For the ver tical recorder axis, similarly, the resistance between CW and S terminals of the, potentiometer should be 1100 ohms when the pen is at the bottom edge (0,5 psla mark) of the chart. The recorder amplifiers are isolated from the cabinet by being mounted with insulating washers under the screw heads, and with rubber covers on the chassis edges. Care should be taken not to lose these parts if the amplifiers are removed for any reason. Mounting arrangement of the recorder amplifiers is shown in Figure '25 2. Timer The timer and attached timer potentiometer are shown in Figure 26. Removal or adjustment of the timer potentiometer can be done without disturbing the timer. To remove the setting knob, loosen the clamp screw (at left center in Figure 26) and slide the 0692764 HARTOLDMONOOQ5299 ?i knob off of the potentiometer body. To remove the potentiometer, loosen the set screw of the potentiometer bushing and slip the potentiometer assembly off the timer. The potentiometer is cou pled to the timer by a tongue protruding into a slot in the screw which holds the timer pointer to its shaft. When replacing the potentiometer on the timer, line up the tongue and groove approxi mately and then work the two parts into contact by simultaneous rotation and forward motion. Pull the potentiometer backward slightly from the position of greatest insertion before tightening the set screw, to reduce friction at the potentiometer and timer shafts. For the method of adjusting the timer potentiometer elec trically after it has been moved, see section Ad. The friction bushing in the face of the timer, on which the potentiometer is mounted, should be tightened enough to avoid chance movement but not so much as to prevent setting of the timer. Access to the mechanism of the timer requires removing it from the panel. To do so, remove the potentiometer and then remove the four screws at the corners of the timer. Disconnect the wiring at the terminals on the rear of the timer, and slip the timer out for ward through the panel hole. Removal of two screws through the back of the flange then permits the dial cover to come off. Re move the pointer screw, and the pointer and dial; remove four screws which were covered by the dial, and the flange will sepa rate from the timer body to permit the cylindrical body cover to be removed. To replace the dial cover-, turn the time-setting pointer to a position in which it is aligned with the remainingtime indicating pointer, and manipulate the dial cover until the pointers engage and the cover is seated on the timer flange. Removal of the timer motor does not require removal of the timer from the panel. It is only necessary to disconnect the motor leads and to remove the two screws which hold the motor to the timer body. 3. Furnace The cell, heater, thermocouple and blower motor are moun ted on a split block under the center of the furnace base. The cell inlet and outlet tubes are clamped between the halves of the block; heater and thermocouple are held in holes through it by socket set screws; and the blower motor Is supported below it on a U-shaped bracket. The limit thermostat is supported from the furnace base shell. The pressure transducer is supported by an angle bracket, and the relief valve is held in place by Its outlet tube. Figure 7 shows the arrangement of components beneath the furnace base. For removal of the heater or thermocouple, an Allen wrench long enough to extend through the slot at the back of the furnace base is useful. Removal of the connecting wires and terminal nuts from the heater permits it to be withdrawn vertically when the set screws are loosened. The thermocouple Is removable downward, after partial straightening of the bends in the furnace portion. If a 0692765 HARTOLDMON0005300 22. defective thermocouple Is to be replaced. It can of course merely be cut off at a convenient point. The blower motor may be removed by loosening the setscrew of the blower coupling, and removing the four screws which hold the motor mount to the split block. The cell holding block is separable when the motor is removed, by loosening the two screws which hold the block together and the four screws which hold it to the furnace base plate. Removal of the intact cell was not contemplated in the design; it is neces sary to cut the cell inlet and outlet tubes, conveniently Just above the furnace shell base. To Install a new cell, cut the in let and outlet tubes to the required length as determined by try ing the cell in place. Then the required right-angle bends below the furnace base can be made and the fittings Installed before the cell holding block is put in place. The pressure transducer is removable by disconnecting the tubes attached to it, removing the electrical connector and slid ing the transducer out of the support bracket. D. Adjustment and calibration " The temperature recording, temperature control and pressure recording sections of the instrument require adjustments to main tain satisfactory operation and accurate calibration. Location of the controls is shown in Figure 28. 1. Calibration of temperature recording system To calibrate the temperature recording system it is neces sary to have a potentiometer capable of supplying voltages which simulate the output of an lron-constantan thermocouple at temper atures from 20'' to 500"C. Disconnect the leads from the furnace thermocouple at the terminal block on the back of the console, Figure 26, and connect the potentiometer to the terminal block with thermocouple wire. Measure the temperature at the potentio meter terminals, and set the reference-Junction compensation of the potentiometer to the corresponding voltage. (This step is of course to be omitted if the potentiometer has automatic referencejunction compensation.) Calibration of the temperature axis requires the setting of three adjustments: the THERMOCOUPLE OFFSET, HORIZONTAL ZERO and HORIZONTAL SPAN controls. The temperature scale is adjusted to register at three temperatures: 5 200 and 400C. First set the potentiometer to 400C. (21,85 mv), and adjust the HORIZONTAL SPAN control until the recorder reads this temperature. Change the potentiometer setting to 200C. (10.78 mv), and adjust the HORIZONTAL ZERO control for this temperature. Repeat these ad justments in sequence until the scale is aligned at the two tem peratures. Then set the potentiometer for 50C., and adjust the 0692766 THERMOCOUPLE OFFSET control to produce agreement at this temper ature. Finally repeat all three adjustments In sequence until the scale Is In agreement at all three temperatures. If the potentiometer is of the type which has both a step switch and a continuous slidewire, it Is convenient to use the three voltages of 2.58, 12.58 and 22.58 mv. These correspond to 50, 253 and 4l3C., and the adjustments are made for these tem peratures. Switching between the three points requires only the changing of the step dial of the potentiometer. 2. Adjustment of the temperature control system The temperature control system must be adjusted to pro duce a temperature program of the correct rate of rise, beginning at room temperature. Mechanical alignment of the TIMER potentio meter and the RATE OF RISE dial, and electrical adjustment of the TEMPERATURE SET control are required. A volt-ohmmeter of the usual type Is needed for making the adjustments. Connect the meter, set for 100 volts range, to the top two terminals on the left side of the temperature control component board (lower left In Figure 23), the uppermost terminal being negative. With the cell at room temperature and the timer not running (it may be set for any time Interval), loosen the TIMER potentiometer bushing setscrew and rotate the potentiometer. At one position the voltage will Jump abruptly from near zero to about 100 volts. Turn the potentiometer counterclockwise a small amount from this position, so that a voltage of a few volts appears. Then turn the potentiometer clockwise until the vol tage Just stops decreasing (it will remain constant at a value which may be a few tenths of a volt either positive or negative). Tighten the setscrew In this position. Note the precaution in section 4C of allowing a small axial clearance between the tongue of the potentiometer shaft and the slot of the timer pointer screw. Note the action of the heater relay while adjusting the TEM PERATURE SET control, with the RATE OF RISE control at approxi mately mid-range. Adjust the TEMPERATURE SET control until the relay plunger is Just pulled down. Because of the delay of re sponse of this circuit, it will be necessary to make the adjust ment slowly, during an interval of about a minute. Turn off the power to the instrument and connect the meter, as an ohmmeter on a 1-megohm range, between the center tap of the RATE OF RISE control and the two end terminals alternately. Ro tate the knob of this control until the resistance readings for the two halves of the control are the same. Then loosen the knob, and reset it to the position representing 4.6 per minute rise. 3. Calibration of pressure recording system The principle of calibration of the pressure recording system is the same as that for the temperature system: adjustmer 0692767 HARTOLDMON0005302 24. so that the scale readings are correct at three points on the scale. The method of operation, however. Is different because Instead of applying standard voltages In place of the thermocouple output, standard pressures now must be applied to the pressure transducer. A source of gas~pressure, such as a cylinder of ni trogen with a high-pressure regulator, and pressure gauges cover ing the range of 0.5 to 300 psia are required. A combination of gauges, such as a test gauge of 300 psia range with a manometer reading absolute pressure up to 1 atmosphere, is sufficient. These gauges should be connected through shut-off valves to a rudimentary manifold system attached to the cylinder pressure regulator (through a valve) and to the sample Inlet fitting of the furnace unit. The vacuum pump normally used with the unit remains con nected to the outlet fitting. Thus any pressure In the range of 0-300 psia may be applied to the pressure transducer in the fur nace unit, and simultaneously measured by the reference gauges. Before the pressure calibration is begun, the voltage applied to the pressure transducer should be checked. Connect a voltmeter of 25 volt range to the counterclockwise terminals of the ZERO ADJUST and ZERO CHECK ADJUST controls (the ZERO CHECK ADJUST con trol terminal being positive). If the voltage is not 25+0.5 volts, adjust to this value with the pressure transducer voltage adjustment on the power supply (Figure 23, upper right). The calibration adjustments are made at 1.0 psia, atmospheric pressure, and 200 psia. However, the calibration Is done with the cell empty, while In use the cell is half-filled with liquid and the pressure transducer therefore supports not only the gas pressure but in addition the pressure of the column of liquid be tween the transducer and the cell. The liquid column adds appro ximately 0.3 psl to the gas pressure in the cell, and the calibra tion pressures must therefore be Increased by this amount -- that Is, an actual pressure of 1.3 psia Is used to calibrate at the 1.0 psia mark of the scale, and so on. This correction may be neglected at the higher pressures, 50 psia and above. Apply 200 psia to the transducer and set the VERTICAL SPAN control for correct reading at this pressure. Release the pres sure to atmospheric, and adjust the VERTICAL ZERO control until the scale reading is 0.3 psl less than atmospheric pressure. Re peat these two adjustments In sequence until both are correct. Then evacuate the cell to 1.3 psia (67.2 mm Hg) and adjust the ZERO ADJUST control for correct indication of 1.0 psia here. Finally repeat the entire sequence of adjustments to secure adequate alignment of the scale at all three pressures. As the last step of adjusting the pressure recording system, the operation of the ZERO CHECK button must be set. Evacuate the cell, hold the ZERO CHECK button depressed and rotate the ZERO CHECK ADJUST control to give a reading of 0.5 psia on the chart. 0692768 HARTOLDMON0005303 Jj. Adjustment of recorder amplifier gain controls The gain controls of the recorder amplifiers must be ad justed for best response as tubes age or after a change of tubes. Turn the gain controls (Figure 25) fully clockwise. If oscilla tion of the recorder pen occurs, turn the controls counterclock wise until the oscillation stops. Test the recorder response by moving the pen bars manually a short distance and releasing them: they should return to the balance position with a slight "bounce. ' If the action appears sluggish. Increase the gain control setting, if there is a tendency to oscillation, decrease the setting. The combination of narrower voltage span of the temperature axis of the recorder, and higher hum level in the temperature amplifier output, normally makes this axis of the recorder stable at full gain setting. If no oscillation occurs when the horizon tal recorder amplifier gain is at its maximum setting, the con trol may be left In this position. 5. Adjustment of logarithmic compressor thermostat The temperature of the box housing the logarithmic com pressor units for both temperature and pressure axes is controlled by a thermostat mounted on the side of the inner compartment. Access to the thermostat is obtained by removing the outer box cover, which is the left side of the box as seen in the photo graph (Figure 23, center). The thermostat adjusting screw is sealed with "Glyptal" cement. To loosen the cement, keep it wet ted with acetone for a few minutes. It is not anticipated that adjustment will be required, but if the. temperature in the inner box (measured by a thermocouple introduced in the interstices of the wire bundle entering the box) is not within the range of 36"1i0oC. the thermostat should be readjusted to provide a tem perature in this range. Change of temperature of the logarithmic compressors produces a change of calibration, which can be cor rected by adjusting the HORIZONTAL ZERO and VERTICAL ZERO ron'rcl? only, without change of the span, offset and transducer zero. E. Trouble-shooting Trouble-shooting of the vapor pressure apparatus depends upon first locating the section of the instrument In which a failure has occurred, and then locating and correcting the defect. Symp toms of trouble may be Incorrect temperature or pressure readings at atmospheric pressure and room temperature! incorrect progres sion of temperature when programming,: irregular (noisy) motion cf the recorder pen; or failure of the pen to move along one or bo'r, axes. It is helpful to remember that the temperature arid pressure systems of the vapor pressure apparatus are Independent except for receiving power from a common source, and having the 0692769 HARTOLDMON0005304 26. logarithmic compressors of the two systems in the same thermo stat. Occurrence of the same malfunction in both systems therefore indicates a failure at one of the two common points. Con versely, a malfunction in one system alone eliminates the com mon units from consideration as points of failure. The first step in trouble-shooting, therefore, is to follow Chart 1 below. The appropriate chart of the other four may then be applied to location of the defect. Trouble shooting chart Is preliminary Indication The same defect is shown on both temperature and pressure axes of the recorder Proceed to chart II Defective operation is shown only on the temperature axis III Defective operation is shown only on the pressure axis IV No defect of recording either temperature or pressure is shown, but the furnace temperature is not being programmed correctly V } 0692770 HARTOLDMDNnnnfi^nci Trouble-shooting chart IIx Failures affecting both recorder axes Indication Recorder dead: pen may be positioned manually at any point on chart Recorder operating, but not giving correct readings a) Pen stays at lower left corner of chart b) Low readings on both axes c) Pen stays at upper right corner of chart d) High readings on both axes Noise in both temperature and pressure readings Probable cause of trouble No a-c power to recorder amplifiers Verifying test Check for lighted tubes No a-c power to temperature and pressure amplifiers Check for lighted tubes; check output voltage of Sola transformer High temperature in log compressor thermostat Check thermostat temp. Defect of LOAD-OPERATE swltchjknob not properly set on shaft Check switch and try knob for correct position Low temperature in log compressor thermostat Check thermostat temp. Excessive a-c line noise; loose connection In line cord or internally In power circuit Check line voltage; _ check wiring ' ro 0692771 HARTOLDMON0005306 Trouble-shooting chart III? Failures affecting the recorder temperature axis only Indication Probable cause of trouble Verifying test Dead: pen can be moved manually to any temperature reading Noise In temperature indicatlon Low reading at room temperature Correct reading at room temperature, low reading at yhigh temperature High reading at room temperature Inoperative horizontal-axis recorder amplifier (bad tubes; shorted capacitor across output; loose plugs) Defective tubes in temperature amplifier Defective battery B1 Defective battery B5 Defect In temperature amplifier Defective thermocouple Defective battery B? Defective battery B6 Defect In temperature amplifier Interchange with vertical amplifier by interchanging plugs Interchange tubes with pressure amplifier Measure voltage: 1.55 volts normal Measure voltage: 8.1 volts normal Measure output voltage (between blue and black jacks, blue+): 8.2 volts normal at room temperature Measure thermocouple output at terminals on back panel, compare with auxiliary', oouple inserted In furnace Measure voltage: 1.55 volts normal Same Measure output voltage as above: 8.2 volts normal at room temp. w CD 0692 772 HARTOLDMON0005307 Trouble-shooting chart IV: Failures affecting the recorder pressure axis only Indication Probable cause of trouble Verifying test Dead: pen can be moved manually to any pres sure reading Noise in pressure indi cation Pen stays below lower chart margin at all pressures: or reads low at all pressures Low reading at atmospheric pressure Inoperative recorder verticalaxis amplifier (bad tubes; shorted capacitor across output; loose plugs) Defective tubes in pressure ampllfier No, or low, output voltage from transducer power supply Defective battery B5 Defect In pressure amplifier High reading at atmospheric pressure Defective battery B4 Defective battery B7 Defect in pressure amplifier Interchange with hori zontal amplifier by interchanging plugs Interchange tubes with temperature amplifier Measure power supply voltage: 25 volts normal Measure voltage: 8-1 volts normal Measure output voltage (between blue and' black Jacks, b3ue+): 4.6 volts at atmpressure, -0.J volt evacuated Measure voltage:. 1.J5 volts normal Same Measure output voltage as above 0692773 HARTOLDMON0005308 Trouble-shooting chart Vi Failures affecting the temperature program only Indication Probable cause of trouble Verifying test Program does not start Heater does not go on Heater Btays on continuously Rate of temperature rise not correct Program does not stop automatically Recorder switch on LOAD position Defective recorder limit switch, or open circuit in limit control system* defec tive STOP switch Defective timer contact Timer motor not running Defective 2D21 tube Defective 12AX7 tube Incorrect power supply voltage Timer motor not running Timer contact not opening at end of cycle Check switch position Check for presence of line voltage between timer terminals LI and 5 when START switch is pressed Check for line voltage between LI and 1 when START button is pressed Observe motor Change tube Change tube Measure voltage across each 0B2 tube: 105 volts normal Observe motor Check for absence of line voltage between LI and 1 of timer when end of cycle is reached 0692774 HARTOLDMON0005309 31. F. Parts List 1. Sample system and furnace a. Valves b. Relief valve c. Pressure transducer d. Heater e. Blower motor f. Limit thermostat g. Thermocouple Whltey 0VS2-A (2) Whitey 0VS2 (l) Circle Seal 5120T-2MP-600 Statham PA401TC-300-1700 Electro-Therm 10-1452-26 Bodlne B8194E Fenwal 16050-0 ThermoElectrlc 5A2120F, 12" immersion, 60" leads 2. Recorder console a. Transformers Sola 20-13.>60 Triad R-4A (2) Triad R-73B b. Amplifiers Phllbrlck UPA-2 (2) c. Log compressors Kane C-7A (2) d. Power supply e. Relays f. Choppers Elcor AT25-75 Ebert EM-1 Potter & Brumfield BS23D Oak 660 (2) (equivalent: Alrpax 172) James Electronics C1800 (2) g. Potentiometer (timer) Hellpot G, 10k, std. tol. h. Potentiometers (recorder) Helipot C, 5k, std. tol. (2) I. Limit switches Aero 1MD1-1A-A18M (2) J. Recorder chassis k. Recorder motors Houston HR92-1 Barber-Colman EYAZ4456 (2) l. Temperature-compensating Texas Instruments TM-l/4 resistor Sensitor, 500 ohm m. Log compressor thermostat Fenwal 32400-0 m. Recorder pen Wrlco No. 7 n. Tubes Standard makes 5-12AX7 4-12AX7/ECC83 or 5751 2-6x4 2-6U8 2-6S4 2-0B2 1-2D21 p. Batteries Mallory or equivalent 5-RM42R 12-RM3R q. Fusefe Buss AGC-10 amp. (2) Buss AGX-0.25 amp. 0692775 HARTOLDMONOOQ5310 U692776 HARTOLDMONOOQ5311 Furnace 069? ? 7 7 HARTOLDMONOOQ5312 06V? 78 HARTOLDMONOOQ5313 Figure Jt. Panel (front) 0692779 HARTOLDMONOOQ5314 0/Sj,/ Figure 5. Block Dlagran 0692780 HARTOLDMONOOQ5315 K3[:U3''C '- Furnace {' - -1 removed) 06977til HARTOLDMONOOQ5316 0697787 HARTOLDMONOOQ5317 Figure 8. Sample System 0692783 HARTOLDMONOOQ5318 . tups wtioco our^toc TOBCS W l PI O fj OP. * 0.0JS WALL , Jo4 ___ j. s tv**- Dto rise tvs- 0 0 4. . , tt o. OJJtyALl, Jo - ._____ S.J. WLOCO T-U0/V0 \ OlVlNS'OhJS >*/ INCHCS SCALl Tm J 0.0. X o OSJ t-SALl t J/4 SJ, TvgfVO . LfMtr. AO &CQO/0P sue Figure 9- Sample Cell 0692764 HARTOLDMONOOQ5319 AMPL Pig. .10. Elementary Circuit Diagram (Temperature Recording and Control Section) 0692785 HARTOLDMONOOQ5320 Figure 11 Phllbrlck Amplifier Component Boards 0692786 HARTOLDMONOOQ5321 TempPress **tS SoK Figure 12. Philbrlck Amplifier Chassis Wiring 0692787 HARTOLDMONOOQ5322 Fig. 13a. Logarithmic Compressor Schematic Diagram Fig. 13b. Logarithmic Compressor Circuit Board 0692786 HARTOLDMONOOQ5323 Figure 14. Elementary Circuit Diagram (Pressure Recording Section) 0692789 HARTOLDMONOOQ5324 Figure 15. Elementary Circuit Diagram (Programming Section) 0692790 HARTOLDMONOOQ5325 o Figure 16. Recorder Amplifiers V4 124X7 vj 124X7 vt 6406 VI 6X4 0692791 HARTOLDMONOOQ5326 VIM B vnttJM t*C BO TIHNMTIMt *C Figure 17. Recorder Chart HARTOLDMONOOQ5327 PRESSURE HARTOLDMONOOQ5328 PRESSURE . P$IA HARTOLDMONOOQ5329 PSIA PRESSURE 400 500 0692795 HARTOLDMON0005330 PSIA PRESSURE HARTOLDMONOOQ5331 0692 79 7 HARTOLDMONOOQ5332 0692796 HARTOLDMONOOQ5333 0692799 HARTOLDMONOOQ5334 069?600 HARTOLDMONOOQ5335 HARTOLDMONOOQ5336 K.1 inn'O . ;J'ib3c connect loriP oomboz HARTOLDMONOOQ5337 Flpurc- ?8. Calibration aajustmentr- 0692803 HARTOLDMONOOQ5338 \ A* * *3 (v-.-''y)*" n A^ A _ __ rq, Ag 0 <M 55 \ sw ' -V -t. .. j THCRmOCO^PlC 9 THCffMO~ CLCC Ttp/r SA2iOF- 12"- o" Figure ?9. Furnace ftir-ng 0692804 HARTOLDMONOOQ5339 0692805 HARTOLDMON0005340 0693806 HARTOLDMONOOQ5341