Document evrmMMYxdEqVBGGYV4eNwG8oM

manufacturing linkages within prec* by using an extremely email pieu tical tolerances. change; that is, the smaller the Another fault in the on-off design is . roent of change, the nearer the reipo " evident in the table, A change in nozzle- will approach sirsight-lin$ performtpe baffle clearance of 0.0005 in. causes an* It appears, however, that We are no output change of 3 psi at one point dealing with extremely small changes1' while et another a change of 0.0085 in. nozzle-baffle clearance because total is required to produce an equal change change represented on the graph is onijr in output. This represents a 17 to 1 0.002 in. Naturally, such minute move ratio in the rate of response.. In other ment cannot be tolerated if controllerwords, if output pressures were plotted action depends solely on a clearaa against nozzle-baffle clearance, the re change. But oa our design continues sulting plot would be a nonlinear curve, we will discover that such smell derii Fig. 12. tions can be used advantageously. Such nonlinearity cannot be permit ted if the controller is to be truly pro portional in its response. Further experimentation with the baffle mechanism reveals that whenever . Alr-Roley Volvo. To make ute'ybl the improved linearity of the low-pig. eure nozzle, yet provide the desired airpressure range to the control valve, V conolude that an amplifier or air relay, backpressure at the nozzle reaches the Fig. 14, is required to transform -i'* upper limit, jet effect from the nozzle nouie impulse of 2 to 4 psi into tn out lifts the baffle (lightly, causing an put range of 3 lo 15 pai. ^ erroneous response In controller action. Supply air .is odriltted on the rigiu To overcome such a condition it appears side end feeds two branches, one to Uu logical to limit backpressure to a rela orifice, another, to the air valve. .TO tively low value, such at 4 or 5 psi. . orifice restricts flow of air to the noa] Orifice Restriction. As a final step in os in the previous designs, but here'tit our analysis, consider the capacity of air line.leaving the restricting block the on-off device to eupply air to a large branches out to feed both the nozzle diaphragm valve. We find the orifice a small bellows below the air valve# employed to restrict air flow to the noz* In this design any change In not& sle also limits the quantity of air de pressure causes a pusher pin, attache, livered to the diaphragm. Flow rate to to top of bellows, to position a small the valve being slow, its response will be boll between two seats, the upper faji sluggish. To gel rapid valve action, we being the supply port, the lower *eil Combtnotlon nozzle and reloy, Fig. have to use some form of booster air acting as an exhaust port As the boll 14, gives on olmost linear response valve. The design of a truly-propor- assumes a position between the. tw lional controller should, therefore, in seats, a given amount of air RowiUn clude the following features: the supply port and a correspond to 0.0015 in., causes backpressure to drop from 15 to 9 psi and move the controlling valve from full-open to 50%open position.. The valve.has thus been moved through botf its stroke by-a slight baffle motion, one-thousandth of op. inch. Mecboeltm Backlash. Linkages bo* tween the recording pen and baffle must be entirely free of backlash or lost mo tion if the baffle is to be accurately positioned to match the pen .arm'a posi tion. In building such a linkage we would undoubtedly introduce friction' because of tightly fitted pivots. As we 1. The baffle should have a total amount flows out the exhaust port'/ movement of at least 0.25 In. to compen The resulting pressure creeled lo^ sate for lost motion in the linkages. valve chamber varies with the Mjl! 2. Output pressure should change position, hence this pressure anil proportionally with baffle movement be utilized es the output (ram the reft. cause its position corresponds directly An air tube connecte the ait-valve chs&g with that of the pen. ber with the final control valve tad. 3. Noale backpressure should below labeled output. Capacity of the supp^ (maximum of .4or 5 psi) to improve the and exhaust ports can be made large linearity of response and to minimize enough to meet the demand lor rapj^ baffle deflection. filling a large-diaphragm chamber/:' 4. Controller should havo ample out Such a relay converts the imoU^i' put capacity to tepidly inflate large pube received from the nozzle to sn out' valve diaphragms. . put pressure of the desired magnitod As a starting point in designing en Craph in Fig. 15 combines the cb(|" improved controller, let's see how a . aeterlstics of e low-pressu*' nozzle (irf sought to reduce friction, the problem of lost'motion would again appear. After rebuilding the linkage system, we would conclude that the teheme was small-diameter nozzle acts when oper ated under low backpressures. Experi menting with a nozzle diameter in the range from 0.020 to 0.030 in. and a relay. The improved linearity ol resas obtained with the low-pressure n? __ zle and relay, will assist greatly I**;* Impracticable for a commercial-type controller and would turn to some means of obtaining greater baffle travel restricting orifice of about 0.011 in. we can plot backpressure versus nozzlebaffle clearance. Fig. 13. final performance of our proportion" controller. "v Second port o/ this article unll to lessen the effecu of lost motion. By It is evident that the response curve sign a system of mechanical (inAaf increasing baffle movement to 0.25 in., is far more linear than that obtained that uses bellows motion to position Iff we could permit an accumulated lost with the previous design where the full notsle and apply this mechanism ~ motion of 0.001 in. with little effect upon range of 3 to 15 psi was impressed on other refinements to a proportional. Instrument accuracy. This would permit the nozzle. We get the greater linearity troller. U (540) POWER S'Dta'nba'/1 Operotiog rotfs titted through font V.'RATIOADJU$TCR$cro!ocotedodjocenttotronformflr windings adjuster froze In position, quick repair was made by-cutting 'obdccrmoct by wood operating rods to external handles. After It free and Installing top Jumped with sotderless crimp lugs Crimp Lugs Speed Transformer Repair j>We recently rouno ourselves in a -\tough spot beesuse of trouble which .developed m the ratio adjusters of Aour power transformers. We have 'three 5000-kva, 27,000/13,600-v, single t phase,' oil-insulated outdoor transformeri in our plant, which are grouped (or ''bank operation. Ratio adjusters, in each 'coil group for changing taps while , transformer is deenergized, are the^.bird-cage type, consisting of studs and l rollers ss shown above. Rerio Adjttiter Detail. .Cables con- 1-nect from taps in transformer winding ,lo individual studs of tap changer. Note cable connections extending. through ^bottom support of ratio adjuster. ^Springs located inside the rollers exert neceisery pressure an studs. Rollers incomplete transformer circuit by ridg- tag two adjacent studs In lap changer. 'Micarta enclosure which surrounds '.ratio adjuster is not shown In above (ketch. Wood operating rods connecting ratio /adjusters to external handles ex- ^tad through cover so taps may be without dropping transformer levL The only possible way adjust- en ^ |s (0 drop .oil ' on lP core laminations and an adjustable periscope with Its mtaiature inspection lamp. "C.-.Tiinaformer load changed recently became necessary lo change Ups S' ow-voltage side. When ratio adjust- By K B HOFFMAN cessible so they could be cut without inglnoor, Burnt and Roe, fn<> removing core and coil assembly from tank. Since the work had lo be done in side'the transformer and in congested ers were operated, two on one trans quarters, it was decided that soldering former required considerable force to the tap leads together would be a haz turn while another was frozen and could ard to the men. not be moved. Compression Lugs Usad. It was de Problem. After taking an oil sample cided to make the connecting jumper ' from both top and bottom ol trans using solderless compression lugs. To former in which ratio adjuster could satisfy ourselves that this type connec not be moved, we found it dark in color tion would be satisfactory with the with carbon porticlet Boating in sample. many-stranded flexible cable, a lug was Oil level was then dropped exposing pressed on a sample section. A hand- ratio adjusters. Inspection showed a operated hydraulic preos was us4d to solid mass of carbon covering rollers exert approximately 8000 pil to press a end studs. In fact, contact surface be % In. slot along both sides of lug. : tween'both was filled with a quarter A similar section of cable was pre inch of carbon residue. pared with a soldered lug. The two , There was a mad scramble to locate .samples were given a tension test, pres replacement ratio' adjusters and as sure being applied between lug end luck would have it, none were In stock. and cable. Sample prepared with the Since the contact condition ol tbo ratio hydraulic press withstood.more tension adjusters was a potential hazard un .than soldered lug. and showed distor der load It would be dangerous to put tion at boll holes in lug before cable the transformers back in service os'they strands pulled free. were. High contact temperature or o .7 As a result we used corapression'lugs' spark could Ignite the oil. The trans lo make tap connections on all trans former bank had to go back in service formers and the bank was returned to as soon os possible since load' condi service in record time. No trouble has tions would not permit an extended out been experienced with the installation age. to date and the transformers have been Solution. After some discussion it was carrying overloads st frequent Inter decided to connect the proper winding vals. Periodic oil testa showed no color, taps foe the new voltage and disconnect chsnge snd no further contamination as the ratio adjusters. Tap leads were sc- evidenced by free carbon particles. tow FR StpUmbor 1948 fStli 87-