Document 4aMVGa3Q2777gMygKa4yZvOkx

TABLE I--OPERATION WITH LINK SET FOR THROTTLING RANGE OF 100/. Pan position on chart Baffte. bellows position, Fig. 7 0A 25 B so C 75 D too B fjota: For every pen petition, there la a corresponding petition Ter the notile Output ond bellowa pressure, paig Vaba opening, 30 6 25 9 so 12 75 IS too and baffle, o corre*ponding output presaure and control-valve opening. TABLE II--STEP-WISE ACTION OF RESET BELLOWS Load, Pressure Preasure of Position of Y, paig lover T X, paig PROPORTIONAL ACTION ONLY (TR-100%): 1. BeRowa Vcharged with air at 9 pti; valve R closed................................................................. 2. Load increased to 75%.................................... ADDITION OP RESET ACTION: 3. Valve R opened to admit output air (12 pti) to bellows V........................................................ 4. Ptoceta reaches new balance and original control point ia restored................................... 50 75 75 75 9 9 plua 12 12 C9 D 12 C 12 plus C 12 Noio: In actual practice, valve R'ia action at a rate that Is consistent with the partly opened in'order to provide reset stability of the controlled process. controller output to lead response rale of the process. That is, if the pen were lo move at a fixed rate away from the desired control point and air flow to bellows X, Fig. 4, were restricted by needle valve P, rate of follow-up motion would be retarded and the nozzle displaced from its normal relationship with the bafflle. Consequently, output pressure would be accelerated in relation to pen me* lion. Accelerated output thus obtained hoe great value in offsetting the ill effects esused by time lag in the pro* cess and measuring element To produce the greatest effect, the process would be arranged to incorpo rate time lag, such as is imposed by the restricting valve between the tanks in Fig. 3. Then using a motor-driven load ing valve to change the system load gradually, improvement in control due to rate response could be vividly demon strated by alternately opening and throttling valve P. When valve P throttles air to bellows X, controller output changee rapidly and causes control valve to supply ex cess water to left-hand tank, which in creases its level, and thereby assists inovercoming resistance to flow offered by the Umedog restriction valve in the pro cess. Without rate response, controller output appears sluggish and the process reaches an unstable condition because the instrument cannot oend an advance impulse to control valve to offset time lag. Finished Design. Our controller is now fully equipped with the three ad justable responses and the instrument is essentially the same as a conventional pneumatic controller. It contains the same elements and operatea on identical basic principles but each manufacturer usea a slightly different mechanical de sign for adjustments. Mechanism to adjust throttling range may be a simple arrangement of pivots and levers or a complicated assembly. Reset action may be accomplished by two opposing bellows or by applying the opposing pressures inside and out side a single bellows. Flow to bellows may be throttled by special needle valves, restricting tubes, or a unit con taining variable passageways. 'Next step is to design a mechanism that detects and corrects improper posi tion of the control-valve stem. But we must ovoid complicated mechanisms be cause the device may be installed on a valve in e _ corrosive atmosphere or where space Is limited. Volvo PoriHenct. Suppose the valve stem travels one-half Inch as diaphragm pressure is Increased from 3 to 15 psi. If an accuracy of 1% is required, the tern must slay within 0.005 in. ol ^ sired Kiting. Such cloie sdjuiimea^' requires a sensitive "feeler" device JK detect voriations of this mignitu^ Let's try nozzle-baffle syrtem u tW'rf "feeler," Fig. 6, j Mounted on right of volve frame is 4" w} bellows A opposed by a range ipring ' This assembly, a duplicate of the fej'vvsj low-up unit in the proportions! cea-'Vh` trailer, is designed so air-pressure change from 3 to 15 psi causes the be!- -Y lows to expand one-half inch. A osule,- C, 2s atioched to o plate mounted be-'**' tween bellows and spring. As bello*) Yi' expands or contracts, plate and noale.;^ move to a position proportional to tbe'-^K air pressure within bellows. An aiiv -supply connection, upper right sketch, feeds sir to a restriction R tad* Y thence to nozzle through a length ef'YY flexible tubing. A branch from the v nozzle line is connected lo diaphragm^ [ Remember that this some KhemeY'{ used in the on-off controller, esn detect','^ a baffle movement of thousandths ef ib`*xh inch. So that valve-stem position ms7^','b become a function of nozzle position, baffle D is mounted on a collar tisched'^ lo the valve stem. Once the baffle Un adjusted to a definite clearance with the nozzle, any motion of the nozzle eauset.'fc a change in backpressure sod Increases or' decreases the air preuuriyj^ transmitted to the diaphragm lop. Suppose the controller supplies thy*-;, impulse of 9 psi to bellows A, correct valve-stem position for pressure is midstroke. As the aoBlel,<d>4 assumes the new position dictated the bellows, baffle is manually adjusted-^! by sliding the collar on the steal uotil^^| the baffle restricts nozzle outflow Moogh ,-j^i to create a 9-pal backpressure. As diaphragm receives this impulse moves valve stem downward to midpoint of travel. Baffle support collar *haljT*3 now bo locked to the valve stem slaflU*^ the mechanism is in proper align It valve friction prevents movement,'baffle movement would temporarily retarded, nozzle would tinue to approach baffle, and back pressure on diaphragm top Increose to full output pressure, ot the force was enough to overcame .^1; lion. Accuracy of valve respond now been improved to equal the controller. Considerable credit is due J* Kowalski for ingenuity end workmanship in building tbs The author also expresses his to R L Clarke, F C Kitstit*. Koenig, E IF Jensen, R C M# - Miss D E IForren for sufttinonI. preparing the demonstration. 92 (498) power :;r 'Shvffor frantfarmar bootup.on remaining two photat %i* mTm { , j.. "T." Copocftfva | rtoefenca botwaon moehlna winding and ground ^.HOOKING TRANSFORMERS ecrots eoch phot* of opon oil ground on one feeder phase, rcsonont condition wo* opprooched circuit breaker is usuolly OK for phote-bur test. Because of with resulting torgo current flow. Finished job with hookup C. Ground Halts Generator Phase Test JOne of several methods may 'bo used in phasing-out a new * ck generator. E J Kelley, of Kansas City, Mo., pitked the one shown in A and ran into * resonant circuit condition insure our s new oc generator in small plant is ordinarily a routine job. Although It requires care it doea not as present any particular problem. *y, e **ihod that U often used is shown r,lB Aetch A. it ^eiitg-Out Procedure. Using the ijheokttp In A, with oil circuit breaker , m#c^ne k brought up to speed t>roPer voltage. Disconnect switches ./j**6 J^en dosed. Phase rotation of new ' tr* lamps, connected in illtnj*I?rTner Mcondary, all grow bright Y, together. When lamps In each ^*bt and go dork alternately, 'eorr -tertk*1, 0 new machine is incase it's necessary to in- .'ivi, *nB ,w of the machine leads, '"uolly done at tho ocb. lii*tme I'^r* ago we had on experience `'ktinctly out of the ordinory. of- *""1 Msanmte hookup, with machine tpw NovmtMr 1948 up to speed and voltage about normal, ing a series resonant, or partially res we found one of the dieconnecte had not onant, circuit. Capacitance of tnsehine's closed properly. With machine stilt winding to ground may be represented running, the disconnect wee opened. We as a wye-connected set of capacitors. received quite a surprise when a heavy Under oormal conditions the voltage arc broke, which indicated a large cur ocroes each capacitor would be the rent flow. On opening each of remain phase or line-to-neutral voltage. But ing phases we noted the same condi with a ground on one phase external to tion. Our first thought was low insula machine, voltage across that capacitor tion resistance in machine winding or would be zero. Voltage across each of leads. After megohmmeter tests showed ' remaining condensers would then be everything clear, we went ahead and the line-to-line voltage. took another crack at phasing out By applying this theory to the condl-^ On the third attempt lamps, con lion at hand, the large current flow can nected across transformer secondary, be explained. If cireuit were in pure followed no regular pattern In going resonance, the only opposition to cur bright and dark. We noted that lamps, rent flow would be the resistance; in when lit, were abnormally bright With most installations of any size this oppo one set exceptionally bright, the others sition is of negligible value. The pure would only glow. resonant condition would occur if the fecund On System. None of (hfs dis machine's capacitive reactance and the turbance seemed to follow the book of inductive reactance of transformer sad rules, so we looked for trouble outside machine were equal. If such a condition the machine proper. Our first clew exists it is possible to have the voltage came when we noticed the ground de across transformer end that between tector on the operator's board indicated machine and ground exceed the applied a ground on one of the feeders. Since voltage. Under that condition the only the system was ordinarily ungrounded opposition (o current flow would be pure we drew a few mental connections and resistance of the series circuit. decided this ground was the seat of our Since the fault was not cleared trouble. promptly we went ahead with our Referring to sketch B, it was evident work using a hookup shown in sketch C. that the grounded phase, primary of This method is used widely in many transformer and capacitance between plants for phasing out as well as syn machine winding and ground, was form chronizing. (699) 93