Trains.com

Locomotive MU systems

15564 views
48 replies
1 rating 2 rating 3 rating 4 rating 5 rating
  • Member since
    June 2022
  • 76 posts
Posted by Pneudyne on Wednesday, July 3, 2024 11:19 PM
As mentioned in the previous post, the notchwise engine speeds for diesel-hydraulic locomotives are not evenly spaced, but get progressively closer as the notch count increments, this due to the propellor curve loading imposed by torque converters.
 
I have not found the engine speed schedule for the Alco DH-643, but I do have that for the Krauss-Maffei (KM) ML-3000 prototype.  I understand that this locomotive in particular inspired the DRGW initial interest in the diesel-hydraulic concept, which then resulted in the DRGW-SP joint order for the KM ML-4000.  The ML-3000 had a pair of Maybach high-speed engines.
 
The ML-3000 engine speed schedule illustrates that the engine speed steps get smaller as the notch position escalates.  The power shown is the total for the two engines.
 
Notch                           Engine Speed                    Transmission
Position                        rev/min                             Input Power, hp
 
1                                  600                                          150
2                                  755                                          300
3                                  865                                          450
4                                  952                                          600
5                                  1046                                         800
6                                  1128                                         1000
7                                  1197                                         1200
8                                  1262                                         1400
9                                  1320                                         1600
10                                1372                                         1800
11                                 1420                                         2000
12                                1465                                         2200
13                                1508                                         2400
14                                1550                                         2600
15                                1590                                         2800
 
 
This sequence generally follows the propellor law cubic curve.  For example, from that curve, and based upon full power at 1590 rev/min, the half power point would calculate as 1262 rev/min, which is in fact the notch 8 number.  For a given hydraulic transmission arrangement, there is a fixed relationship between engine speed and engine load.  On the other hand, for a diesel-electric locomotive with some form of non-inherent load control, there is scope to vary the relationship.  That allows the use of equal notchwise engine speed increments, with appropriate loading then determined somewhat independently for each speed.
 
 
It may be noted that the propellor law curve applies only to transmissions where there is a torque converter in the circuit at all times.  There have been examples where say instead of three torque converters, each covering a section of the speed range, a combination of torque converters and fluid couplings was used, for example two converters and one coupling as used on the early examples of the DB, Germany V160, and on the SNCB, Belgium Types 211 and 213.  The fluid coupling, used in the top speed range, was considered to be more efficient.  But it also compromised power utilization, in that it forced the engine to run along its torque curve.  In those cases, speed control governors were not used.  Rather the governors were of the min-max type with power control, either stepwise or continuous, by varying the rack position.  From an MU viewpoint though, it made no difference.  The same throttle notch control input could be used for either form of control.
 
 
Returning to the DH-643, on the power control side, it was evidently designed to respond to the standard AAR MU protocol, but with the addition of one control wire from the local master controller that allowed access to the “odd” notches.  I imagine that this additional control wire was trainlined, using a non-conflicting spare in the 27-wire bundle, but that is unconfirmed.  (And, one imagines, a different spare to that which was used for the “SN” function in the GE 16-notch system.)  If so, a DH-643 trailing in a consist with a DH-643 leading would operate in 15-notch mode, whereas a DH-643 trailing in a consist led by a standard diesel-electric locomotive would operate in eight-notch mode, that is, using the “even” notches only.  (As a corollary, it would appear that a 16-notch diesel-electric locomotive, GE or Alco C-855, and a DH-643 would see each other as eight-notch locomotives were they to be MU’d.)
 
 
Cheers,
  • Member since
    June 2022
  • 76 posts
Posted by Pneudyne on Sunday, June 30, 2024 12:08 AM

Pneudyne

 

 
 

 

Absent hard information on the Alco control approach, one is left to speculation.  Quite possibly it did use the GE KC99 master controller on the DH643, which as I understand it was also used for the C855.  And given the precept of maximum commonality with the existing diesel-electric fleet, I should expect that it would not have departed from the use of its standard 251 engine governors of the time, namely the Woodward PG and the GE MG8.  The PG could be configured for up to 15 engine speeds.  The MG8, with electrical resistance speed control, had no inherent limit, so probably could have been configured with 15 or 16 speeds.

 

 

 

 

 

I have since learned that the DH-643 was fitted with a GE KC102 16-notch master controller.  In power mode, notch 1 was dead and notches 2 through 16 provided the 15 steps.  In hydrodynamic braking, notches 1 through 16 were used.
 
Furthermore, it used Woodward PG electrohydraulic engine governors modified to use the overriding solenoid (ORS), not required with hydraulic transmission, as an additional speed control.  Exactly how it was configured is unknown, but most likely is that it provided an intermediate engine speed increment.  Thus, the usual eight speeds – notches 2, 4, 6, 8, 10, 12, 14 and 16 in this case. were provided by the A, B, C and D solenoids using the standard compatible pattern, with ORS providing the increment for notches 3, 5, 7, 9, 11, 13 and 15.  Thus the 15 notches could be provided with just one extra trainwire, controlling ORS, and no translation mechanism was required to/from the standard AAR pattern.  Although the A, B, C and D solenoids by themselves could be configured to provide 15 speeds, a translation device (sequence converter) would have been required to match the resultant pattern to the AAR pattern.
 
(15 speeds with A, B, C and D would require:  A = 1, B = 4, C = 2, D = negative 0.5.)
 
Note that the 15 engine speeds would not have been evenly spaced, unlike the typical diesel-electric case which has even spacing.  The hydraulic transmission loads the engine according to the propellor curve, approximately cubic.  As an example, for a 1000 rev/min engine, the half-power point would be a little under 800 rev/min.  In the AAR sequence, notch 5 is around half power, so that is notch 10 on the DH-643 throttle handle.  So, notches 2 through 10 cover say 400 to 800 rev/min, and 10 through 16 cover 800 to 1000 rev/min.  The engine speed increments get progressively smaller as the notch count increases.  As far as I know, this is done by appropriate choice of the governor speeder spring law.
 
 
Cheers,
  • Member since
    June 2022
  • 76 posts
Posted by Pneudyne on Saturday, June 29, 2024 11:19 PM

Pneudyne

 

In that regard, I understand that Lima-Hamilton did use a different solenoid pattern, namely:

 

Shutdown         D

1st speed           no solenoids active

2nd speed          A

3rd speed          C

4th speed           A, C

5th speed           B

6th speed           A, B

7th speed           A, B, C, D

8th speed           A, B, C

 

I do not know the associated speed increments associated with that solenoid pattern, nor the respective engine speeds.  But assuming that the 7th speed was somewhere between the 6th and 8th speeds, then it is clear that the magnitude of D must have been less than the magnitude of C, as well as opposite in sign.  In this case compatibility was maintained by using a sequence converter between the master controller, which generated an AAR-type output, and the governor.

 

 

 

 

To add to that commentary, I understand that Lima-Hamilton did use eight equally-spaced engine speeds.  In that case, its particular solenoid pattern would produce that result with the following values:
 
A = 1
B = 4
C = 2
D = negative 1
 
Nonetheless, why it departed from established practice is unknown.
 
 
Cheers,
  • Member since
    June 2022
  • 76 posts
Posted by Pneudyne on Friday, June 14, 2024 2:46 AM

Erik_Mag

The "odd" control sequences look a bit similar to a Gray Code, which strictly has only one bit changing between steps. This is a bit safer than straight binary when using a rotary encoder (throttle controller in this case), where a slight error in positioning whatever encodes the individual bits would lead to short term spurious codes being sent when the encoder/controller was moved.

 

 

There is evidence that simple binary rotary encoder errors were a potential problem in locomotive control systems.

 

An example was the Westinghouse UK Westcode electropneumatic braking system (for MU cars) which provided seven braking steps using three solenoids in various combinations.  To quote from the pertinent ILE paper:

 

As has been mentioned above, the control of pressure graduations in the brake cylinder is achieved by means of three train wires which are energised in different combinations, eight of which are available if the “off” position with all train lines de-energised is included. The control of these train lines lies in the brake valve, where three cam operated snap action switches are fitted. The switching sequence, which is indicated in Fig. 4 is so arranged that only one change is made at a time, i.e. either a switch is made or it is broken. If two or more simultaneous changes were attempted with a simple switch it would be inevitable that in the transition stage an incorrect combination could occur, unless impracticable limits were put on the manufacture and adjustment of the mechanism. The choice of a single change sequence therefore results in simplified construction of the brake valve, and a switch assembly which is not critical and in fact can be assembled into the brake valve without the need for any adjustment.

 

Another example was provided by the JNR, Japan DE50 diesel-hydraulic prototype built by Hitachi.  This had 14-position throttle control using four solenoids, I think generally following an established JNR  pattern.  The Google translation of the appropriate Hitachi commentary gives:

 

This is the first attempt to use an alternating binary signal for the four types of notch signals output from the main controller.

 

The electromagnetic solenoids for notch control of the engine and dynamic brake operate with a pure binary signal, so the segment method of the main controller has traditionally also been a pure binary system, but it has taken a great deal of effort to manufacture and adjust the segment method so that it does not enter a notch that is too far away when crossing a notch.

 

This suggests that unacceptable transient errors were possible with a simple binary system.  “Alternating binary signal” does suggest one change at a time.

 

On that basis it is reasonable to deduce that the AAR throttle solenoid sequence was devised at least in part to minimize transient switching errors that could arise with a pure binary system.  Another possibility was to provide the “D” unit with some “exercise” during normal throttle operation, so that it was not just a shutdown device.  That may have been more of an issue with the external electropneumatic governor operators – whose individual cylinders might be inclined to “stickiness” if not much used - than with the internal solenoids of the electro-hydraulic versions of the Woodward New SI and PG governors.

 

 

 

Cheers,

  • Member since
    June 2022
  • 76 posts
Posted by Pneudyne on Tuesday, June 11, 2024 11:59 PM

Rainhilltrial

Pneudyne ... would like to know what materials you may have re the GE UM12C for Thailand.  Speed v TE curve? Any manuals? Drawing of the GSC truck?

 

The major source material that I have on the GE UM12B for RSR, Thailand is a Railway Gazette 1964 June 05 article, most likely based upon material submitted by GE and then used by the RG editorial staff.
 
The three-and-a-half page article included an elevation drawing, overhead and side-on equipment layout drawings, and a three-quarter view picture of a truck.  There is a tractive effort-speed curve.
 
The C truck was of the GE semi-cast, floating bolster, non-equalized type.  It had non-standard wheelbase of 12’2.25”, split 5’11” outer and 6’3.25” inner.  At the time, the standard C truck wheelbase for export Universals was 11’6”.
 
It appears to have been an early, if not the first application of a non-equalized C truck to an export Universal.  At the time, U20C production was still using the equalized C truck, but I haven’t checked the U13C case.  Abandoning the equalization might have been a weight-saving measure in the case of the UM12C, which was built to an axle loading of around 27 500 lb.  It may be noted that the “lightweight” UM13C built for Pakistan Eastern later in 1964 also had non-equalized C trucks, in this case of 11’6” wheelbase.
 
The GE semi-cast floating bolster trucks date from the 1961 export Universal range adjustment, if not before, and appear to have superseded the original fabricated box frame type.  The fully cast option date from c.1957, apparently because South African Railways insisted upon GSC cast trucks for its U12B order.
 
From the Moser List, the engines (two per locomotive) were shown as Cummins VT12-825B, the main generators as GT605, and the traction motors as GE761.
 
To stay aligned with this thread, it may be noted that the UM12B was said to have standard GE control equipment.  It had a 75-volt auxiliary electrical system and almost certainly eight-notch power control with the standard throttle solenoid pattern.  Load control was of the inherent characteristic type using a split-pole exciter, as GE did with the smaller, Caterpillar-engined export Universals.  It appears to have been fitted for MU.
 
 

 

 
 
 
 

 

Cheers,
 

 

  • Member since
    July 2010
  • 13 posts
Posted by Rainhilltrial on Sunday, June 9, 2024 1:59 PM

Pneudyne ... would like to know what materials you may have re the GE UM12C for Thailand.  Speed v TE curve? Any manuals? Drawing of the GSC truck?

  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Tuesday, September 13, 2022 10:42 AM

One fundamental problem with MU-cable wreck damage is very similar to a problem at the Brown's Ferry nuclear powerplant.

You may recall this as the plant where not-so-bright design engineers optimized the number of 'penetrations' in the containment by running most of the cabling, including supposedly redundant systems, through one channel, and then more not-so-bright personnel tried checking an air leak in the channel with a candle.  The insulation on the cables did not sustain combustion... but it supported it, and the candle flame went back along the channel, burned off the insulation en masse... and all the little wire conductors started touching each other.

At random

With heaven knows what shorts between what conductors.

When an MU cable is pinched in an accident, the conductors may not be severed or attenuated; they may easily be clamped together in such a way that the insulation on the wires is displaced so the conductors come into contact with each other.  I can remember at least one accident report, many years ago, where MU cable damage of a specific kind was reported as a causative factor -- someone here will remember what it was.

  • Member since
    December 2007
  • From: Georgia USA SW of Atlanta
  • 11,919 posts
Posted by blue streak 1 on Tuesday, September 13, 2022 8:11 AM

The Saturday NS deailment brings up a  question.  Is there any way the 27 point MU cable damage could disable some functions of an loco?  What is the effect of the MU air hoses going to atmosphere when locos have a wreck separation as what did happened ? 

  • Member since
    December 2001
  • From: Denver / La Junta
  • 10,820 posts
Posted by mudchicken on Tuesday, September 13, 2022 6:30 AM

jeffhergert

 

 
SD70Dude

You can get the MR and trainline gladhands to go together, but you have to twist one hose around 180 degrees, and even then I don't think the gladhands will 'lock' together properly. 

 

 

Our local mechanical trouble shooters have rigged up a hose that has a gladhand that fits the brake pipe on one end, the other end's gladhand will fit the main reservoir hose.  They did this so they could recharge the MR on a dead DP equipped with the air start system.  This way they don't have to get a live engine next to the dead one to restart it.

It comes in quite handy when the mid-train DP has the auto stop shut down, but then doesn't automatically restart and then the MR bleeds down.  That happened to me once.  I figured we would have to set over half the train to the other main track to get the head end next to the dead DP.  The mechanical guy came out, applied the compromise hose and was able to get enough pressure in the main and the auxilary air start reservoir to restart it.

The bad habit of DPs shutting themselves down, but not starting up again when needed is why I don't care for the Auto Engine Stop/Start systems.  At one time, engines in DP mode weren't supposed to have the AESS systems active because of that problem.  It doesn't happen everyday, but way more often than it should.

Jeff    

 

 

A fair number of roadmasters/MTM's have MR-to-AB brake hose connector gladhands for use with airdumps to have better control of the airdumps (esp the leaky ones with bad air cylinders/rings)  and to keep the train from going into emergency and dumping the air while in use. Those rascals are guarded jealously by the track forces to keep them from getting lost because they are not common. Being that the airdumps are generally the last to be maintained by the mechanical department, are subject to hard use in worktrain service and are extra maintenance items to the mechanical department, adding a little extra pressure from the main air reservoir is often the only way to get some of the big piston drums to get enough air to charge/lift and dump their loads.(trainline air won't do it alone, especially if there are too many cars with small air leaks)...Gratefull there is more than one way to safely unload those cars in the field.

There are plenty of instances where you want to get the load off the airdump doors by simply moving the train.

Mudchicken Nothing is worth taking the risk of losing a life over. Come home tonight in the same condition that you left home this morning in. Safety begins with ME.... cinscocom-west
  • Member since
    June 2022
  • 76 posts
Posted by Pneudyne on Wednesday, August 24, 2022 4:57 PM

Pneudyne
The situations where the standard system was interfaced with control/MU systems that were somewhat different or quite different.

 

Some examples:

 

The SP diesel-hydraulic case.  Finding the details is a work-in-progress, see:  https://cs.trains.com/trn/f/111/t/65687.aspx?page=2#3410266.

 

 

By way of background, the need for MU compatibility between diesel-hydraulic and diesel-electric locomotives had been called out several years before the DRGW and SP experiments.

 

In a 1953 ASME paper (53-A-122) by J.S. Newton of Baldwin-Lima-Hamilton, Hydraulic Torque-Converter Transmissions for Locomotives’, the author had provided an 11-point list of desiderata for diesel-hydraulic locomotives intended for freight of road switcher service in the USA, as follows:

 

(1)  The drive should be designed to develop constant horsepower over a minimum speed range of 5 to 1 and 8 to 1 is desired.

(2)  The efficiencies at all loads and speeds should be equal to those of an electric transmission.

(3)  The power unit should be mounted in the locomotive underframe—not on the trucks.

(4)  The design should permit quick and easy removal of individual axles as well as complete trucks.

(5)  Trucks should be suitable for high-speed service—swing bolster or the equivalent.

(6)  The locomotive with train must negotiate curves of 21 deg.

(7)  Multiple operation of units must be possible.

(8)  Multiple operation with locomotives having electric transmissions is desirable.

(9)  The drive should permit braking equal to the dynamic brake of the diesel-electric drive where braking power is usually greater than pulling power.

(10)  Tractive effort must be equal in either direction of motion.

(11)  The drive must be arranged for zero torque at the wheels when the locomotive is stationary.

 

B-L-H would have been well aware of the consequences of being the “odd man out” when it came to MU compatibility in the US domestic market.

 

Nonetheless, the 1961 K-M prototypes for DRGW and SP were not built to be MU compatible with the US DE fleet.  At the time the DRGW and SP locomotives were ordered, the German industry had not yet adopted a common control and MU system for diesel-hydraulic locomotives, and in fact was still just coming off a learning curve when it came to the degree of “fineness” that was actually required for power control.  The main customer, DB, was buying only diesel-hydraulic locomotives, so there would have been no need to seriously consider mixed MU operations.  Perhaps the German builders thought that such would have been difficult to arrange, and even if it could have been done, was undesirable.

 

Be that as it may, both DRGW and SP found the lack of MU compatibility to be a drawback, and in 1963 modified their diesel-hydraulic locomotives accordingly.  Apparently DRGW developed a two-way interface that worked at trainline level, effectively doing both electric-to-pneumatic (DE to DH) and pneumatic-to-electric (DH to DE) translations.  That would have been an interesting device, but I should not have much expectation that any detailed information about it remains available.  On the other hand, SP changed the master controllers, installing the same GE KC99 type as used on the GE U25B and as had been specified for the forthcoming production fleet of K-M locomotives.  This required just a one-way interface, from the GE/AAR protocol to the K-M pneumatic power and dynamic brake control protocol, the details of which we may reasonably assume will surface sooner or later.

 

Then there was the Alco DH643.  Alco wrote an SAE paper about this, SAE 650420, ‘The Design-Development of a Mainline Diesel Hydraulic Locomotive’, by H. W. Schmidt

 

The opening statement of that paper was:

 

‘THE DESIGN AND DEVELOPMENT of a mainline diesel-hydraulic locomotive was undertaken with the potential of greater adhesion, lower possible continuous speed than a diesel-electric for the same horsepower per axle, and lower maintenance on the transmission system.

 

Design parameters were that:

 

1.     Locomotive must be compatible and operate in all functions in multiple with existing and future domestic diesel-electric locomotives.

2.     Locomotive must be maintained and serviced with existing facilities and personnel.

3.     Driving controls and arrangement must be such that crew instruction is minimized.

4.     Locomotive to use existing expendable parts such as brake shoes, wheels, oil filters, air filters, drawgears, couplers, springs, etc.

5.     Locomotive construction and arrangement be as close as possible to existing units to facilities routine servicing.

6.     Locomotive to be of road-switcher type commonly used domestically and be of conventional appearance.

7.     Locomotive to use a medium-speed, heavy-duty diesel, engine of proven performance and readily available spare parts.

8.     Locomotive to be equipped with standard-type running gear suitable to American roadbeds and service.’

 

MU compatibility had been elevated to numero uno.  Unfortunately, the paper did not explain how that was achieved.  It did mention that the hydrodynamic brake had 16-step control.  But the diagram of the transmission and hydrodynamic brake assembly showed the eight-step (three piston) electropneumatic brake controller, labelled as ‘8-point filling regulator’  Oops!  I think that was a stock Voith diagram of the time.  There was a later Voith diagram showing the 16-step (four piston) brake controller, and it is known that Voith offered eight-step, 16-step and stepless options.

 

Absent hard information on the Alco control approach, one is left to speculation.  Quite possibly it did use the GE KC99 master controller on the DH643, which as I understand it was also used for the C855.  And given the precept of maximum commonality with the existing diesel-electric fleet, I should expect that it would not have departed from the use of its standard 251 engine governors of the time, namely the Woodward PG and the GE MG8.  The PG could be configured for up to 15 engine speeds.  The MG8, with electrical resistance speed control, had no inherent limit, so probably could have been configured with 15 or 16 speeds.

 

 

Cheers,

  • Member since
    June 2022
  • 76 posts
Posted by Pneudyne on Saturday, August 20, 2022 11:49 PM

Some further background on some elements of the development of the MU common standard was provided by a couple of brief items in ‘Railway Locomotives and Cars’ 1954 October, referring to both Baldwin and Fairbanks Morse locomotives.

 

The Baldwin item (p.75) referred to developments over the preceding year, and was titled ‘B-L-H Diesel MU With Those of Other Builders’.

 

Quoting the pertinent part of the item:

 

>>>>>>>>>>

 

1,600-HP Locomotives for Operation with Other Makes - During the past year a dynamic braking controller and field loop circuit was developed and units were placed in service equipped for dynamic braking of B-L-H locomotives in multiple with General Motors units.  The braking excitation (field loop) circuits were designed to provide equal braking effort on either of the manufacturer's units.  The new controller produces a lower braking excitation (field loop current) on all units over the first portion of the braking range.  This requires greater movement of the dynamic braking lever and thus a smoother braking control at high speeds.  Units also were placed in domestic service that will operate in multiple with General Motors units in braking and motoring and with Alco units in motoring only.  In order for these units to multiple with certain locomotives making use of manual transition, the master controller has a transition drum.  A transition lever (selector) which rotates the drum controls manual-transition locomotives operating as trailing units in multiple with a Baldwin unit.

 

The transition lever also controls dynamic braking of the Baldwin unit operating singly or dynamic braking of the Baldwin unit when the lead unit is in multiple operation.  An intermediate dynamic braking stop position prevents complete lever swing to maximum braking without some delay to manipulate the catch release as an instantaneous maximum application may cause serious damage.

 

>>>>>>>>>>

 

Presumably that referred to Baldwin units fitted with the alternative Wemco 8-notch electric throttle control, which I think used the XM-781 master controller.  A careful interpretation of what was a said was not that there had not been previous use of that system on Baldwin locomotives, but that such would not have had an MU’d dynamic brake control.  Thus the 1953 innovation seems to have been Baldwin’s (or more likely Westinghouse’s) development and application of a field-loop controlled dynamic braking circuit that suited the Baldwin electrical equipment, and whose ‘control curve’ could be arranged for reasonably compatible MU operation with EMD units.

 

Also in 1953 was the apparent addition of a transition control handle to Baldwin version of the Wemco master controller that allowed the control of trailing Alco and EMD units that had manual transition.  The fact that Alco units could be controlled in motoring only indicates that with this round of changes, Baldwin had not introduced a compatible potential wire dynamic brake control.

 

Additional to the above, it was noted that a humping control option had been added for the 1600 hp locomotive.  Whether or not that was MU’able was not said, but one could take the lack of clarification as a ‘no’, assuming that MU’d humping controls were not yet generally in use.

 

 

The Fairbanks Morse item (pp.77,78) was ‘F-M Improves Its Locomotives’.  The last two of a long list of improvements were:

 

>>>>>>>>>>

 

Hump Control.  -With this deice, the power of a locomotive can be very gradually reduced to keep the speed constant as cars are cut off in a humping operation.  Conversely, it can be used to apply power slowly when starting a train and so minimize wheel slipping.

 

Dual-Circuit Dynamic-Brake Control.—Units with the field-loop type of dynamic braking control would not MU in dynamic braking with units using General Electric load control, and vice versa.  With a dual-circuit control, which can now be furnished, a leading Fairbanks-Morse unit can control the dynamic braking in a trailing unit employing either type of braking.

 

>>>>>>>>>>

 

The second of those supports the notion that F-M was the first to introduce a ‘two-way’ dynamic braking control.  Possibly that development was done in conjunction with Westinghouse, although I understand that F-M had some in-house electrical equipment capability.

 

Regarding the hump control, F-M patented an MU system for this, described in US2908852 of 1959 October13, filed 1956 September 10.  It looks as if the MU capability was a later addition to the original, which was probably not MU’d.  The MU control system was basically a servomotor-driven rheostat controlled by a variable voltage, 0 to 74 volts.

 

At some stage, probably later, I understand that AAR trainwire #1 was used for hump control purposes.

 

Cheers,

  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Wednesday, August 17, 2022 9:44 AM

blue streak 1
did creep control MU?

Be no point in it.  The control has to be fast-loop to the motors as they are either monitored carefully for slip 'as soon as detected' or actually operated in microslipping with the actual slips being quickly (within some fraction of a degree of wheel rotation which Pneudyne will know) "arrested" by power modulation.

Since it is individual to the locomotive, there is no particular point in trainlining actuation, other than perhaps to set up creep mode for other locomotives so equipped before they reach a particular bad section of track.  This would be an effective waste of a wire in the 27-pin cabling for little if any real-world gain.

I first saw this in action in the mid-Nineties, on KCS in Shreveport, where a couple of switchers that were nearly visually identical were trying to pull an impossible number of heavy empties up a steep approach grade.  They were in MU, with the 'creep control' on one audibly operating (the wheels 'ring' or chitter as adhesion is broken and made) and the other not.

"Hump control" (which is very fine output load reduction as cars are progressively released as the cut comes entirely onto the hump) is something different.

  • Member since
    December 2007
  • From: Georgia USA SW of Atlanta
  • 11,919 posts
Posted by blue streak 1 on Tuesday, August 16, 2022 4:36 PM

did creep control MU?

  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Tuesday, August 16, 2022 10:50 AM

Thank you.

This thread alone makes it worthwhile to keep reading the forum threads.

  • Member since
    June 2022
  • 76 posts
Posted by Pneudyne on Monday, August 15, 2022 9:04 PM

Overmod
I thought it would be easy to find 'manual' pictures of the connection arrangements for BLW power, or to locate some of Matthew Imbrogno's comments on modernizing the system for shortline and special use.  I am sorry to have failed in this so far.  I am sure there are people here, or following the Classic Trains forums, who will know; someone might consult the nooks and crannies of the Bakdwin Diesel Zone site or a Wayback version of it to see what's there.

 

Probably not quite what you are looking for in that it does not address the modernization aspect, but some information on the BLW control and MU system as it was, is available in this book:

 

Diesel-Electric Locomotive Handbook – Electrical Equipment

George F, McGowan

Simmons-Boardman, 1951.

 

The chapter on Baldwin (46 pages) provided a description for and schematics of the pneumatic throttle control system, including dynamic braking, based upon the CE100 master controller as typically used on road locomotives.

 

A good description, with diagrams, a schematic and a main generator curve, of the pneumatic throttle power control system is provided in AIEE paper 47-36 of 1947 January, ‘A 3,000 Horsepower Diesel-Electric Locomotive for the Seaboard Air Line Railway’, by D.R. Staples, Baldwin, T.L. Weybrew and C.A. Atwell, both Westinghouse.

 

To quote just a little of the detail, throttle line pressure is zero at idle and 7 lbf/in² at power on.  Up to 21 lbf/in², it progressively increased excitation to maximum, via the carbonstat, with the engine still at minimum speed.  Thereafter, the engine speed was progressively increased to maximum over the range 21 to 60 lbf/in².  The load control was of the single point type, acting at full rack to back off the excitation.  That was quite different to say the EMD approach, which assigned unique load control points, monotonically increasing, to each throttle notch/engine speed.  That the Baldwin engine was of the flat torque curve type probably facilitated the chosen power control arrangement.

 

That control and MU system, in conjunction with the Woodward UG8 governor, was used by Baldwin’s Belgian associate and licensee, Cockerill-Ougree, on new builds into the 1960s.  I think that in the USA, Baldwin might have adopted the Woodward PG governor for some of its late builds, but that is unconfirmed.

 

 

Cheers,

  • Member since
    June 2022
  • 76 posts
Posted by Pneudyne on Monday, August 15, 2022 9:02 PM

Pneudyne
The GE 16-notch system.  Only partial information seems to be available.

 

The type of short description often given for the GE 16-notch system is typified by this from the ‘Trains’ magazine 1968 December article:  ‘Incidentally, the 16-position GE throttle introduced with the U25B does not break the pattern of compatibility.  The throttle still has only seven engine speeds above idle, with every other notch merely providing more finely graduated generator field excitation.  Eight-position throttle engines M.U.’ing with the GE’s disregard these excitation-only stages.’

 

I have not found a comprehensive treatment of the topic.  The GE Educational Manual ‘General Electric Model U25B Diesel-Electric Locomotive’ (GEJ-3815) certainly helps quite a bit, but it leaves some gaps and does not, for example, contain any main generator or load control curves.  What follows is derived and inferred from various sources, and not guaranteed to be correct.

 

The main objective of the system appears to have been to obtain finer graduation of starting and low-speed tractive effort, rather than finer graduation of running power, than could be provided with eight notches, this being considered desirable for locomotives that had a relatively high power per axle.  That same rationale had been used with the GTELs, which were provided with 20-notch controls.  One imagines that in the U25B case, the choice of 16 notches was made to provide a system that was readily compatible with the existing 8-notch system.

 

Accordingly, there were 16 steps of main generator fixed excitation, corresponding to throttle notches ½, 1, 1½, etc, through 8.  This provided 16 basic convex and heavily drooped main generator curves, the maximum current point of each chosen to provide appropriate and closely spaced steps of starting tractive effort where they intersected the standstill line.

 

There were eight engine speeds, set by the customary Woodward PG governor, which also set the engine load point for each speed, and through the action of the load regulator rheostat, applied eight supervening hyperbolic constant-power curves to the 16 basic generator curves.

 

1st engine speed was used for notches ½, 1 & 1½, 2nd speed for notches 2 & 2½, 3rd speed for notches 3 & 3½, and so on through 7th speed for notches 7 & 7½.  8th speed was used for notch 8 alone.

 

The net result was that adjacent notches, such as 4 & 4½, shared the same supervening constant power curve, so that over some of the operating speed range they would not have been any different, although at the extremes they would have differed.  At the low power end, where load regulator had but little effect in typical GE control systems, the overlap would have been less, and perhaps notches ½, 1, and 1½ were largely distinct across the operating range.

 

GE must have been satisfied that finer graduation of standstill/low speed tractive effort, but not of running power, was sufficient, and that additional complexity was not warranted.

 

The 16-notch control was achieved in an interesting way.  The throttle handle controlled a 16-step rheostat in the master controller.  This set and supplied the exciter battery field voltage not just for the leading unit, but for any trailing U25B units.  This voltage, and the requisite current, was relayed to any trailing units via the dynamic brake excitation XB trainwire.  Effectively, the motoring excitation current for all units in a GE 16-notch consist was supplied by the leading unit.  Insofar as the XB trainwire was not otherwise used during motoring, its use in this role did not create any conflicts.

 

Of course, trailing GE units needed to 'know' that a GE unit was in the lead.  To this end, a spare trainwire, designated SN was used to signal to trailing GE units that there was a GE 16-notch unit in the lead that signalled trailing units to source their excitation from the XB trainwire.  The SN trainwire was live whenever the throttle handle was in any position from ½ through 16, and controlled the MR relays on individual units.  This relay, when operated, switched excitation source to the XB trainwire.

 

Engine speed was controlled by the AV, BV, CV and DV trainwires from the throttle handle in the usual way.

 

If a GE 16-notch unit found itself in a consist headed by a regular 8-notch unit, the SN trainwire would have zero volts on it.  Thus it would switch to local resistor ladder control of excitation, in 8 steps, controlled from the four throttle trainwires, AV, BV, CV and DV.  That form of motoring excitation control had previously been used by GE on the Export Universals fitted with Cooper-Bessmer engines and three-field control.

 

The same 16-step rheostat in the master controller was also used to control excitation during dynamic braking, again via the XB trainwire.  Also again, excitation current for the trailing units was supplied by the lead unit, with the SN trainwire switching the MR relays to achieve this.  The same general form of dynamic braking excitation control had also previously been used on the Export Universals, albeit with continuously variable potentiometer operated by the selector lever.

 

If a GE unit was trailing a regular unit, lack of a signal on the SN trainwire would cause it to revert to local control of dynamic brake excitation, using its own load regular rheostat controlled from the XB trainwire from a micropositioner.  I imagine that that approach recognized that not all dynamic brake potential wire control systems could supply sufficient current for excitation purposes, only enough for control purposes.

 

Thus the doubling of the notch count apparently was achieved with the use of just one extra trainwire.  Presumably the SN wire was an existing spare in the 27-wire bundle, although which one is unknown.

 

It is conceivable that the later advent of the GE Type E excitation system allowed improvement to the 16-notch control.  With Type E, close approximations to constant power hyperbolic curves could be constructed electrically, with the load regulator then serving simply as a trim and protection against abnormal conditions.  Thus, each of the 16 notches could have had its own constant power hyperbola.  Whether this was actually done I do not know.

 

What would be useful here would be the finding of the main generator curve sets, covering all notches, say for the U25B and for a later GE model with 16-notch control and Type E excitation.  And also perhaps for the Alco C855, which is understood to have had 16-notch control.

 

 

Cheers,

  • Member since
    June 2022
  • 76 posts
Posted by Pneudyne on Monday, August 15, 2022 8:58 PM

Mentioned in my June 26 post as one of the sidebar aspects of the AAR MU system was:

 

“The UP system that allowed leading GTELs to control trailing diesel-electrics.  There seems to be no detailed information available as to how this was done.”

 

 

Starting in 1958, most but not all of the UP GTEL4500s and GTEL8500s were fitted to MU with trailing diesels.  It appears to have been a UP initiative; whether GE was also involved is not known.

 

I have not been able to find any information as to how the 20 notch GTEL throttle control was mapped to the eight diesel notches.  Maybe that level of detail has never made its way into the railfan domain.  My guess, and it is only that, is that the mapping was arranged so that diesel notch 8 was reached somewhat ahead of GTEL notch 20.  This would have recognized that the load control system on the GTELs could sometimes supervene before notch 20 was reached, in which case the throttle handle would be held at a lower notch.  This item in the GTEL4500 operating manual (GEJ-1943) is pertinent:

 

“Notching Guide – Red Band – When throttle handle is advanced, the pointer swings toward or into the red band.  If pointer indicates in the red band continuously, throttle handle must be notched back until pointer indicates in the yellow band.  No reduction in power results from this action.”

 

(The notching meter was additional to the conventional electrical load meter.)

 

It was probably desirable that trailing diesels could reach notch 8 even when the throttle handle had to be moved back from notch 20.  Having diesel notch 8 correspond to say notch 15 or thereabouts on the GTEL control would have achieved that.

 

Power control on the GTELs was done by a potentiometer in the master controller, which supplied a variable voltage to the excitation system (Amplidyne for the first GTEL4500 batch, Type E (or an early precursor thereof) for the GTEL8500 fleet, and unconfirmed, but I think Static for the second GTEL4500 batch).  For control of trailing diesels, a straightforward approach (but not the only option) would have been the addition of cams and switches in the master controller to provide the AV, BV, CV and DV governor control signals.  Forward, reverse, generator field control, etc., probably could be taken from the appropriate GTEL trainwires without the need for any interfaces.  Similarly the GTEL could have provided control voltage for the diesel, unlike the Milwaukee electric + diesel case where there was full separation of the respective auxiliary electrical systems.

 

The GTELs had the same potential wire dynamic brake control as on Alco-GE diesel-electrics, so they were directly compatible with (some) diesels on that front .  Although if field loop dynamic brake control compatibility had also been required, then the brake control potentiometers would have needed to have adequate current capacity.

 

 

Cheers,

  • Member since
    June 2022
  • 76 posts
Posted by Pneudyne on Friday, August 12, 2022 11:12 PM

The previously discussed Milwaukee Wylie throttle used to allow DC electric locomotives to control trailing diesels was an early example (electromechanical era, one might say) of mapping from many notches to eight diesel-electric notches.

 

From the same era, the EMD FL9 dual-mode locomotive of 1956 provided an example of mapping the other way, allowing control of what was effectively a multinotch DC electric locomotive from an eight notch diesel controller.

 

The details were provided in the book Diesels to Park Avenue, by Joseph R. Snopek and Robert A LaMay, as follows:

 

Diesel Notch          Electric Notch

 

1                                                    Series-parallel manual 1

2                                                    Series-parallel manual 2

3                                                    Series-parallel manual 3

4                                                    Series parallel manual 4

5                                                    Series-parallel automatic steps 1 to 17.  Last step (17) is full series-parallel.

6                                                    Parallel automatic steps 18 to 26.  Last step (26) is full parallel.

7                                                    Parallel automatic step 27, first field shunt.

8                                                    Parallel automatic step 28, second field shunt.

 

The automatic steps were under the control of a current limit relay, whose setting was variable via an engineer’s control.  Automatic acceleration was needed to 'condense' many notches into few.  It was a well-established technique in EMU practice, where often just four controller notches were used.  So one could say that EMD expanded it to eight notches.

Thus counting both the manual and automatic steps, the DC control involved a total of 32 notches, of which four, namely 5 through 8, were running notches.  My guess is that the second field shunt corresponded to field shunting in the diesel-electric mode, with the first step included to minimize the magnitude of the notching current spike.  In the diesel-electric mode, the running back of the load regulator took care of that.

 

The FL9 case may be compared with that of the New Haven EP-5 class electric locomotive when operating in DC mode.  Therein it used 31 of the 35 controller notches, all manually switched, with just two being full-field running notches, field shunting not being used on DC.

 

 

Cheers,

  • Member since
    June 2022
  • 76 posts
Posted by Pneudyne on Friday, August 12, 2022 1:35 AM

Upthread I said:

 

“Dynamic braking – EMD field loop vs. GE potential wire – somewhat resolved with dual systems in the mid-1950s, probably led by railroad initiatives more than the locomotive builders, then finally addressed by EMD’s adoption of potential wire control in 1961, with retrofits available for older units.  Incidentally, this was never a problem in export markets, as EMD adopted potential wire control for export models from the start.  There compatibility did not seem to have been the primary driver; rather the field loop control was found to be functionally less suitable for the export models.”

 

An example of a railroad initiative was that of Western Maryland, which modified its Alco-GE units (fitted with potential wire dynamic brake control) and its EMD units (fitted with field loop dynamic brake control) so that they could work in mixed MU sets with either exercising full control of the dynamic braking of the other.

 

How it was done was described in an article in “Railway Locomotives and Cars” (RLC), 1957 April, p.68ff, “Western Maryland Ups Locomotive Utilization…Alco and EMD Units Run M-U”.

 

 

EMD’s adoption of potential wire DB control was recorded in RLC 1961 April, pp.40,41, “EMD Field Loop Can Be Eliminated”.  Details were provided as to how older EMD units could be converted from field loop to potential wire braking.

 

EMD’s earlier use of potential wire dynamic braking in export locomotives was cover by US patent 2745050 of 1956 May 08, filed 1952 January 02.

 

 

Evidently Fairbanks-Morse claimed to be the first locomotive builder to offer “universal” dynamic brake control as a factory fit.  This was recorded in the book “Train Master” by Diesel Era with David R. Sweetland, Withers 1997, on p.53.  It applied to the 1956 Train Master batch built for the Southern, fitted with GE Amplidyne equipment, and which were required to operate in MU with Southern’s existing EMD F7 fleet.

 

Something I haven’t been able to determine is when Alco first offered “universal” dynamic brake control as a factory option.  (It was available on the GE U25B from the start, as far as I know.)

 

 

Thus was the dynamic braking control incompatibility variously resolved on the pathway to a standard diesel MU system.

 

 

Cheers,

  • Member since
    January 2019
  • 1,686 posts
Posted by Erik_Mag on Wednesday, August 10, 2022 10:31 PM

Pneudyne

A more abbreviated description was provided in Railway Locomotives and Cars, 1958 September, p.58ff in an article “Diesel Boosters for Electric Locomotives”.

That issue is available for viewing and downloading from "The Internet Archive". In addition to that issue, the complete collection of "Railway Locomotives and Cars", "Railway Mechanical Engineer" and "American Railway Journal" are available for viewing and downloading.

The viewing option became a lot more palatable for me after getting fiber internet service as the new pages load almost instantly.

  • Member since
    June 2022
  • 76 posts
Posted by Pneudyne on Tuesday, August 9, 2022 2:38 AM

In my June 26 post, I mentioned “The Milwaukee “Wylie Throttle” that allowed leading DC electrics to control trailing diesel-electrics.  Good information on this is available.”

 

The main source is a 1959 April AIEE paper “Multiple-Unit Operation of Diesel and Electric Locomotives on the Milwaukee Road”, by Lawrence Wylie.

 

A more abbreviated description was provided in Railway Locomotives and Cars, 1958 September, p.58ff in an article “Diesel Boosters for Electric Locomotives”.

 

The scheme was simple but really quite ingenious.  Coupling between the electric and diesel control systems was entirely mechanical, with no electric interconnections.  And the link could be disconnected to allow independent control of the diesel if desired.  Supply current for the diesel control was obtained from the diesel locomotive itself.

 

The throttle coupling was by a rack-and-pinion device that provided a variable ratio by having an eccentrically mounted pinion.

 

The electric controller had 37 notches (plus separately control for two stages of field shunting), and the mapping to the 8 diesel notches was as follows:

 

Electric                         Diesel

 

1 - 3                             1

4 – 6                             2

7 – 9                             3

10 – 12                        4

13 – 15                        5

16,17                            6

18 – 23                        7

24 – 37                        8

 

(Electric notches 16 (series), 26 (series-parallel) and 37 (parallel) were the running positions.)

 

The objective was to have the trailing diesel-electric consist in notch 8 at about 15 mile/h, by which speed slipping was less likely.

 

As built, there was no control of trailing diesel dynamic braking, that that was noted as a future possibility.

 

 

This was an early example of mating two different control systems that were not originally designed to work together.

 

 

Cheers,

  • Member since
    March 2016
  • From: Burbank IL (near Clearing)
  • 13,540 posts
Posted by CSSHEGEWISCH on Tuesday, July 12, 2022 10:05 AM

Susie Q's numbering plan has continued to this day but it has become more of an anomaly than anything else since everything on the roster is MU-equipped anyway.

The daily commute is part of everyday life but I get two rides a day out of it. Paul
  • Member since
    February 2011
  • 117 posts
Posted by sandyhookken on Monday, July 11, 2022 4:01 PM

In the early days of dieselization, locomotives could be purchased with or without MU equipment (the non-MU option was probably cheaper).

One railroad that did this in the late 1940"s was the NYS&W. To easily determine which locomotives did or did not MU, the railroad used odd numbers on the non-MU locos, even numbers on the MU locos. The NYS&W continues this practice to this day. 

When the Delaware Otsego system purchased the NYS&W around 1980, they assigned an odd numbered locomotive (with MU) to the Little Ferry yard. The unit suddenly developed minor, and a few major, operating problems. The DO ended up transfering it to another one of their railroads, where it ran fine.

The NYS&W occasionally leases locomotives from CSX or NS. I've seen a few leased units with odd road numbers, but haven't heard anything yet about any "strange" problems. 

  • Member since
    June 2022
  • 76 posts
Posted by Pneudyne on Saturday, July 9, 2022 6:55 PM

 

I had wondered if the AAR sequence was chosen to simplify the solenoid changes required at notch transitions, although I had not made the connection with the Gray Code.  In any event, the A solenoid alternated between off and on for the odd and even notches respectively.  But the C solenoid was on from notch 3 through notch 8, which would not have been the case for a straight binary sequence, where it would be on only for notches 3 & 4, and 7 & 8.  Either way, the B solenoid was on from notch 4 through notch 8. 

Nonetheless, there were other railroad applications that did use the straight binary sequence. 

One example was the throttle control used in the Alco 539-engined switchers when equipped for MU.  These had a Woodward SI governor with external speed control and solenoid shutdown of the energize-to-run type.  The shutdown solenoid played no part in engine speed control.  The sequence was: 

Shutdown         no solenoids active

1st speed           no solenoids active

2nd speed          T1

3rd speed          T2

4th speed           T1, T2

5th speed           T3

6th speed           T1, T3

7th speed           T2, T3

8th speed           T1, T2, T3

 

I don’t have the engine speed schedule, but I suspect that the relative speed increments were, or were proximate to: T1 = 1, T2 = 2 and T3 = 4.

 

As I understand it, the governor operator was a version of the GE 17MK electropneumatic unit devoid of one of its customary four operating cylinders, but I have not been able to confirm this.

 

The 17MK3 operator, with all four cylinders, was used on the standard GE 70-tonner when fitted for MU.  The non-MU version had a mechanical throttle control.  Here a Woodward UG8 governor, with solenoid shutdown, of the energize-to-run type; and not used for engine speed control, was employed.  Seven-notch rather than eight-notch control was used.  Seven notches appear to have been an established GE norm for industrial switchers when the 70-tonner was introduced.

 

The sequence was:

 

Shutdown         no solenoids active

1st speed           no solenoids active

2nd speed          T1, T2

3rd speed          T1, T3

4th speed           T1, T2, T3

5th speed           T1, T4

6th speed           T3, T4

7th speed           T1, T2, T3, T4

 

I have no information about the associated engine speeds.  If T1 is set aside, then T2, T3 and T4 appear to form a normal binary sequence except that the T2+T4 step is missing, with a jump from T4 alone to T3+T4.  T1 might have provided a small speed increment, or perhaps a small decrement.  Which of those it was would have affected where 6th speed sat in relation to 5th and 7th speeds.

 

 

Cheers,

 

  • Member since
    December 2017
  • From: I've been everywhere, man
  • 4,269 posts
Posted by SD70Dude on Friday, July 8, 2022 1:43 PM

Erik_Mag

....where a slight error in positioning whatever encodes the individual bits would lead to short term spurious codes being sent when the encoder/controller was moved.

Newer units display the throttle position on the computer screen.  If you move the throttle halfway between notches on GEs the computer display will briefly show notch 1, even if you were between, say, 4 and 5.  

Greetings from Alberta

-an Articulate Malcontent

  • Member since
    December 2007
  • From: Georgia USA SW of Atlanta
  • 11,919 posts
Posted by blue streak 1 on Friday, July 8, 2022 2:41 AM

From posts on other sites there appears that the new Siemens ALCs and ACS-64s have MU issues due to fact Siemens does not use the relay based 27 pin operation. Uses some kind of voltage sensing that can be affected by moisture on the 27 pin connectors.

  • Member since
    January 2019
  • 1,686 posts
Posted by Erik_Mag on Friday, July 8, 2022 12:10 AM

The "odd" control sequences look a bit similar to a Gray Code, which strictly has only one bit changing between steps. This is a bit safer than straight binary when using a rotary encoder (throttle controller in this case), where a slight error in positioning whatever encodes the individual bits would lead to short term spurious codes being sent when the encoder/controller was moved.

  • Member since
    June 2022
  • 76 posts
Posted by Pneudyne on Thursday, July 7, 2022 12:13 AM

Both of Lemp’s basic load control approaches, namely the 1914 speed control with a load regulator rheostat, and the 1924 inherent characteristic type, in which by differential decompounding and other means the main generator curves were made to approximate the hyperbolic ideal, have been widely used in practice.  In some cases, elements of both have been used together.  The original speed control system had a single load setpoint, appropriate for the engines of the time with their relatively flat torque curves.  Later developments allowed multiple load setpoints, either proportional to engine speed or independent of it.

 

The inherent characteristic type was mostly used on lower-powered locomotives, say under 1000 hp, although in the 1950s both GE and Alsthom used it on some of their more powerful designs.  With this type, the key control parameter was engine speed, which is perhaps wherefrom came the concept that each throttle notch represented a specific engine speed, with a monotonic relationship.  Some of the systems derived from the 1914 patent, such as that from Brown Boveri, assigned more than one load setpoint to each engine speed, thus there might be for example four engine speeds but eight throttle notches.  Where each engine speed had its own load control setpoint that was proportional to speed, then a single throttle operator could be used to set both, engine speed directly and engine load via a floating lever.  So again the one engine speed for each throttle notch was a logical arrangement.

 

The choice of eight notches might not have been derived as a standalone answer to the question - how many steps are reasonably needed to have fine enough control - but rather because in the first instance that number was seen as the highest that could be implemented without undue complexity, and so was accepted as a reasonable trade-off point.  The eight steps could be provided by a three-unit binary operator, which in turn required three trainwires for speed control alone.  Once established by early use, the eight-notch control essentially became a fixture, regardless of the underlying details.

 

On the classification of actual excitation control systems according to Lemp type, one could say that EMD’s 1950s approach was essentially 1914, but with a hint of 1924 in that the main generator self-field winding provided some measure of decompounding.

 

GE’s Amplidyne system is more difficult to characterize.  It constructed highly drooping basic main generator curves, so in that sense there was a 1924 element.  But it also used load regulation (1914) to provide the necessary hyperbolic portions of these curves.  It appears that just six load control setpoints were used, though, notches 1 through 3 sharing the same setpoint.

 

GE’s three-field system, used on the Cooper-Bessemer engined export Universals of the 1950s (except the UD18B, which had the Static system), and as far as I know on the Alco DL541/543 export models, seems to have been basically a 1914 system with a significant element of 1924 through differential decompounding of the exciter.  The small export Universals with Caterpillar engines had a pure 1924 system.

 

On the other hand, the Alco export models of the 1950s and 1960s that were fitted with the 6-251 engine had what was basically an inherent characteristic system assisted by a load regulator rheostat, so might have been characterized as majority 1924 with some 1914.  Possibly the same system had been used on earlier 6-251-engined locomotives, such as the GE White Pass & Yukon shovel-noses and the Alco S-5 and S-5 6 switchers.  Earlier 539-engined Alco switchers I think had a 1924 system.

 

An earlier GE approach, used for example on the New Haven Alco-GE DL109, combined the differential exciter (Lemp 1924) with a centrifugal speed switch.  The latter moved to reduce excitation whenever the engine rotational speed reduced below its maximum due to the load, so in principle conformed to the 1914 system.  I suppose that it was a rudimentary form of chopper, effectively doing pulse width modulation of the supply to the exciter battery field.

 

The advent of electronics, such as the GE Type E system, allowed the construction of very close approximations to hyperbolic constant power main generator curves without the help of the customary load regulator.  In such cases the latter served simply to trim the curve, and to protect the engine from overloading in the event of a malfunction.  An interesting early application was in the UP GTEL8500s, in which a series of constant power hyperbolic main generator curves were constructed electronically.  Doing this by progressive load regulation as in the diesel case was not possible, although there was supervening exhaust temperature load regulation/limiting.  The earlier GTEL4500s had a set of “natural” (convex with respect to the origin) main generator curves, with only the supervening exhaust temperature load limiting providing a hyperbolic element.

 

I am not aware that relative power outputs, or relative power output ranges have been formally assigned to the notches in the AAR system.  Rather, whatever convention might exist, e.g. that notch 5 is around half-power, has more likely arisen from actual practice, with subsequent desire, in the interests of compatibility, not to depart too much therefrom.

 

Anyway, I think it is fair to say that the eight-notch control was a product of what was reasonably possible back in the early days of the electromechanical era.

 

One puzzle though is the standard AAR solenoid pattern for eight-step speed control, which is as follows:

 

Shutdown         D

1st speed           no solenoids active

2nd speed          A

3rd speed          C

4th speed           A, C

5th speed           B, C, D

6th speed           A, B, C, D

7th speed           B, C

8th speed           A, B, C

 

The respective relative speed increments are:

 

A = 1

B = 4

C = 2

D = negative 2 and shutdown

 

Thus it may be seen that at the 5th and 6th speeds, the B and D solenoids cancel each other out.   5th speed could be obtained by B alone, and 6th speed by the A and B combination.  And in fact that pattern was used in some non-US applications of the Woodward PG governor.

 

The inclusion of the D solenoid in the array for shutdown purposes was I think driven by the use of governors with rod shutdown, where the speed control rod was moved in the direction opposite to that for speed increase in order to effect shutdown.  Thus it could also be used to provide a speed decrement in association with the other solenoids, and when it was set not to exactly offset the increment provided by one of the latter.  For example, to obtain 15 speeds on the Woodward PG governor, the relative settings would be:  A = 2, B = 8, C = 4, and D = negative 1.  Apparently, the settings for each solenoid were variable over a wide range.

 

So why were the C and D solenoids, cancelling each other out, used together for the 5th and 6th speeds in the AAR system?  As the answer is nowhere to be found, I imagine that that whatever the reason was, it goes back a very long way.  Perhaps there was an early use of an electropneumatic speed setter where the eight speeds were nor evenly spaced, and C and D differed in magnitude?  And when eight evenly-spaced speeds were to be used, it just happened to work out that C and D had the same magnitude.

 

In that regard, I understand that Lima-Hamilton did use a different solenoid pattern, namely:

 

Shutdown         D

1st speed           no solenoids active

2nd speed          A

3rd speed          C

4th speed           A, C

5th speed           B

6th speed           A, B

7th speed           A, B, C, D

8th speed           A, B, C

 

I do not know the associated speed increments associated with that solenoid pattern, nor the respective engine speeds.  But assuming that the 7th speed was somewhere between the 6th and 8th speeds, then it is clear that the magnitude of D must have been less than the magnitude of C, as well as opposite in sign.  In this case compatibility was maintained by using a sequence converter between the master controller, which generated an AAR-type output, and the governor.

 

 

Cheers,

  • Member since
    January 2019
  • 1,686 posts
Posted by Erik_Mag on Saturday, June 25, 2022 9:34 PM

Supposed Herman Lemp quote: "To hell mit der volts, it's der amps dat count".

Trains had a series of articles ca 1979 on the Lemp system of control, basically wiring feedback windings on the generator to give it an approximation of constant power output for a given shaft speed.

Join our Community!

Our community is FREE to join. To participate you must either login or register for an account.

Search the Community

Newsletter Sign-Up

By signing up you may also receive occasional reader surveys and special offers from Trains magazine.Please view our privacy policy