• U23B nitpicking questions

  • Discussion of General Electric locomotive technology. Current official information can be found here: www.getransportation.com.
Discussion of General Electric locomotive technology. Current official information can be found here: www.getransportation.com.

Moderators: AMTK84, MEC407

  by Allen Hazen
"Extra 2200 South" had an article on the U23B and U23C by Bob Carman and Bill Peterson in issue#75. (Cover date April-May-June 1982, but since therre are news items and photos from June 82, it may have actually been published a bit later.)
In the "Spec Profile" it says that DC/DC U23B had 581 generators (specifically, GT581D or GT581F1) and that AC/DC units had GTA11 (specifically GTA11A1 or GTA11C1).
QUESTION: Were there any other variants?
For DC: The GT581 was catalogued for Alcos of 1600 to 2400 hp, so would be the obvious generator of choice for a 2250 hp 4-motor unit with an engine speed comparable to that of the Alco engines. A certain number of Alcos in the 2000 to 2400 hp range, however, were built with generators from trade-ins: 566 (from PA/PB, used in-- for example-- some LI C420 and NYC RS32) or 564 (from FA/FB-1 and RS-2, used in other NYC RS32 and in GB&W C424). I believe the GT581 was a derivative of the GT564, of-- I assume-- similar dimensions. So there should have been no TECHNICAL reason why GE couldn't have used it in some U23B, and a fair number of DC/DC U23B went to impecunious customers who may well have had early Alcos to trade in.
QUESTION: does anyone know of an application of the GT564 generator in a U23B?
For AC/DC: Early GE locomotives with AC/DC transmission used the (overengineered) GTA9. The cheaper GTA11 seems to have replaced it in production around 1969: EL's first order of U33C (built in 1968) had GTA9, and their second (built in 1969) had GTA11. Most AC/DC U23B were built in 1972 or later, but the first order-- C&O 2300-2329, GE construction numbers 37228-37257-- were built in September and October 1969. Probably GTA11, but
QUESTION: Does anyone know for sure?
((There are inaccuracies on this sort of thing in published railfan literature and internet rosters; I have a vague ambition of trying to sort the details out for GE models I am particularly fond of!))

  by Typewriters
There IS an answer at the end of this, but I have to qualify it first -- and this is a good tip for anyone who is attempting to collect or catalog old GE manuals.

Each manual will have a part number and a publishing date on the front page. This is not surprising; but what is surprising is that many manuals were reprinted later, after the original stock ran low -- and the reprints will have LATER dates while retaining EXACTLY the same content, word for word.

(Believe it. Right here in front of me are two totally identical manuals, which are GEJ-3866, Educational Manual for Model U33 Locomotives. One has a print date of 1-71 and the other 2-73. They're perfectly the same otherwise.)

Now that we know that, we should also remember that when changes are made to a manual, it receives a different part number. Again, right here are manuals GEJ-3849 and GEJ-3849A, both Educational Manual for Model U30 locomotives; the second one incorporates changes and improvements.)

What to we get from this? We can see that it is possible with some knowledge to figure out if a given manual in-hand is the original one for a type, or range, even if the printing date is later.

HAVING SAID ALL THAT, I note in response to your GTA-9 vs. GTA-11 question for the U23B that the Diesel Engine Service Manual GEJ-3869 which covers units from U23 to U36 lists GT-581 or GTA-11 for the U23B and GT-586 or GTA-11 for the U23C. I am of the opinion that this is the original (or first) diesel engine manual which incorporated information for the U36. (Print date on this example is 5-70.)

This does leave a gap between introduction of the U23 and U36 which I cannot specifically cover with the manuals at hand, even though they're a pretty extensive set. However, with that late date of construction for the C&O U23B units, I'd bet on GTA-11.

-Will Davis

  by Allen Hazen
Thank you!

(i) Further evidence that C&O 2300-2329 had GTA-11: according to the U33C article in "Diesel Era" v. 7, n. 1 (Jan-Feb 1996), Erie lackawanna's 1969 U33C -- the ones which had (source: E-L diagram book, on the WWWeb at George Elwood's "Fallen Flags" site) GTA-11 -- were built in August 1969, a month before the C&O locomotives. ... This isn't conclusive (GE might not have made a sharp break in their construction between GTA-9 and GTA-11 units, C&O might have said "Give us ALL the extra-cost options you can think of" (Grin!)), but if asked to bet, I know where i'd put my money.

(ii) At a guess, the GE manual you have gives the catalogued options. It seems to me that it would still have been POSSIBLE for a purchaser later (as of May 1970, the only U23B built were those of D&H, C&O and Monon (ATSF's first order started the next month)) to request a non-standard variant: so it's still POSSIBLE, though perhaps unlikely, that a few U23B were built with GT-564.
(iii) Interesting that GE offered the GTA-11 on the U23C: according to the "Extra 2200 South" article, all the U23C built -- both for U.S. and for Brazil -- had DC/DC transmission with GT-586 generators.
((I appreciate the notes on the dating and numbering of GE manuals. My father taught "Descriptive Bibliography" at the Columbia University Library School, and the application of information like that to solve scholarly problems was one of his academic specialties. In his case they were usually problems of literary history rather than railroad history, but the principle is the same.))
  by Allen Hazen
Since the last activity in this string...
One or another Brazilian railroad has rebuilt U23C with AC generators: somewhere on this forum there is a string that discusses that.
And now-- JUST now-- the WWWeb has a document answering the question about C&O's U23B, which I thought just might have had GTA9 generators, though it was much more likely that they had GTA11.
George Elwood, for his "Fallen Flags" rail image site (I haven't said this for a few days, and it deserves frequent reiteration: this site is a GREAT resource for anyone interested in railroad history!!!!!!), has scanned a 1969 Chessie System locomotive diagram book: not great images, but Chessie included LOTS of detail on the pages about particular classes: two pages of data. And, for the U23B, it confirms that they had the standard (for AC-equipped U23B) GTA-11 main generators. As we all assumed was most likely, but it's nice to have a document specifically confirming this.
(And thanks, again, Will Davis, for your earlier answer! Over the years you have given a lot of well-documented answers to detailed questions, for which I am very grateful.)
  by Pneudyne
Allen Hazen wrote:"I believe the GT581 was a derivative of the GT564, of-- I assume-- similar dimensions. So there should have been no TECHNICAL reason why GE couldn't have used it in some U23B, and a fair number of DC/DC U23B went to impecunious customers who may well have had early Alcos to trade in.
QUESTION: does anyone know of an application of the GT564 generator in a U23B?
The 1971 GE main generator list included the following entries under GT564:

GT564E1: Conversion of model GT564C1 to fit the Alco 251 diesel engine. This Includes different spider and frame machining, also the use of a GT581C1 fan.

GT564E2: Conversion of GT564C1 similar to form E1 except will use spur gearing to drive auxiliaries.

GT564F1: Conversion of GT564C1 similar to form E1 except for addition of air ducts, revised framehead, and various miscellaneous changes for 2400 hp operation.

That confirms that GT564 variants originally fitted to Alco 244 engines were later modified to suit Alco 251 engines.

Also from the GE list in respect of the GT581:

GT581C1: Same as model GT581A1 except designed to fit Alco 251 or Cooper-Bessemer FV-12 for 1800 hp locos. Can be used in place of model GT581A1 on Alco 244 engine with small modification.

GT581C2: Conversion of model GT581A1 to fit Alco 251 engine instead of Alco 244 engine. Interchangeable with model GT581C1.

From that two points may be inferred:

Firstly, that a generator designed to fit the Alco 251 engine would also fit the CB engine – and by extension, the FDL engine.

Secondly, that some GT581s were modified in service to fit Alco 251 engines.

The apparent interchangeability of GE main generators between Alco 251 and CG/GE engines indicates that GT564s modified to fit the Alco 251 could also have been used with CB/GE engines. So it was possible that some of the U23Bs might have been fitted with reworked GT564s.

On the other hand, that GE mentioned the CB engine, but not the FDL engine in connection with the modified GT564, suggests that what might have happened, and technically could have happened did not in fact happen. But that assumes that the documentation was fully accurate, and that is not guaranteed.

On balance, and as a reasonable working assumption, I’d say that the use of the reworked GT564 in the U23B did not happen, and that the best way to disprove this would be to find an actual example, as you originally requested.

The GT581 was the standard main generator fitting on the 12-cylinder export Universals, U18C (1956) through, U20C (1961) and for the additional models U22C and UM22C (mid-1970s), which ran in parallel with the U20C. But the larger U26C (1971) and U30C (1982) export models used the GTA11 alternator, as did the GE do Brasil version of the U22C, which in some ways was closer to the U26C than it was to the standard U22C.

Thus it was not so surprising that the GT581 was used on the domestic U23B. But whereas the GT581 could handle six GE761 motors, evidently six GE752s were two much which would explain the use of the higher capacity GT586 on the domestic U23C.

  by Typewriters
I can now confirm that the Lehigh Valley U23B units used the “constant speed” (actually three speed, usually) throttle control. This was stated clearly in a video, of the “Railfanning with the Bednars” series narrated by Mike Bednar who was an LV employee. He said that Conrail reset them to “normal” and the locomotives operated better that way.

Usually GE constant speed worked as follows: The engine ran at idle speed with the throttle in IDLE. When moved to RUN 1 however the engine came to what would normally be RUN 4 speed and stayed there with the throttle lever in RUN 1, 2, 3 or 4. Moving the throttle to RUN 5 sped the engine up to RUN 8 speed where it stayed until you were wide open.

This is how the U50C’s were set up as well.

-Will Davis
  by Allen Hazen
Good to hear from you!
... I will now ask idiot questions (to see how badly I misunderstand things). U-series GE units were unpopular with many locomotive crew. One complaint was that they loaded (increased power) very slowly (compared to EMD and Alco units). Am I right in thinking that the "constant speed" option was an attempt to address this issue? (Suppose you are running the locomotive at Run 1 after encountering a yellow signal, or at some low speed after crossing points or diamond, and you want to get back to normal train-pulling. A priori it seems to me likely that you'll get back to the power output of Run 4 if the engine's heavy rotating components are already rotating at the Run 4 rpm. And similarly in the higher band.)
And running the engine at much higher rpm than needed for a given power... would seem likely to increase wear and tear, so...
(Come to think of it, weren't some of Penn Centrals U23B at least initially set up with non-standard throttle control to make them load faster in switching? Was this the same system as used on LV's? And (i) did LV get its units with the "constant speed" option by telling GE "we'd like them set up like PC2750-2776... except for the long-hood ahead thing"? (ii) Did Conrail reset the throttles on its ex-PC units the same way they did on their ex-LV?)
  by Pneudyne
Allen Hazen wrote: Mon Jan 03, 2022 1:54 am And running the engine at much higher rpm than needed for a given power... would seem likely to increase wear and tear, so...
That certainly used to be a concern amongst the European locomotive engine builders, including English Electric (EE) and Sulzer. EE used to claim that its control system ensured that the engine never ran at a higher rotational speed than it needed to for the power required. In fact the engine stayed at minimum speed up to 30% load, thereafter steadily increasing to maximum speed at full load. (Basis information recently found, the Baldwin pneumatic throttle control did something similar.) In part the European concern was with fuel consumption, given the higher cost of fuel there. Another aspect was that full starting tractive effort could be obtained with the engine still at minimum speed.

In general though, the North American engine builders seem to have been much less concerned on this point. Indicative is the fact that in some cases, engines are often run at well above idle speed, even at full speed, during dynamic braking in order to provide adequate cooling air (from mechanically-driven blowers) for the traction motors. And needing to advance the engine speed to obtain full starting tractive effort seems not to be an issue.

The GE Type E electronic excitation control system probably helped facilitate greater diversity in the relationship between engine speed and power output. It electronically constructed close approximations (using a series of tangents) to constant power rectangular hyperbolae, such that for “normal” throttle notches, the load regulator was required to do only a final trim. I understand that the load regulator rheostat mostly stayed mostly in the 4 o’clock to 5 o’clock range. Logically, it was also possible to construct, for any given engine speed, lower power curves in which the load regulator simply stayed at its full position except in abnormal conditions. I imagine that this was what was done in the case of the 3-speed control. Another datapoint I have (uncorroborated, though) is that the New Zealand U26C used the same engine speed for notches 4 and 5.

Type E excitation, or at least an early version of it, was used on the UP GTEL8500 to construct a set of 20 constant power hyperbolae, given that a diesel-type load control system could not be used. It also occurs to me – although I have no knowledge that it was actually so - that Type E could have been used to improve GE’s 16-notch control system. One may deduce that the original version, with three-field excitation, would have had 16 basic main generator curves (convex to the origin), but only 8 constant-power hyperbolae, obtained by the usual load regulator action. Thus over part of the range some of the notch and half-notch positions would have coincided; perhaps not a big drawback given that it was the non-overlapping basic curves that dropped down on to the zero speed line to provide close graduation of starting and low-speed tractive effort. But with Type E, it would have been possible to have 16 constant-power hyperbolae, and so avoid overlaps.

(The European approach to a 16-notch control probably would have been to use 16 engine speeds. In Japan, it was done with 14 engine speeds, with the addition of two fixed-excitation notches at minimum speed.)

Where two adjacent notches share the same engine speed, then a fairly rapid upward power change should be possible, limited by electrical inertia (machine winding inductance, etc.). But a slower ramp than that actually possible might be preferred to limit the tractive effort jerk rate.

Perhaps somewhere in the GE literature there is a detailed explanation of the reasons for and benefits deriving from the use of three-speed control and other departures from the one engine speed-per-notch regime. But if so, it appears to be fairly well hidden.

  by Typewriters
There were a number of “constant speed” experiments here in the US in those days, for a variety of reasons - efficiency, response, wear being factors that were explored. Of course a few years later GE would slow down its FDL because efficiency became paramount and friction losses were enough higher at 1100 rpm than 1000 to warrant the speed reduction (once pistons were designed that could take the increased firing pressure and resultant stress.)

Of course on these fixed speed setups power change between notches at a given engine speed is by excitation only. Pneudyne is right on the Baldwins; there were Baldwins where the throttle advance (once the throttle switch picked up, either by air on a D-1 throttle or electrically on a CE-100 throttle) for the first quarter or so of travel simply acted (again by air) on the Carbonstat load regulator without the engine yet speeding up.

Will Davis
when i was running on conrail. i ran several of the conrail (ex-lv) u-boats. they had a switch on the control stand for switchng operations. when you flipped that switch on, the engine rev'ed up to run 4. this was used (in theory) to kick cars. those "switching" switches (as they were labelled) were still operative in the early 90's. over run 4 the throttle accelerated the diesel the same as any other engine, one notch at a time.
  by AllenHazen
Thank you! (One of the benefits of this site for amateurs like me is that we can receive information from professionals like you!)
Conrail was established in 1976. So... your experience with these locomotives was at least several years after that?
It sounds as if by the time you were running them, these locomotives could be run with the LV-specified "constant speed" set-up (probably better for switching, where rapid acceleration was desirable) or with the "normal" set-up, at the engineer's choice.
I suppose that this sort of thing would be routine on more modern units with microprocessor control, but it sounds as if Conrail and/or GE was able to make it work with old-fashioned "analogue" electronics!
  by Pneudyne
I suspect that that arrangement would have been doable in even the days of electromechanical control equipment. In the “switching” position, notches 1 through 3 effectively would have been of the constant excitation type, with corresponding main generator curves not “cut” by associated constant power hyperbolae. (Actually, the notch 4 constant power curve might have taken a small “bite” out of the notch 3 curve.) For switching purposes, I think that would have been fine. For example, some European and Japanese multinotch systems had the first few notches as constant excitation rather than constant power.

For general operation, as with the 3-engine speed control, It may have been preferred that all notches were of the constant power type. The Type E excitation system could have done that. This chart is illustrative.

from US3105186.PNG

It is from US patent 3105186 of 1963 September 24 (filed 1959 August 26) “Horsepower Limit Control and Function‘ Generator Therefor”.

The basic objective of the system was stated as follows:

“Accordingly, it is an object of my invention to provide a horsepower control system for a prime mover driving a generator wherein the output characteristic of the generator is shaped to provide accurate horsepower control of the prime mover.

“It is another object of my invention to provide a horsepower control system for a prime mover driving a generator, wherein the excitation of the generator is so controlled as to establish a horsepower characteristic for the generator which does not exceed the capacity of the prime mover.

“It is still another object of my invention to provide a horsepower control system for a prime mover driving a generator wherein an electrical network generates ‘a function closely approximating a ‘desired horsepower characteristic of the generator and controls the excitation of the generator therewith.

“It is a ‘further object of my invention to provide a novel electrical network which generates a close approximation of a desired horsepower output characteristic of a generator.”

In short, essentially GE wanted to construct “Lemp” curves by electrical means, rather than relying wholly on feedback from the engine via the governor.

The “Lemp” curves were constructed in series of tangents. In the chart, the tangents DE and EF are shown as approximating the adjacent/overlapping constant power curve B, although in practice, a multiplicity of tangents were used.

It was also said that the locomotive operator may select a characteristic such as B, H or I of by varying the reference current by means of the tap on a control potentiometer.

Thus multiple constant power curves (or at least very close approximations thereto) wee available at any given engine speed,

With conventional load control relying on governor feedback via floating lever and pilot valve, it was not so easy to provide more than one power setting for each of several speed settings, although it could be done. Sulzer showed a version of its 1939 eight-speed governor in which the pilot valve sleeve had two positions, electrically controlled, thus allowing two power settings for each speed, although I can’t trace that it was ever used as such. The Regulateurs 1100 governor was also available with a controlled movable pilot valve sleeve. But GE’s electrical/electronic approach looked a lot neater. GE used pulse width modulation (PWM), which is sometimes describe as obtaining an analog result by digital means.

You do not have the required permissions to view the files attached to this post.