GE U50 and GTEL informations

I did though find the GE patent for the floating bolster truck, pertinent here because that was what was used under the GTEL8500.


Cheers,

Pneudyne wrote: ↑Sat Mar 06, 2021 7:14 pm The history of triple-truck locomotives is quite diverse. Westinghouse patented a single-frame, articulated truck, B+B+B layout in the early 1920s.

But GE had at least one earlier patent for single-frame, triple-truck locomotives, namely US 1160715 of 1915 November 16.
It also had US 1511064 of 1924 October 07.

So the idea goes back to the early history of electric locomotives.

Cheers,

Regarding the disposition of the span-bolsters and trucks from the UP GTEL4500 fleet, as discussed upthread, this excerpt from an article on the Alco C855 in “Railway Locomotives and Cars” 1964 July indicates that all of the UP Alco C855 and GE U50 fleets were recipients.





From that we may deduce that the SP U50 fleet had new-build span-bolsters and trucks. Given that GE specified the same running gear for its catalogued, but never-built U56 model, evidently there was no problem with sourcing new span-bolster assembles.





I still have not found any patent related to the GTEL4500 running gear, nor “hard” confirmation of its presumed supplier, GSC.


Cheers,

Pneudyne wrote: ↑Tue Jan 26, 2021 9:58 pm The GE GTEL4500 prototype appears to have been the first to use independent span bolsters, that is, not connected by an articulation joint.

Maybe not. Photographs available on the web of the Illinois Terminal C class units, built from late 1924, suggest that their span bolsters were not interconnected. That needs more research to confirm or deny as the case may be.

On the other hand, The Piedmont & Northern 100-ton unit described in Railway Age, 1925 January 17 did have articulated span bolsters.

So perhaps the articulated and non-articulated variants of the span bolster four-truck running gear appeared at about the same time.

That they were established types might explain why there were no apparent GE patents associated with its later 1940s used of both forms of the span-bolster running gear, articulated for the VGN EL-2B and non-articulated for the GTEL4500. In both cases, GE did not seem have made any significant changes to the basic concepts.


Cheers,

Some information on the running gear for the GTEL 4500 was provided in a 1949 AIEE paper, “The Alco-GE 4,500-Horsepower Gas-Turbine Electric Locomotive”, by A.H. Morey of GE.

Firstly, from the general description:

“The Alco-GE gas-turbine electric locomotive, shown in Figure 1, is designed for freight service. It is an 8-axle 8-motor B-B-B-B locomotive, weighing 253 tons (average) and rated at 4,500 horsepower for traction (at 80 degrees Fahrenheit and 1,500-feet elevation). The locomotive is 83-feet 71/2-inches long, and 14-feet 3-inches high over the roof sheets. It will negotiate curves of 288-feet radius.”

Secondly, from the running gear description:

“The trucks are of the conventional 2-axle swing-bolster type, as indicated in Figure 3. Each pair of trucks is connected by a span bolster which, in addition to spanning the two trucks and supporting the locomotive cab, also acts as a traction-motor air duct. This arrangement gives a high degree of flexibility on curves. It also has sufficient vertical flexibility to give satisfactory riding qualities even on comparatively rough track. Resonant vibration between the truck equalizers and the swing bolsters was encountered at low speeds. Damping means were provided in the spring system and no further trouble has been encountered.”

No detailed rationale was provided, one being left to infer that once the desired power output had been chosen, 8 axles were deemed necessary both to accommodate the required machinery, and deliver the requisite power to the rails using a standard traction motor. And that in turn, the chosen running gear was seen as the best way to do arrange the required 8 axles.

Some comments on the in-service characteristics and performance of the GTEL 4500 were provided in an article in “Diesel Railway Traction” for 1954 February. This was based upon a 1953 AIEE paper that I have yet to track down. Here is a pertinent excerpt:

“The 4,500 h.p., eight-motor, 233-ton design has proved to be satisfactory. There was much concern at the beginning over the high h.p. per axle and the high per axle deigned into this locomotive. By comparison, it has the same horsepower as a three-unit diesel-electric, yet it has only two-thirds as much weight and two-thirds as many motors. The h.p. per axle may be as high as 675, compared to a maximum or about 400 on the diesel-electric freight units. The weight per axle goes as high as 72,000 lb as compared to a maximum of about 60,000 lb for the diesels. However, these fears have proved unnecessary in that both the high h.p. ratio and the high axle weight have been quite satisfactory. It was feared that the high h.p. per axle would lead to excessive slipping of the wheels. This has not developed. In fact, the locomotive has developed just the opposite reputation for being anything but slippery. The high axle weight has not caused trouble. In regard to wheel wear and flange wear, the performance has been exemplary.”

Regardless of its origins, the UP was evidently happy with the span-bolster running gear, basis its re-use on its later GE U50 and Alco C855 diesel-electric fleets.

For the U50, GE simply replicated the GTEL 4500 running gear spacing, with 41’6” between span bolster centres. The Alco C855 was a longer locomotive, with 44’0½” span bolster centres.

The respective CTE and MCS numbers may be compared:

GTEL 4500 4500 hp 105 000 lbf at 12.9 mile/h
U50 5000 hp 106 000 lbf at 14.7 mile/h
C855 5500 hp 106 000 lbf at 16.6 mile/h

As far as I know, all had GE752 motors with 74:18 gearing and 40-inch wheels. Thus from the CTE/MCS numbers one could infer that, to a first approximation, the traction motor maximum continuous current was similar for all three, with the MCS then more-or-less increasing proportionally with power output.

Now to the wheel arrangement designation. Originally, and as shown in the above-mentioned 1949 AIEE paper, the GTEL 4500 wheel arrangement was described as “B-B-B-B”. The same designation was used by Railway Age magazine in respect of the Alco C855 and GE U50 in its 1963 October 07,14 issue, pertinent page attached. But later, the designation “B+B-B+B” seems to have been used. Was this a formal change by AAR, or did it have some other origin, such as from the UP, or from the “railfan world”. I am much more familiar with the British Commonwealth system, in which these locomotives have always been described as “Bo-Bo-Bo-Bo”. In that system, and within the context of a single frame, the “+” sign is reserved for articulated truck locomotives in which all buff and drag forces go through the running gear, completely bypassing the main frame. Specifically “+” refers to articulated couplings between adjacent trucks (or adjacent span bolsters) which communicate all buff and drag forces. It looks as if the same guideline was once applied by the AAR, but that that may no longer be the case. Nonetheless it would seem that the AAR designation for say a single unit of the VGN EL-2B electric locomotive was, and still is described as “B-B+B-B”, the “+” signifying that the two span bolsters are coupled by an articulated joint.


Cheers,

Regarding the origins of span bolster running gear, as suspected the Illinois Terminal C class locomotives did not have connected span bolsters (e.g. see: https://www.flickr.com/photos/rocbolt/48597697277). Thus they appear to have been the first of the span bolster B-B-B-B type. I imagine that the non-connection was to allow for maximum flexibility in very tight street trackage curves.

Piedmont & Northern (P&N) does look to have been the originator of the articulated B-B+B-B version of span bolster running gear, also subsequently used by some other interurbans. The earliest reliable reference I can find is the previously mentioned Railway Age 1925 January 17 article, which had a picture of #5601. One may find other internet references to P&N span bolster locomotives with earlier dates (e.g. https://www.wvncrails.org/uploads/5/8/1 ... 059985.jpg), but whether these are accurately dated is unknown. Later, GE built at least one example for P&N, #5611 described in Railway Age 1942 April 04 p.705.

Outside of the interurban sphere, first use of the articulated span-bolster B-B+B-B wheel arrangement appears to have been for the IC EMC/GED/St. Louis diesel-electric centre-cab transfer locomotive of 1936. Then followed the GE-built Virginian EL-2B class motor-generator electrics in 1948, which looks to have been the last application.

Again outside of the interurban sphere, the independent span-bolster B-B-B-B form was next used for the Alco-GE GTEL4500 prototype, then the GE production GTEL4500, and then the GE U50 and Alco C855, in all cases with couplers mounted on the span-bolster outer ends, and the span bolster pivots taking the full buff and drag shear forces. Around 1969-70 came the CEM (French) monomoteur B-B-B-B (4B, as it was called) design, along with a B-B-B (3B) with a span bolster unit at one end and a single B truck at the other (really!- it was the oddest form in the tribo world). In this case couplers were mounted on the mainframe. Then followed GE, and later EMD (neat low profile design) with export B-B-B-B running gear for some of their respective diesel-electric models, coupler mounted on the mainframes.

For completeness perhaps the solitary Baldwin-Westinghouse C-C-C-C steam turbine electric unit for the N&W might be mentioned. The span bolsters were independent, but carried the couplers. And amongst never built proposals may be mentioned the Brown Boveri GTEL proposal – for US use – of an A1A-B-B-A1A unit. And the Westinghouse proposal for a unified set of standard electric locomotives included articulated span bolter types with B-B-B+B-B-B and B-B-B-B+B-B-B-B wheel arrangements. Also in the range was a B-B-B-B type with four independent lateral motion trucks and no span bolsters, which arrangement was used for its solitary GTEL prototype.

I have not found any patents relating to early span bolster running gears. Such were also used for very large freight cars, but I do not know whether these predated or post-dated their use in interurban locomotives.

Regarding electric locomotives, I suspect that where eight axles were required, it was found easier and more economical to use to a pair of single-cab B-B units coupled (semi)-permanently back-to-back, hence the lack of span-bolster units in the 1950s and onwards. The GE writings of the early 1950s particularly about the development of the E2b prototypes for the PRR, certainly point in this direction, and against designs such as the GN W-1 and VGN EL-2B. (it was intended that the E2b would also be built in DC and MG versions if required).

Previously noted, but worth restating in context, is that the NYC T class B-B+B-B electrics of 1913 and up were not of the span bolster type, as the inner axle pairs were rigidly mounted to their respective beams, with the outer trucks then acting as pilots. Apparently NYC wanted an improved version of its S-3 class 2-D-2 rigid-frame units with powered pilot trucks for higher adhesive weight, and with those pilot trucks spaced further from the main wheelbase for better stability. The latter would have been difficult with the 2-D-2 arrangement. The answer was to split the D wheelbase into two separate but articulated B sections, which then allowed the use well-spaced pilot trucks. Interesting though was that much later GE considered B-D-B rigid-frame running gear for the VGN El-2B, before landing on the span-bolster B-B+B-B form, which allowed the use of four identical trucks.

Against all of that, it may be seen that the GTEL4500 prototype represented the introduction of the independent span-bolster B-B-B-B wheel arrangement into the heavy-duty line service locomotive arena, and that wheel arrangement remains in the inventory for some export applications.


Cheers,

As previously mentioned, the UP GTELs all had 20-notch power control. GE reasons for this were given thus:

“Control of power plant speed and generator power demand (locomotive power) is provided in 20 steps on the controller. This number of throttle notches was selected, instead of the customary eight, to give finer and better control of tractive effort as well as better locomotive performance when working near limiting adhesion. The throttle handle operates a potentiometer which feeds a d-c signal into the excitation system of all four generators. By the proper selection of taps on the potentiometer, equal increments of tractive effort are obtained.”

The outcome is shown in the main generator curve chart for the GTEL4500 prototype, which shows evenly spaced tractive efforts along the standstill (IR) line, with supervening load control imposing a constant power hyperbolic curve only at the full-power end.



The GTEL4500 prototype and the first production batch had Amplidyne excitation control, derived from established GE diesel-electric practice. The second batch had static excitation, although I imagine that the curves were similar.

With the GTEL8500, a change was made to what was an early form of electronic excitation. This allowed the construction of hyperbolic constant power curves of varying power level throughout the range.





The basic means by which this was achieved is shown in this diagram. Note that as with the GTEL4500, throttle control was by a (stepped) variable voltage input to the excitation control circuitry.





Cheers,

Continuing from the previous post, this approach to excitation control appears to have been covered by US patent 3105186:



It also looks like the forerunner of the system later used on Alco (Type E) and GE diesel-electric locomotives.

As far as I know, dynamic braking control on the GTELs was of the potential wire type, continuously variable and operated by the selector handle.

Most of the GTELs were eventually fitted with equipment that allowed them to control trailing diesel-electric locomotives in MU. This appears to have been a UP initiative, although probably with GE knowledge and possibly involvement. I have not found any details about it, though. Probably some of the existing GTEL master controller trainline outputs could have been used, such as those for forward, reverse, generator field and dynamic brake. But the AV, BV, CV and DV diesel throttle control solenoid trainlines would need to have been added, perhaps by fitting additional cams and contactors within the master controllers. There would need to have been appropriate mapping from 20 GTEL to eight diesel notches. My guess is that diesel notch 8 would have been reached ahead of GTEL notch 20. Firstly, under some conditions (ambient and/or mechanical), full available GTEL power could be reached before notch 20, and it was probably undesirable to also limit trailing diesels to less than notch 8. Secondly, the normal acceleration patterns may have been different, with the diesels getting into notch 8 at lower speed than the GTELs would reach notch 20. Maybe the diesel control reached notch 8 at around notch 15 or 16 of the GTEL control. There was a precedent there with the Milwaukee “Wylie throttle” that allowed the EF-4 class “Little Joe” electrics to control trailing diesels. There the diesel control reached notch 8 when the electric controller was only at notch 24 of 37, reckoned to be at around 15 mile/h at which speed the diesels would be expected not to slip in notch 8. The Wylie throttle coupling system was purely mechanical, and non-linear. Control current for the trailing diesels came from the diesel. I’d guess that in the UP system, there was not a problem with diesel control current coming from the GTEL auxiliary electric system.


Moving from the GTELs to the U50, insofar as the latter was in many ways a double U25B, then one would expect it to have been fitted with the same 16-notch throttle control. But given that the U50 was more-or-less the diesel counterpart to the GTEL4500, it probably wanted a multinotch control anyway. The fact that the same was specified for the Alco C855, and not otherwise used by Alco, was indicative.

The 16-notch control used the GE KC99 master controller, which was an interesting unit. Without any prior knowledge of its workings, one might expect to find that as well as the usual power control trainline outputs, FO (forward), RE (reverse), GF (main generator field), AV (governor A solenoid), BV (B solenoid), CV (C solenoid) and DV (D solenoid), there was an additional trainline that was activated on the half-notch steps.

But from what I can piece together from various sources – not necessarily correctly so – that was not the case. Also, the fact that notches ½, 1 and 1½ all shared the same (minimum) engine speed was a complication that would appear not to have been addressed by a single extra trainline.

Apparently there was an extra trainline, labelled SN, activated in all power notches ½ through 8. What this did was to operate what was known as the MR relay, which “told” any 16-notch units in a consist to operate in 16-notch mode. Absent a closed MR relay, they defaulted to eight-notch mode.

Also, the KC99 contained a large rheostat that was switched in 16 steps and which provided an output to the XB (dynamic brake excitation) trainwire, with the output voltage increasing stepwise from notch ½ through 8. This applied during motoring as well as during dynamic braking. And the XB trainwire was the source for exciter battery field current during motoring as well as during dynamic braking, not only for the leading unit but also for any 16-notch trailing units in the consist. The lead unit supplied excitation current for all 16-notch trailing units. Non-16 notch trailing units in the consist simply ignored the XB trainwire during motoring, and “did their own thing” according to the customary trainline signals.

Overall, that gave three notches (½, 1, 1½) at first (minimum) engine speed, two notches (2, 2½, etc,) at second through seventh engine speeds, and one (8) at eight (maximum) engine speed. One imagines that the notches were arranged to provide relatively closely spaced tractive effort increments along the standstill line, following the GTEL example.

When a 16-notch unit was trailing in a consist led by a regular 8-notch unit, the SN trainline was dead, so the MR relay did not operate, and it defaulted to eight-notch mode. Here motoring excitation was apparently controlled locally by a resistor ladder switched by the AV, BV, CV and DV trainwires. (Whether this ladder was integrated with or separate from the KC99 internal rheostat I don’t know). In dynamic braking, excitation was controlled in a continuously variable way by the load regulator rheostat, in turn operated by a positioner relay controlled by the XB trainwire.

With the three-field system as assumed to have been used for the U50, I think that there would still be just eight hyperbolic constant power curves as determined by the governor and load regulator action. So adjacent notches would share part of the same curve over some of the speed range, although not at lower speeds.

But the Type E excitation system used in the Alco C855 had the ability to construct electrically/electronically close approximations to constant power curves, as shown in the GTEL8500 case. Here then, for normal curves representing the maximum desired power at each engine speed, the load regulator rheostat served mostly as a trimming device, and also as a supervening control in abnormal conditions. But it was also possible to construct constant power curves of less-then-maximum power at any engine speed. If that facility had been used in the Alco C855 case – and whether it was done is unknown - then there could have been 16 constant power curves as well as 16 basic excitation levels. In that case, the XB trainline feed from the KC99 master controller rheostat probably provided the appropriate control voltage input to the excitation system.

The later GE U50C had regular eight-notch throttle control, despite its very high power-to-adhesive weight ratio, higher than any of the GTELs. By this time, improved wheel slip control systems meant that minimizing the notching up tractive effort jump was less of a concern, and there may have been some controlled current ramping at upward notch changes (that I don’t know). As I understand it, the U50C derived eight constant power curves from just three engine speeds, using the capability of the electronic excitation system to construct such curves without help from the load regulator.

Some of the foregoing commentary is based upon dependable source materials, but some, in the absence of such sources, derives from deduction and speculation on my part, so could well be wrong.

Nonetheless, it may be said the subject locomotives do not lack for interesting and unusual features, the running gears and control systems particularly.


Cheers,

Pneudyne wrote: ↑Wed Feb 08, 2023 11:26 pm
But the Type E excitation system used in the Alco C855 had the ability to construct electrically/electronically close approximations to constant power curves, as shown in the GTEL8500 case. Here then, for normal curves representing the maximum desired power at each engine speed, the load regulator rheostat served mostly as a trimming device, and also as a supervening control in abnormal conditions. But it was also possible to construct constant power curves of less-then-maximum power at any engine speed. If that facility had been used in the Alco C855 case – and whether it was done is unknown - then there could have been 16 constant power curves as well as 16 basic excitation levels. In that case, the XB trainline feed from the KC99 master controller rheostat probably provided the appropriate control voltage input to the excitation system.

Well, in any event, in the Type E case, the KC99 master controller rheostat output would simply have provided a control input to the excitation system, with the latter supplying the actual excitation current, in contrasts to the three-field case, where that current came from the rheostat.


Cheers,

Pneudyne wrote: ↑Wed Feb 08, 2023 11:26 pm Continuing from the previous post, this approach to excitation control appears to have been covered by US patent 3105186:


US3105186 19590826 GE Power Limit & Function Generator p.01.png



It also looks like the forerunner of the system later used on Alco (Type E) and GE diesel-electric locomotives.

As far as I know, dynamic braking control on the GTELs was of the potential wire type, continuously variable and operated by the selector handle.

Most of the GTELs were eventually fitted with equipment that allowed them to control trailing diesel-electric locomotives in MU. This appears to have been a UP initiative, although probably with GE knowledge and possibly involvement. I have not found any details about it, though. Probably some of the existing GTEL master controller trainline outputs could have been used, such as those for forward, reverse, generator field and dynamic brake. But the AV, BV, CV and DV diesel throttle control solenoid trainlines would need to have been added, perhaps by fitting additional cams and contactors within the master controllers. There would need to have been appropriate mapping from 20 GTEL to eight diesel notches. My guess is that diesel notch 8 would have been reached ahead of GTEL notch 20. Firstly, under some conditions (ambient and/or mechanical), full available GTEL power could be reached before notch 20, and it was probably undesirable to also limit trailing diesels to less than notch 8. Secondly, the normal acceleration patterns may have been different, with the diesels getting into notch 8 at lower speed than the GTELs would reach notch 20. Maybe the diesel control reached notch 8 at around notch 15 or 16 of the GTEL control. There was a precedent there with the Milwaukee “Wylie throttle” that allowed the EF-4 class “Little Joe” electrics to control trailing diesels. There the diesel control reached notch 8 when the electric controller was only at notch 24 of 37, reckoned to be at around 15 mile/h at which speed the diesels would be expected not to slip in notch 8. The Wylie throttle coupling system was purely mechanical, and non-linear. Control current for the trailing diesels came from the diesel. I’d guess that in the UP system, there was not a problem with diesel control current coming from the GTEL auxiliary electric system.


Moving from the GTELs to the U50, insofar as the latter was in many ways a double U25B, then one would expect it to have been fitted with the same 16-notch throttle control. But given that the U50 was more-or-less the diesel counterpart to the GTEL4500, it probably wanted a multinotch control anyway. The fact that the same was specified for the Alco C855, and not otherwise used by Alco, was indicative.

The 16-notch control used the GE KC99 master controller, which was an interesting unit. Without any prior knowledge of its workings, one might expect to find that as well as the usual power control trainline outputs, FO (forward), RE (reverse), GF (main generator field), AV (governor A solenoid), BV (B solenoid), CV (C solenoid) and DV (D solenoid), there was an additional trainline that was activated on the half-notch steps.

But from what I can piece together from various sources – not necessarily correctly so – that was not the case. Also, the fact that notches ½, 1 and 1½ all shared the same (minimum) engine speed was a complication that would appear not to have been addressed by a single extra trainline.

Apparently there was an extra trainline, labelled SN, activated in all power notches ½ through 8. What this did was to operate what was known as the MR relay, which “told” any 16-notch units in a consist to operate in 16-notch mode. Absent a closed MR relay, they defaulted to eight-notch mode.

Also, the KC99 contained a large rheostat that was switched in 16 steps and which provided an output to the XB (dynamic brake excitation) trainwire, with the output voltage increasing stepwise from notch ½ through 8. This applied during motoring as well as during dynamic braking. And the XB trainwire was the source for exciter battery field current during motoring as well as during dynamic braking, not only for the leading unit but also for any 16-notch trailing units in the consist. The lead unit supplied excitation current for all 16-notch trailing units. Non-16 notch trailing units in the consist simply ignored the XB trainwire during motoring, and “did their own thing” according to the customary trainline signals.

Overall, that gave three notches (½, 1, 1½) at first (minimum) engine speed, two notches (2, 2½, etc,) at second through seventh engine speeds, and one (8) at eight (maximum) engine speed. One imagines that the notches were arranged to provide relatively closely spaced tractive effort increments along the standstill line, following the GTEL example.

When a 16-notch unit was trailing in a consist led by a regular 8-notch unit, the SN trainline was dead, so the MR relay did not operate, and it defaulted to eight-notch mode. Here motoring excitation was apparently controlled locally by a resistor ladder switched by the AV, BV, CV and DV trainwires. (Whether this ladder was integrated with or separate from the KC99 internal rheostat I don’t know). In dynamic braking, excitation was controlled in a continuously variable way by the load regulator rheostat, in turn operated by a positioner relay controlled by the XB trainwire.

With the three-field system as assumed to have been used for the U50, I think that there would still be just eight hyperbolic constant power curves as determined by the governor and load regulator action. So adjacent notches would share part of the same curve over some of the speed range, although not at lower speeds.

But the Type E excitation system used in the Alco C855 had the ability to construct electrically/electronically close approximations to constant power curves, as shown in the GTEL8500 case. Here then, for normal curves representing the maximum desired power at each engine speed, the load regulator rheostat served mostly as a trimming device, and also as a supervening control in abnormal conditions. But it was also possible to construct constant power curves of less-then-maximum power at any engine speed. If that facility had been used in the Alco C855 case – and whether it was done is unknown - then there could have been 16 constant power curves as well as 16 basic excitation levels. In that case, the XB trainline feed from the KC99 master controller rheostat probably provided the appropriate control voltage input to the excitation system.

The later GE U50C had regular eight-notch throttle control, despite its very high power-to-adhesive weight ratio, higher than any of the GTELs. By this time, improved wheel slip control systems meant that minimizing the notching up tractive effort jump was less of a concern, and there may have been some controlled current ramping at upward notch changes (that I don’t know). As I understand it, the U50C derived eight constant power curves from just three engine speeds, using the capability of the electronic excitation system to construct such curves without help from the load regulator.

Some of the foregoing commentary is based upon dependable source materials, but some, in the absence of such sources, derives from deduction and speculation on my part, so could well be wrong.

Nonetheless, it may be said the subject locomotives do not lack for interesting and unusual features, the running gears and control systems particularly.


Cheers,

I think you have this all correct. This information is covered in the early GE manuals, such as the Educational Manual for the U25B and later U25B and C in one manual.

I have the U25B Educational Manual, GEJ-3815, of 1962 February, and that is the foundation for many of my comments and deductions. It certainly lays out the basics, but there are some gaps. For example, nothing substantial is said about motoring excitation control when a U25B was trailing a non-GE locomotive. There is though one somewhat cryptic paragraph: “The DR relay is picked up whenever the DV solenoid governor valve is energized. It has two functions in locomotive operation. It returns engine speed to idle if the ground relay trips when throttle is in notches 5 through 6½. Its other function is to assist in controlling excitation whenever trailing behind a non-U25B locomotive.”

Clearly, in that trail condition, motoring excitation control must be local, as the XB trainline would be “dead”. We may obtain a clue from the three-field excitation used in some of the export Universal models of the time, namely the U9, U12 and U18, the schematic for which was:


That shows that motoring excitation was controlled by a simple resistor ladder switched directly by the AV, BV and CV throttle control trainlines, and indirectly by the DV trainline via the DR relay. A reasonable assumption is that the U25B had a broadly similar setup to control excitation when it was in “trailing non-GE” mode. Speculation on my part, but possibly in this mode the eight excitation steps did not exactly overlay on any of the 16 steps in leading mode, except perhaps at notch 8.

The dynamic braking control schematic for the subject export Universals was thus:



That form of dynamic braking control was covered by US patent 2879462 of 1959 March 24, filed 1955 March 21. The GE U25B primary system was essentially similar, except that instead of a continuously variable braking rheostat controlled by the handle lever, a stepped rheostat controlled by the throttle handle was used.

That the U25B used a different control system, namely a load regulator positioner relay, when in “trailing non-GE” mode, is rather curious. Perhaps it was done in deference to the early Alco-GE units when such were leading. Their potential wire DB control systems handled currents that were probably in the milliamp range. On the other hand, where excitation current was provided from a leading GE unit, the required currents would have been higher, say up to around 0.5 amp per unit, maybe more. This may have been beyond early potential wire capability. Hence the use of the load regulator positioner, which would have quite a small current draw. Later Alco units may have had DB control rheostats with higher current capability; such would have been needed anyway with the introduction of dual systems with field-loop control capability.


I also have the GE U33 Educational Manual, GEJ-3866, of 1969 July. This shows a 16-notch throttle handle, also used for dynamic brake control. The workings of the electronic control system are well described, including its ability to construct close facsimiles of constant power curves without the help of the load regulator rheostat. It is not explicitly stated that “intermediate” curves were so constructed for the half-notches, but this is a reasonable assumption on the basis that it was easily done and would have been logical. In this case, excitation currents were always developed locally, with the trainlines serving only as control inputs. But the U33 XB trainwire output surely would have been capable of delivering the excitation currents required by the U25B and like locomotives, including the U50.


Cheers,

This group of locomotives forms an interesting progression in terms of control and excitation systems:


The GTEL4500 prototype and first production batch used a 20-notch version of the GE Amplidyne excitation system, originally developed in eight-notch form for Alco-GE diesel-electric locomotives (1).

The GTEL4500 second production batch used a 20-notch version of GE’s Static excitation system, whose major early use in eight-notch form, beyond GE prototypes and small production batches, was in Alco diesel-electric locomotives.

The GTEL 8500 used GE’s “electronic” excitation system in a 20-notch form.

The U50 used the 16-notch version of GE’s three-field excitation system that had been introduced with the U25B, that system being derived from the basic eight-notch three-field system used on the larger GE export Universal models from 1956, and also by Alco on some export models, e.g. the DL541.

The Alco C855 used a 16-notch version of GE’s Type E excitation system, which followed the same basic principles as what was used on the GTEL8500.

The U50C used GE’s “electronic” excitation system in eight-notch form, but with fewer than eight engine speeds.


The U50 and C855 may be viewed as somewhat more powerful (roughly 10 to 20%) diesel-electric counterparts to the GTEL4500, with identical running gear.

The U50C was a somewhat more powerful diesel counterpart to “half” of a GTEL8500, again with identical running gear.


A curious facet is that the GTEL4500 prototype was an Alco-GE product, albeit conceived by GE and built at Erie. But the first production batch, although mostly built in the Alco-GE era, was evidently strictly a GE product.

Possibly the Alco-GE agreement did not cover gas turbine locomotives, so that GE had a choice of allowing Alco to participate or not. Electric locomotives also appear to have been outside of the agreement, as were export diesel locomotives – even those fitted with Alco 244 engines, of which quite a few were built within the Alco-GE era. Curiously, though, the Alco DL500 export model, announced just before the end of the Alco-GE agreement, was initially billed as an Alco-GE model.


(1) The Amplidyne system and its associated controls was described in a cluster of AIEE papers of 1947 January, namely:

47-37 “Developments in Diesel-Electric Traction-Generator Excitation Systems”, by C.A. Brancke and G.M. Adams.
47-38 “Braking Resistors and Control for Diesel-Electric Locomotives”, by E.F. Weiser.
47-39 “A Power Plant Regulating System for Diesel-Electric Locomotives”, by C.B. Lewis.
47-40 “Developments in Control Systems for Diesel-Electric Locomotives”, by M.D. Henshaw.

From the first of those papers:

“Almost all Diesel engines can operate successfully over a 3-to-1 speed range. Advantage is taken of this variable speed characteristic to obtain variable power output easily and economically. Standard practice for several years by most manufacturers of Diesel-electric road locomotives has established the use of eight throttle handle notches for obtaining eight different engine operating speeds on units which are operated in multiple. The amplidyne excitation control circuits therefore were designed on this same basis in anticipation of probable requests by ultimate customers to operate in multiple with existing locomotives.”

And from the third:

“The combined speed and fuel limit setting resistors and commutating relays - Four relays are used to commutate resistance in the regulator circuits to set the speed and the fuel limit in a manner to give eight steps of engine power. While it is possible to provide the regulator with the proper signal for either a step or a modulated control in a variety of ways, the relay scheme is the one that was first used on Diesel-electric locomotives and must be adhered to as standard if the newer locomotives are to be trained with the older ones without duplication in systems.”


The latter indicates that accommodating the 20-notch GTEL control was not outside of the ambit of the basic Amplidyne system.



Cheers,

For whatever it might be worth the following is from GEJ-3849, "Educational Manual Model U30 Diesel-Electric Locomotive" pub. 4-67, page 1035:

"If other than a General Electric locomotive is leading the consist, the XB wire is not energized during motoring. However, excitation will be received on the trailing General Electric units from the XC wire. As the leading (other than General Electric) locomotive notches out on the throttle, no signal will be put on the XB wire. The engine speed signals will be trainlined on the GF, AV, GV, CV and DV wires. Current flows into the resistor combination (R1, R2, R3, R4 and R5) through the blocking diodes (RT1, RT2, and RT4.) The various combinations of resistors, when energized, produce a variable voltage on the XC wire. As the throttle handle is moved to successively higher notches, the voltage on the XC wire increases just as it did on the XB wire. The voltage on the XC wire is proportional to the throttle notch position. Since the load potentiometer is connected to the XC wire, the reference current and excitation will correspond to the throttle-handle notch position."

This is of course the earliest AC-DC drive configuration with pulse width modulator excitation.

-Will Davis

That certainly does help. The five resistor, three diode motoring excitation control matrix used on the U30, switched by the GF, AV, BV, CV and DV trainlines, looks to have been very similar to that used on the export Universals from 1956. That tends to confirm that the previous assumption that the U25 was similar.


Cheers,

Pneudyne wrote: ↑Wed Feb 08, 2023 11:23 pm
The GTEL4500 prototype and the first production batch had Amplidyne excitation control, derived from established GE diesel-electric practice. The second batch had static excitation, although I imagine that the curves were similar.

Perhaps not. In assuming curve similarity, I had overlooked the fact that the with the GE Static excitation system, it was possible to construct approximations to constant power hyperbolae curves. It was probably not done in diesel-electric locomotives given that the load control function was taken care of by the governor-controlled rheostat.

The load control possibility was not mentioned the pertinent GE AIEE paper (1), but it was covered in the pertinent patent (2). But I discovered it first in a Sulzer paper (3) on diesel-electric locomotive control.

Here is the pertinent diagram from the Sulzer paper:


The associated commentary was:

“The [load control] windings artificially give the main generator a manner of drooping “natural characteristic” with two different slopes to follow as closely as possible the constant load hyperbola. Thus the load regulator LCR only operates for final adjustment of the load.”

And this is the diagram from the GE patent:


Explained in the text were the additional reactors, etc., required to change the straight line RS into an approximation to the hyperbolic curve RUS.


It would appear that it was possible, and maybe probable that GE used the load control facility in the version of its Static excitation system that was applied to the second (“veranda”) GTEL4500 batch, in order to obtain notch curves that approached the constant power type. If so, it might have been the only Static Excitation application that included “artificial” load control.


(1) AIEE Paper 54-190, 1954 May, “Static Excitation for Diesel-Electric Locomotives”, by S.W. McElhenny and R.M. Smith (both of GE).

(2) US patent 2883608, filed 1955 January 03, issued 1959 April 21, “Static Excitation Generator System”.

(3) ILE Paper #623, 1961 October, “Control of Diesel-Electric Locomotives”, by O. Schlaepfer (Sulzer).

One might say that the folks on the locomotive engine side of Sulzer were students of diesel locomotive control systems, and looked at the wide range of systems in use worldwide. Sulzer always gave due credit to the pioneering load control work of Hermann Lemp, something that some others – including GE - who used his basic system did not do. In this case the GE source quoted by Sulzer was publication GET2573, which I have not seen.



Cheers,