• Baldwin vs. Boxpok driver centers

  • Discussion of steam locomotives from all manufacturers and railroads
Discussion of steam locomotives from all manufacturers and railroads

Moderators: Typewriters, slide rules

  by Allen Hazen
Boxpok -- a proprietary name for the General Steel Castings design -- were the most commonly used disc driving wheels on late American steam locomotives, but the Baldwin Locomotive Works had its own type: recognizable from the raised "lips" around the openings and, if you get a good view, from the words "Baldwin Disc Driver" cast on the face of the driver center. I think they were broadly comparable in physical characteristics. Baldwin drivers were used in a number of locomotives of the late 1930s -- Santa Fe Hudsons and Northerns, for example -- but a reasonable number of late Baldwins, notably the (production) T1 duplexes for the Pennsylvania -- had Boxpok. Is there some (known and interesting) reason why they used Boxpok instead of their own version on these late locomotives? (The obvious thought is that there was a cost advantage. But -- given the collapse in the market for steam locomotives in the 1940s -- I would have thought that Baldwin would have had plenty of capacity at their own foundry, so I would be surprised if they were unable to produce in-house driving wheels as cheaply as they could buy them from a subcontractor.)
  by Allen Hazen
Hmm… I ***think*** General Steel Castings was a joint subsidiary of Alco and Baldwin. And one of its foundries was in… Eddystone, PA. So maybe the financial aspects are a wash: buying from GSC is close to being an internal sourcing for BLW. (??? I'm sure more could be said about this by someone who knows more than I do!)

With regard to the Pennsylvania Railroad's T-1… The prototypes, 6110 and 6111, built in 1942, had Baldwin Disc Drivers: very recognizable in photographs, given the lips around the openings and an additional ridge op the middle between any two openings. The production series, built in 1945, had Boxpok. So maybe BLW decided, during the war, that it wasn't worth while trying to sell their own model of disc driver in competition with Boxpok?


The PRR T-1 is what made me think about the issue: the December 2015 issue of "Railfan and Railroad" has a two-page article by spokespersons of the Pennsyvania Railroad T1 Steam Locomotive Trust, the organization that is trying to build a new T-1 (inspired by the successful British project to build an all-new locomotive to a 1945 design). Their project has apparently gotten to the stage of making a wooden pattern for casting new Boxpok drivers, and the reference to Boxpoks made me wonder why a BLW locomotive wouldn't have had BDDs instead of Boxpoks.
  by Pneudyne
Allen Hazen wrote:So maybe BLW decided, during the war, that it wasn't worth while trying to sell their own model of disc driver in competition with Boxpok?
Could the WPB have been involved here, perhaps directing that there should be only one maker of disc drivers? The timing seems to be about right.

Other supporting evidence of Baldwin builds with Boxpoks from a quick skim through the library:

The Santa Fe 2900 series 4-8-4 looks to have had Boxpok drivers, whereas its immediate predecessors had Baldwin discs.

The Lehigh & Hudson River 4-8-2 had Boxpoks, whereas the B&M R-1d from which it was cloned had Baldwins.

The B&O EM-1 2-8-8-4 had Boxpoks

The WM 4-8-4 had Boxpoks, even though the earlier4-6-6-4 with which it was essentially homologous had Baldwins.

The MoPac 2200 series 4-8-4 had Boxpoks.

I think that there would be other examples. That is not to say that one could not find say a post-1942 Baldwin build with Baldwin disc drivers.

The WM case suggests that even after WWII, and whatever the reason for abandoning its own design, Baldwin stayed with Boxpok drivers where discs were required.

I have never seen even a hint that there was a material technical difference between the Baldwin and Boxpok designs. (I’m not so sure about the Scullin version, though.)

  by mp15ac
What I find interesting about disc type drivers is that the C&O 2-6-6-6 H-8's and the entire N&W roster ran on regular spoked drivers. Neither railroad felt that the disc drivers offered any advantage of spoked drivers.

  by Allen Hazen
(Quick comment: will have to think and research more.)
I had a feeling the later Santa Fe 80-inch driver 4-8-4 had Boxpok drivers, and I know the first series had Baldwin Disc. In maintenance… somewhere on the WWWeb I came across a photo of a Santa Fe 4-8-4 with mixed drivers: Boxpok on some wheels, B.D.D. on others!
  by Allen Hazen
Norfolk and Western, I think, had their own foundry. They probably thought the advantage of disc drivers, if any, wasn't enough to justify the added cost of buying from outside. (But I'm not sure they had their own foundry: if they were buying from G.S.C. anyway…)
C&O 2-6-6-6 is puzzling: Lima liked high tech, and C&O was usually willing to spend money on extras!
What's really puzzling… is that one of the prime advantages claimed for disc drivers was that they allowed better counterbalancing, which I would have thought was particularly desirable for comparatively SMALL driving wheels. Yet most of the disc drivers I recall were on high-wheeled passenger power!
Sorry, those are VERY off-the-cuff responses. I'll try to think more and get back to you with a more considered response.
  by Pneudyne
On disc driver diameter range, shooting from the hip, I think that original fitments spanned the range from 63 to 84 inches. No examples of lesser diameter come to mind, although there might be one or two somewhere.

I suspect that better balancing including the ease of cross-balancing, coupled with lower mass, made disc drivers attractive for high-speed locomotives, where any lessening of track pounding and wheel-lifting tendencies would have been welcomed by most roads. As an example, the NYC J-3a (with Boxpok or Scullin disc drivers) was said to have around 50% lower dynamic augment than the J-1e, originally with spoked drivers. On the other hand, with say a 63 inch-drivered 2-10-2, whilst lightweight rods and disc main drivers would help, they would not make a silk purse out of a sow’s ear, and by the time disc drivers were being used for retrofitting and rebuilds, such locomotives were typically in secondary, lower-revenue service.

N&W probably did cast its own spoked drivers. By way of indirect evidence, TSC #51 included a reprint of the RME 1945 June article on the N&W J. The partial list of materials and equipment customary with such articles unusually did not include a supplier of the driver centres, which leads one to infer that N&W “rolled its own".

The C&O seemed to adopt a “brute force” approach to its steam locomotive designs, including an apparent “heavier is better” philosophy. Apart from the Allegheny weight growth story, its version of the Van Sweringen 2-8-4 was a lot heavier than the others. And the C&O was accustomed to track-pounding heavy locomotives such as its 2-10-4, which had brought with it a quantum increase in axle loading compared to previous norms. I have wondered whether the high (at the time) axle loading of the 2-10-4 was chosen in part to keep the dynamic augment-to-axle load ratio within reasonable bounds. The Allegheny had a dynamic augment, at wheel diameter speed, of 16 380 lbf. That compares with the UP Big Boy number of 7590 lbf. Against that background, the benefits of disc drivers were probably of little interest to the C&O.

Returning to Baldwin, whilst it had its own design of disc drivers, it also fitted the Boxpok type from quite early on in the disc driver era, presumably at customer request. The 1938 version of the WP 2-8-8-2 looks to have been an example.

Regarding the Santa Fe 2900 class 4-8-4, this from the Hundman Locomotive Cyclopedia Volume I: “The 2900s were built with BoxPok drivers and tapered rods. Post-WWII modernization brought new lightweight roller bearing rods and the substitution of a Baldwin disc driver for the fourth axle to improve driver/rail dynamics.” That suggests that the Baldwin disc design had very specific benefits in this case, so there was a technical difference between the Boxpok and Baldwin designs.

Finally, it would appear that all of the RF&P 4-8-4s had Baldwin disc drivers, including the 1942 and 1945 batches. That weakens the WPB intervention theory, although the 1942 batch may have slipped through basis having been ordered and in production by the time that the WPB intervened, and by 1945 the WPB may have been loosening the reigns somewhat. Those 1945 RF&P 4-8-4s may well have been the last locomotives fitted from new with Baldwin disc drivers.

  by Pneudyne
On disc driver diameters at the smaller end, LeMassena, in his book on the American 2-8-2, shows a picture of a 1936 Baldwin 2-8-2 #15 with what appears to be Baldwin disc main drivers. This same locomotive appears in TSC #45; the driver size was 57 inches. The photo therein, which I suspect shows it in original condition, is somewhat indistinct in respect of the main drivers, but on balance I’d say that it was of the disc form.

So this may an example of a relatively small diameter disc driver original installation. If so, I think that it would also have been one of the very early Baldwin disc installations.

  by Pneudyne
This subject gets more complicated the more I look at it. You have opened a can of serpents, Mr. Hazen!

In terms of Boxpok driver initial fitments to production locomotives, early – but not necessarily the earliest – include the Alco-built DLW Q-4a 4-8-4 of 1934, and the Baldwin-built NP A-2 4-8-4 of 1934-35. The former had what might be called the customary Boxpok design, whereas the latter what could be called the “big hole” type.

So Baldwin had an early acquaintance with Boxpok drivers, probably starting before its own design became available. And some customers, such as NP, SP and WP, continued to specify Boxpok pre-WWII even after Baldwin had released its own design.

Evidently there was some kind of external factor at work during WWII, otherwise it would be difficult to explain why roads like the AT&SF, previously committed to the Baldwin disc driver for new builds, accepted the Boxpok type.

Lima used the Boxpok type. The earliest example that comes to mind is the SP GS-2 4-8-4 of 1936.

Scullin seemed to be a minor player when it came to original installations, being more involved in retrofits. Its biggest original equipment sale seems to be its half share (25) of the NYC J3a fleet. The fourth make of disc driver, the LFM Universal, seems to have been for the rebuild/retrofit market only.

Was Scullin the first to offer a disc driver? Innovation sometimes comes from “outsiders”, so it is plausible. Thus the Boxpok could have been GSC’s necessary response, given its major role as casting supplier to the locomotive industry. I wonder how much control Alco and Baldwin exercised over GSC. Possibly the Boxpok was a GSC “internal” design with not much other than general guidance from Alco and Baldwin. Then Baldwin thought that it could do differently (and presumably in its mind, better), and did so.

  by Allen Hazen
Thank you for … sorting through the can of slithery things! You've researched this much more systematically than I have. I don't know whether I will be able to contribute anything further, but if I do it will just be a footnote to your posts!
  by Pneudyne
The following excerpt from Reed, Loco Profile #20 “The American 4-8-4” provides some information on the chronology of disc drivers. It was part of the description of the NP Class A-2, in which its differences as compared with the A-1 were enumerated:

“Another variation was the use of Boxpok drivers, a form just then being introduced. This wheel was due to General Steel Castings Corporation, Alco and American Steel Foundries. Derivation of the name was ‘box-spoke’, the spokes, much fewer in number, being of box section, as was the rim, this giving great lateral strength and rim stiffness. About 26,000 Boxpok wheels are believed to have been applied throughout the world.

This wheel had been preceded by the Scullin double-disc centre (1932) which was used mainly by the N.Y.C., and was followed by the Baldwin disc type, as illustrated in the front-cover close-up of one of the A.C.L. 4-8-4s.”

Given that there was the hand of Alco in the Boxpok design, Baldwin’s motivation for developing its own design may have been less that of doing it better and more that of doing something that was not connected with Alco.

  by Allen Hazen
Hmmm…. The same names keep cropping up!
The Wikipedia article on General Steel Castings begins
"The General Steel Castings Corporation was a steel casting corporation in the United States established in 1928 by the Baldwin Locomotive Works, American Locomotive Company, and American Steel Foundries."

Without more information about what "established by" signifies, I take it that this means GSC was a jointly owned subsidiary of the three named companies. So, if the Boxpok driving wheel centre was "was due to General Steel Castings Corporation, Alco and American Steel Foundries," it is credited jointly to … GSC and two of GSC's three main owners! … It would be interesting to know (and there's probably a relevant database available on the internet, but I don't know how to look) who the patents were assigned to!

(The W-ia article also says that GSC, the year after its founding, merged Commonwealth Steel Castings. Commonwealth: as in truck frame and tender bottom and…)
  by Allen Hazen
Not much joy yet on the patent front…
The Scullin driver seems to have been an idea of the Scullin Steel Company (of St. Louis, Missouri), which appears to have gone out of business around 1980 (strongly suggesting that it WASN'T at any stage acquired by or merged into GSC) and (remarkably!) doesn't have a Wikipedia article!
http://www.freepatentsonline.com/2177693.html" onclick="window.open(this.href);return false;
if not the patent for the Scullin driver(*), seems to be a related development…
((*) Granted 1936: sounds much too late, given the introduction of the type in 1932. But I think it was possible, in the first half of the 20th C, for companies to take out patents some time after making an advance: one of GE's diesel-electric locomotive patents wasn't taken out until some time after a detailed description of the control system involved had been published in the trade press. So…)
  by Allen Hazen
Maybe progress on the patent front… though not details that seem helpful on the "Why did Baldwin…?" question. GSC sought patent coverage in Canada as well as in the U.S.A., and this
http://patents.ic.gc.ca/opic-cipo/cpd/e ... mmary.html" onclick="window.open(this.href);return false;
seems to be the Canadian patent for the Boxpok driver. Issued 1934, inventor was Harry M. Pflager (working alone? or as the head of a design team? I don't know), but the assignee ("owner") of the patent was General Steel Castings.
(Found by putting "General Steel Castings patents" into a search engine: there is a list…)
  by Pneudyne
Allen, I wrote this before seeing your last two posts, but I'll post as written anyway despite the overlaps and redundancies.

Searching ‘Google Patents’ sometimes works, but it can be tedious finding one’s first “hit” on a topic. Usually though, that provides words and phases for a refined search that brings up more possibles more quickly. Once the patent numbers are known, then the respective patents are easily recalled in a Google Patents search.

In this case, from a not-too-extensive search US1960039 and US2042160 appear to apply to the GSC design. Both were assigned to GSC, but with no mention of the GSC participants.

US2065217 and US2065217 might apply to the Baldwin design. In neither case is there an assigned corporation, but in both cases the location is shown as Philadelphia, PA, which I think where was the Baldwin HQ.

US2177693 is a Scullin patent.