I just read the CT article by the late Dave Ingle. RIP, Dave.
I appears that ATSF numbered A-B-B-A sets of F units as a single locomotive. That seems like a practice that would have had built-in problems. Let’s say it’s #48. I assume the rear-facing A was also physically numbered 48 on the side, as well as using the numberboards.
Were the B’s also numbered 48?
It seems like they were seriously limiting their flexibility with this set-up. Or were there changeable numberboards on the sides of the Bs also? IE, could one of thise Bs be #46 the next day? Were there different, actual equipment ID numbers that were less visible?
How did this all work? I mean, if you need to swap out a unit, re-painting its number seems a bit excessive!
The best way to imagine this is to think of an engine numbered this way as a single articulated ‘locomotive’ – in the case of an FT, a 5400hp locomotive composed of four ‘units’. (Ignore for a moment which of those units have semipermanent drawbar connections, couldn’t be operated separately, etc.) In part this was done because union rules hadn’t yet determined whether a full or partial crew was required on each locomotive ‘unit’ (or separable section) controlled in MU.
As such, of course there would only be one locomotive ‘number’ and it would be on both ends as it’s a long single locomotive that happens to have three bends in the middle.
For servicing, these might have subnumbers. ATSF went through some peculiar conventions in this respect – take the first PA sets, which were numbered from the arbitrary consist ‘front’ as road number, say 51, suffixed by “L” for lead, then “A” (which happened to be a B unit), then “B” (worse, if you have no aspirin, an A unit in this case). Your four-unit hypothetical would be 48LABC in this case (with L and C being the units with cabs).
Now it would not take long for the railroad to realize that keeping a 5400hp steam locomotive out of service during maintenance was a necessity, but that for an EMD was a waste. So pretty quick they’d be looking at the standard convention of coupling A-B drawbarred sets in pairs to get the 5400hp … now perhaps with a separate road number for each ‘half’ so you wouldn’t find 48L on one train meeting 48C on its counterpart the other way. But still no severe reason to treat the booster different from its cab, except that usually only one unit would need work and the other sat idle in the meantime.
Mind you, actual systems could be, and were, remarkably more complicated, and not always very effective long-term.
Once the agreements on firemen in trailing units, weight equ
Unfortunately, Overmod’s explanation is oversimplified…
The Santa Fe system was even more complex and there was numerous examples of renumbering, all the way back to “Amos and Andy” which started off as 1a and 1b, went through 1 and 10 as separate units and became 1L and 1B. 1B became 2611 as a switcher (with new Blomberg B trucks) but 1L was “converted” to an E8 before becoming 2610 and in due course 2611 became an E8B.
I think Santa Fe had couplers on all units of their FT units right from the beginning, but there was concern that only one crew should be used on FT sets and after the first two sets (100 and 101) were delivered, the sets were reorganised as ABBB sets (so there was only one cab).
Santa Fe’s PR department was working hard on the Canyon Diablo bridge and at least three photos appear in the CT “Photo of the day” collection. This one shows two trains with four unit FT sets carefully posed on the two bridges, with that in the background on the new bridge visibly stretching into the distance.
The nearer FT is formed up ABBB while the further unit is ABBA. I think this was at the stage where the unions had agreed to two A units per set of four units.
So ATSF purchased sets with three B units, then a batch of A units, all involved in a complex and unnecessary series of renumberings. Later FT sets were broken up into three and two unit sets and the numbers got out of hand.
Also, it should be noted that FT A units had illuminated numbers on the sides as well…
Santa Fe kept using subnumbers right up until the last F7s hit the Cleburne shops to be rebuilt into CF7s. By convention the “L” units did not carry the subletter in the indicators, but “C” units did. There were some “A” and “B” cabs, mostly among the PAs (Mike and Ike were 1 and 1A, with 1A becoming 10)
Other roads followed various systems. Chicago Great Western had “B” units sublettered up to “G” since they had many more Bs than As. GGW cabs were “A” and “C”. Freight Bs were “C” and “D”. The passenger Fs included FP7s 116A and 116C, and F7Bs 116B,D,E,F and G.
Soo Line started out with no Bs, numbering their F3s/F7s/FA1s as A/B sets. This broke when the passenger boosters arrived, initially “B” but later “C”. Freight Bs were “C” from the start. Soo eventually stopped trying to keep units together, and even had one unit 2204C with no corresponding A or B (and hostler controls, the only Soo F7B with them). Soo F-units kept their subletters until retired, so 2203A could be expected to encounter 2203B at some point, though of course not 2203C which would be trailing…
SP started very early numbering B units in a separate series. E units in the 6000s for As, 5900 for Bs. On the SP and some other western roads the indicators were used for train numbers with “X” preceding the engine number for extras.
There has been some argument about this over the years, and I don’t even pretend to be an authority on what was actually done.
I believe the demonstrators were built with drawbars between A and B units, and the B couldn’t be operated separately even if it had been given hostling controls (my memory saying there were components of some systems that were not duplicated on both units). It would be perfectly logical for ATSF to order ‘special’ versions with couplers, but the accounts I dimly remember had ATSF doing that work later, replacing the drawbar with couplers and making other changes as required – there was some discussion about what the result ought to be called by ATSF but of course it was inconclusive. We recall that the F2 was originally specially designed to make 3-unit sets out of FT ‘halves’ which I think were then assumed to be a drawbar-connected A-B pair – of course, that could be wrong, too.
MKT added to the confusion in a general renumbering in which F-unit cabs had A and C suffixes while boosters had B,D,E,F etc. suffixes. MKT had four FP7-F7B-FP7 sets originally numbered 78ABC to 81ABC. The F7B’s were later renumbered as 78BDEF.
When the ABBA F3 set #16 arrived, it just showed “16” in the numberboards on both ends. Likewise with all the ABBA sets of F3s and F7s. I assume the F3As and F7As always had that small “L” or “C” painted on the nose above the coupler, but I’m not enough of a SFe expert to be sure.
So that means the 16 set was never broken up until … bet no one here knows when. In, say, 1955 SFe passenger trains had one of the ABB sets of F7s, or one of the ABBA sets of F3s or F7s. If the lead unit said 35, the trailing A-unit did too.
In January 1958 the Super and El Cap were combined – dunno if SFe considered 6000 hp enough for an off-season train. If that was when five-unit sets of Fs began to appear, then maybe that’s when some A-units got numberboards with a “C”. But far as I know, no numberboard "C"s until a few years later.
No SFe unit ever carried an “L” in the numberboards – right? (Or maybe there was one exception, after 1970?)
According to Lawrence W. Sagle in ‘The Picture History of B&O Motive Power’, the B&O assigned the 4 unit A-B-B-A FT locomotives road numbers of 1, 3, 5, 7, 9 and 11. The cab units ‘in small letters’ were 1 and 1a, the B units were 1x and 1ax. The B&O’s FT’s did have the drawbar between the A and B units. Later the locomotives were renumbered starting at 101 with the same pattern being followed.
If they were more or less permanently attached, the I sort of get this. But it seems to me it would have been obvious from the start that it would be a lot simpler if each unit had its own unique number (like almost everything else on a railroad).
Let’s say 2-3 units of the A-B-B-A set craps out enroute, bringing the train to a halt, like in hilly country. Did they just send out the posse, IE other locos to pull the whole mess along? And then when they got to a shop, the whole foursome was out of service until the ailing ones were fixed?
Seems crazy. Although, if they had not yet worked out disconnectable MU cables, etc., then I guess that was that.
By frame numbers, AT&SF F3 #16 was delivered as 16ABCL. PAs and early E units had lighted numbers on the sides (PAs also had the number over the windsheld center), DL109s and Eries had the numbers painted on their sides. For some reason neither freight nor passenger Fs had obvious numbers on the carbodies. FTs had the number plate on the front below the headlight.
My personal observation - the number plates below the headlight seem to have been something that started in the late 40’s and was out of favor by the late 50’s. Many roads, not just ATSF, did it and then moved on.
The display of engine numbers on the early diesel cab units, to my eyes, appeared to be very hard to read under any circumstance especially at night.
As I recall, it wasn’t the engineer that was the problem on an MUed locomotive; the rate would go up based on the controlled engine weight, just as it would for a Henderson quad or quint if you were to be assigned one. It was the firemen, one per unit just as there would be for separate steam engines called for a consist…
In truth, this might not have been as bad … in the 1940s … as, say, the 81-car limit that was being pushed as a national standard in the Thirties, the thing that formally killed the N&W Y7. In full-crew states, that would have resulted in enormous duplication under the balance of applicable rules and agreements.
On the other hand, what a different decision would have done would be to change the rules of engagement for large single-unit locomotives with high horsepower, a ‘poster child’ being the Baldwin development with the multiple transverse 408-engined gensets that became the prewar demonstrator of the Essl locomotive. If an economic basis for 6000hp in a single fairly short carbody had come to exist, we might have seen evolution of effective select-a-power gensets many decades early, together with relatively high-speed experience for Baldwin in parallel with whatever they might have done with free-piston alternatives.
Maintenance on those things was interesting. You took some pipes and cabling loose, and the whole unit, including radiator circuit and part of the roof, slid out like the works-in-a-drawer on a RDC. Get a spare from the shop and you’re back on the road in 20 minutes or so …
Just to make things even more complicated, here’s another example. CN numbered their F3 and F7 B-units in the same number series as the A’s.
Initially two A-B-A sets of F3’s were purchased, numbered 9000-9005. 9001 and 9004 were B-units.
For the F7’s, even numbers were A’s and odd numbers were B’s. As CN bought far more A’s than B’s this resulted in a whole bunch of higher odd numbers being skipped.
CN later switched to giving passenger B-units their own number series, but those earlier units kept their original numbers until retirement or rebuilding.
Outside of ATSF … some railroads adopted additional conventions for discriminating A units from B units. We’ve discussed some already, including alternating with As (say) being given even numbers and Bs odd. On the Southern (and its subsidiary CNO&TP, the infamous Rathole) the DL109 A units were assigned a block of consecutive numbers, and the ‘corresponding’ DL108 boosters got corresponding numbers 50 higher – on CNO&TP this was 25 higher. This implicitly tells you how many of these locomotive ‘sets’ the system intended to acquire… it also indicates that some railroads had ‘number series to burn’, like CN, when they assigned series from different makers to different reserved blocks.
Someone like rcdrye will have exhaustive examples of this sort of thing, and the strange ways it could evolve over time (as I recall, there were some interesting details of IC E-unit numbering).
Every company has their own numbering conventions for their locomotive fleets. These conventions will change over time as the locomotive fleet changes in character. Some locomotive series going to the scrappers and new and different locomotives appearing enmass. In today’s world of computers the numbers are arranged in series so the the computer can be interrogated to generate various statistical reports measuring the various aspects of locomotive operation as well as costs of maintenance etc. One company’s numbering and naming conventions are not necessarily the same as another company’s.
I don’t remember any in America, but there were some weird ‘early computer compatible’ numbering systems in Europe, with parity and check digits and other stuff that European computer engineers might find like catnip, but that anyone actually trying to operate a railroad would suffer from.
The closest thing I can think of, and it isn’t very close, was the relatively short-lived use of reflective button numbers, before Scotchlite was invented, to give you numbers that could be read easily both in the day and night without onboard electric lighting.