JimC

I think everyone who has written books or magazine articles knows the temptation to use a good story that has slightly dubious provenance or has been exaggerated just a tad. Lets face it we've all come across tales that have grown in the telling. And its also going to be the case that not every story told in the first person actually happened to that individual. As for Tuplin: well he definitely has, shall we say, a weakness for a good story. The other one I take with a bit of a pinch of salt is Gibson's "critical appreciation of GWR locomotive design" where I've found a number of places where what he has to say literally doesn't add up!

How does that work with clearances vis a vis the wheels? Obviously there are ways and means to juggle it a bit, but isn't the main limiting factor the need to have the connecting rods in line with the piston centre and reasonable bearing widths for connecting and coupling rods?

I suppose one should add that problems of water supply, quality and cost were a major preoccupation to the railways. From my readings of GWR minutes the cost of water supplies was always discussed, and water supply contracts etc were a significant preoccupation at director level in the GWR. Improving water quality was also a priority, because better water meant better boiler life and had a very great influence on running costs. In his book Ken Cook records how Swindon made big improvements in boiler life through both selection of supply and installation of water softening plants.

According to RCTS 157 was Northern division based, and that boiler was fitted at Wolverhampton. I don't know if its just a freak angle, I suppose it must be, but it almost looks as if those rivet heads are polished metal! That photo is in RCTS but dark and muddy, that's a far better version with much more detail visible. I may attempt a sketch of that form. Very hard to distinguish between colours in that photo though isn't it. Really useful post, thank you.

There were actually two 157 classes. The first, above, was specified by Gooch and built by Sharp Stewart in 1862. They could be regarded as a development of the earlier 69 class with larger driving wheels. They were numbered 157-166. They were little altered in their lives, with only one receiving a new boiler, from an Armstrong Goods. They did receive weatherboards and it is possible that some may have been given open cabs. Most were scrapped in 1878/9 when the new 157 class took over their numbers. The last three stayed in service until 1881; these survivors were renumbered 172-4 for their declining years. I don't usually put drawings up here that are featured in my book unless very heavily updated, but I thought both 157 classes should be in this piece. Normally 19thC heavy rebuilds, even those reusing "only the space between the wheels" were classified as renewals, but RCTS tells us the second 157s were classified as new. Its interesting that even when renewals were effectively all new locomotives they tended to keep the same frame type as their predecessors as these did. As with most of the later 19thC classes they had a healthy variety of different boilers fitted over their lives, domes and domeless, belpaire or round topped box. This sketch is intended to be representative for around 1900. A little prototypical note though: I typically draw my locomotives bearing the class number, so 157 here. But to be strictly correct then as I read RCTS 157 herself never actually ran with a domeless Belpaire firebox boiler. And finally a Wolverhampton variant of the class, discussed in comments below.

There was also the load problem. AIUI most vans ran at best part full, hence the later development of partitioned vans. No point in a 20T van with 6T of merchandise in.

Well, I've made a small amount of progress on my 3521 tank drawing. I found a detailed drawing of the trailing axle assembly of the 0-4-2T version in Ahrons "British Steam Locomotive". Just need the rest of the locomotive now. GWS, who have a considerably stronger 19thC drawing collection than the NRM, have a fair collection of odd components - much of the trailing bogie is available in detail drawings by the look of things - but no general arrangement or frame plan. Amazing what strange things survive. The NRM do have a couple of weight diagrams, but at 27+ quid a shot and with 19thC weight diagrams typically having next to no detail... Suppose I had better bite the bullet... See what I mean... Here's where I am. It will be that boiler, although not sure what position, , frame outline on the trailing wheels drawing is from the drawing in Ahrons, frame outline on the driving wheels is from the frames as reused on 4-4-0s and the two don't line up... A bit of perusing of drawings tells me that the leading bogie on the odd 4-4-0T pannier tank 1490 is pretty much the same as the trailing bogie on the 3521s, so there's another source...

I think a drawback with superpowered 1840s locomotives would have been the same one that largely prevented superpowered locomotives in the 30s - the rest of the infrastructure. Double the train weight and how are you going to stop it? I don't suppose there was much in the way of refuge sidings etc at that time either.

Fair enough, the specific example doesn't work on inspection, but I think the point remains that different technology /operating patterns could have led towards different working patterns and different solutions. I'm not one of those who believes our ancestors were fools, but sometimes with the benefit of hindsight we might be able to imagine feasible alternatives.

OTOH that's a ridiculous way to run a railway. Desperately expensive. Now supposing the mineral business could be migrated to 12 ton or 16ton containers loaded in 3s and 4s onto converted coach chassis as flats? Existing coal handling facilities modified to suit the containers, which aren't so very different from the PO wagons in size. Maybe the same to general freight traffic. Now we have a freight traffic that can run at stopping train speeds. What's going to be required?

Pannier tanks largely rest on the frames with brackets front and rear, but I guess that would still reduce twisting. With the big side tank classes you also have the issue that a substantial part of the water capacity is in a tank under the coal, and the connection between could be awkward to keep tight. There was that welsh class which would derail, fracture the link and dump all the water on the track. They didn't last long!

I've been trying to see if one could bodge a 4'1 wheel 10 coupled wheels into a 56xx platform to get a Welsh banker, but not only is it a considerable struggle to fit everything in without a big rework of the valve gear, but also, when I did the sums, it turns out that, even with a 225psi boiler, the tractive effort is no greater than a 5205 and it certainly wouldn't handle curves any better. I suppose one could look at a 10 coupled 4'1 wheeled 5205, but I'd start worrying about clearances. This is the imaginary topic, and we can be as impractical as we like, but its getting hard to argue with the proposition that in real terms the GWR ran the traffic as efficiently as was practical.

A dedicated banker... Hmm. I suppose the thing about the Lickey was that it was a busy line with trains going up and down all day, and especially with the low powered MR locomotives there was a day's work without too much time sitting idle. I guess we have to page @The Johnster here. How many iron ore trains a day were there? And also how long did it take them to make the run? If there isn't a reasonable enough work to occupy the crew for a reasonable turn then as well as the capital expense of the dedicated banker you're also paying a crew to sit around and smoke and burn coal for no benefit half of the day. And the length of the run matters for the design too, because can you have a smallish boiler that's mortgaged to the hilt for the trip up, and then recovers waiting for the next train, or do you need full sustained output?

But are Mallet's practical under a GB loading gauge? I suppose the 2-8-0+0-8-2 Garratt Beyer's proposed to the GWR would have done the job. But the factors listed in the posts above might well have precluded a small fleet of specialist locomotives.

To the (substantial) limits of my knowledge I think a 4 Cyl 4-8-0 or 4-10-0 will work, but not a 2-8-0 or 2-10-0. However I have doubts about whether a 4 cylinder locomotive would be much more capable than a 2 cylinder one at the specific task of slogging up Welsh valleys at low speed with a heavy train of iron ore. A mixed traffic or express 4-8-0, on the other hand would most likely be a viable proposition

Its not actually photoshop, I use a vector drawing package, but anyway... I don't think I can make a 4 cylinder 2-8-0/2-10-0 work in the GWR style. As you're aware all the 4 cylinder GWR classes had leading bogies. Assuming divided drive then all the weight of the inside cylinders and about half the weight of the outside ones would come on the pony truck, which would be way too much. And if you imagine inside cylinders driving the second axle then the room available for the boiler is greatly reduced due to the valves (unless you end up with an arrangement as heavily compromised as the Lickey Banker in which case its hardly worth adding the extra cylinders). A 3 cylinder layout might be more viable because the valve can go alongside the cylinder instead of above it, but by this stage its departed so far from the GWR kit of parts that its me designing the locomotive, and that's not a viable option! So we could possibly have a 4 cylinder 4-10-0, which would be a mighty beast on paper but... Another consideration is that the GWR 4 cylinder layout used Walschaerts valve gear, which meant only two eccentrics between the frames, which could just about be squeezed in, but the Churchward/Pearce implementation of Stephenson's was considered better at starting a train. That would mean four eccentrics, and are are you going to get all that in and still have decent size big end bearings? It would probably be possible to work out an outside Stephensons with two return cranks like the odd Black 5, but then you're left with having the rockers to drive the inside cylinders the opposite side of the cylinders from the valve gear, giving you timing problems with heat expansion... So all in all the 4 cylinder freight engine for the valleys looks like a rather difficult prospect, and after going round in circles like this I think you end up with the King boilered two cylinder 2-10-0... On the other hand a four cylinder 4-8-0 cross between a 47xx and a King as an ultimate fast mixed traffic locomotive might be a viable option. I wonder if it would be possible to squeeze in Hall sized 6 foot driving wheels. One up on a Hall - call it a Palace class?

But better to put a proper boiler on it! This is a King boiler on much the same chassis. If I were doing it a bit more thoroughly I would juggle with the wheel positions to get the fixed wheelbase shorter, this was (obviously) a quick hack job - oops, look at the top feed!

I suppose in theory it would look something like this. In practice though it would be of very limited utility, since there would be little weight for the rearmost wheels to carry. A 2-10-0 with a 47xx, Castle or even King boiler, on the other hand, would be a more useful proposition.

It varied. At worse string (well hopefully not in the 20thC). At best the GWR introduced Zeiss optical alignment kit in I think the 30s, and Cook (NE and E CME, ex Swindon works manager) introduced optical alignment to the ex LNER post war. I don't know about the SR and LMS.

I think the shade of Ken Cook would tell you that if you build locomotives accurately to narrow tolerances they run higher and much more predictable mileages between overhauls and don't shake and hammer themselves to bits.

An excellent question! I've been wondering about timescales too. In 'Chronicles of Steam' Cox talks about LMS staff brainstorming the general outline of the new class while the exchange was actually in progress, and makes no mention of Castle drawings or purchase. Mind you Cox also manages to completely avoid any mention of the Nelson drawings. I have to say I'm not convinced Cox is always the most dispassionate of reporters. I suppose you wouldn't expect the LMS board to trial a GW locomotive and then borrow SR drawings unless the GW had turned them down. Holcroft doesn't seem to mention it on a quick browse of his books, although he was SR at the time and talks about the Nelson drawings. He also reports that the LMS had already decided on 3 cylinders and the drawings were sent straight to North British, who he says mainly used the boiler drawings, adapted for LMS standard tube sizes.

And was done. When the GWR wouldn't supply Castle drawings the LMS got Nelson drawings from the Southern. And of course Stanier took GWR drawings with him when he went to the LMS.

I have a feeling that the Castle detail design, geared for Swindon's sometimes expensive and finicky maintenance, might not have fared well at LMS sheds and factories. And what's the betting that after the drawings had passed through the LMS drawing office they would be subject to the same kind of 'improvements' as the Garrett design received? We know that a number of GWR features didn't transfer well to the LMS even with Stanier there to keep things in hand. Without him, well, I think the GWR board probably did the LMS a favour in refusing the drawings. Look what happened when Inchicore had a set of Churchward Star drawings.

The F class was very similar to the A class except for the saddle tank. The F class is one of the trickier ones to sketch out, because there were several different batches from builders, and variations between the batches, front overhang for example, definitely existed. There are two styles of foot plate valance too. The first five at least had a straight valance, the remainder curved as drawn. This second sketch shows a lightly swindonised version of the F class, still with the Barry boiler and bunker, but GWR safety valves and cover and several other standard Swindon fittings.

Yes, but this is what statistics are for. You can pull out surprisingly small differences. 5-10% should not be too hard if wind strength and direction is adequately recorded.