JimC

The other factor with US style locomotives is weight. Its all very well if you have endless prairies and few bridges, but those monsters were *heavy* - axle loading on the notorious Big Boy was 30 tons if Wikipedia is correct and I've done the calc to Imperial tons correctly. Not going to take that across many UK bridges. I think that's another way the Garratt scores over the Mallet types.

I was idly looking at Cape gauge the other day, and the example loading gauge I found (Malaysia - https://twitter.com/malayanrailways/status/1201281457294954496) was very much the same size as UK standard gauge - 9'3 across. So there wouldn't actually be much, if any scaling up possible. But it gets worse, because if the frames are widened to standard gauge is there enough room for the cylinders? , Rail to platform is 2'10.75 on Malayan Railways, but 730mm (2'4.74") on Network Rail.

The trouble is a big Mallett on the UK loading gauge would be a nonsense. The boiler would look like a pipe cleaner. I've very crudely and quickly hacked a weight diagram to give some kind of idea. On the UK gauge the only articulated config that makes sense, to my mind, is a Garratt. Imagine it as a proper Mallet with a tender and its even more ludicrous...

On a completely different topic, I'm not inclined to go back and replace missing images in this thread. I take the view they are mostly 'of the moment' and in any case I post the ones that enthuse me on my Web page (link in sig). Also I haven't kept all of them. Anyone disagree?

Very challenging to get one in the loading gauge and weight restrictions though.

There's more than one version of the genesis of the 47s. Another one is that the 43s could run short of steam on faster timed freights - with only a Std 4 boiler to feed 18x30 cylinders perhaps unsurprising - and so the 47 was an enlarged 43. Also some good sources state that the 47 was always planned with a larger boiler, but the design wasn't finished in time so the Std 1 was a stopgap. Cook also tells us more 47s were asked for, but Collett elected to build Castles instead. On 47 boiler delays, I note the studies for Std 7 boiler 4-6-0s, and wonder if they tried and failed to get a boiler down to weight for the Star chassis and that was the delay? Almost zero evidence for that speculation though. Under Collett they opted to make a smaller diameter barrel for the upboilered 4-6-0. If Churchward had done that rather than stick with the Great Bears boiler barrel tooling maybe the 47 would have had a Castle sized boiler?

Would it be worth putting a 301 permanent redirect on rmweb/community? That would help out people with outdated bookmarks. Jim C

However, as was said long ago, it still leaves you with the problem...

Lucky indeed... In the early days before organisation wide WAN links there was the burglary (server and all tapes) - "Oh, is that why you said we had to take a tape home every night? It seemed like too much trouble so we stopped doing it" There was a week long recovery which I think involved disk controllers corrupting the disks. RAID 1 with two controllers, we *ought* to have been protected. Never really got to the bottom of that, it looked like a disk problem at first, and even when I started considering the controllers it didn't really make sense. Had to put a complete new server in, and that costs time in procurement. Once the new box was in I lost interest in diagnosing precisely what the fault was! That was when I discovered just how **** poor the new enterprise back up library was over a WAN at DR restores from incrementals which involved multiple tapes. Daily incrementals are fine for BAU restores but in those days at least were not great for DR. That one was also complicated by the need to prioritise urgently needed data. The archive system was desperately inflexible, and while in theory I could have relocated the server to the DC for faster restores, in practice I also needed the server on site for a skeleton service. Hopefully the tools are better now. There used to be a saying in my NetWare days - NW admins come in two flavours: those who have Backup Exec, hate it, and wish they had Arcserve, and those who have Arcserve, hate it, and wish they had backup exec.

The 47 had heavyish axle loading, but no worse than a Castle. AIUI cylinder clearance and minimum curve radius were more restrictive. The first two drivers were quite lightly loaded, it was the last two that approached the limit, so putting a bogie on the front would be of limited value!

AFAICS all bookmarks to the site are obsolete, so its probably easiest to start at the front page and follow the menus to where you want to go.

I submit there are various reasons why its wrong, but the most important one is "its not the right b****y time". Not to put to fine a point on it you do reviews when the dust has settled, the folks have caught up with their sleep and the after effects of eating nothing but cold pizza for however long it was and the emotions and adrenalin have calmed down.

Easier said than done when it comes to terabytes, even low terabytes of data over long distance links. And in some circumstances a regular on line backup may even be enough line data to affect the hosting costs. Also restore times over the wire may be horrendous. I recall one disaster recovery I had to do over WAN links which took over a week to get everything back. Feel for the guys, been there more than once, got the T shirts. Its something I wonder about with RAID5. The chances of two separate failures may be tiny, but when all the disks are identical and from the same batch maybe they aren't as tiny as one might hope.

I considered that, but felt the weight distribution would be wrong. It was done on the rather odd 3901 2-6-2Ts though, which are somewhat congruent to this discussion. The story behind these is said to be that Swindon works was short of machining capacity, so they took some fairly new Dean Goods and crossed them with 45xx, using cylinders, motion components and wheels from the Dean Goods and fitting pony trucks, new frames and cabs and (presumably new) Std 5 boilers.

There were even a couple of experimentals back in the 19thC with outside valve gear and inside cylinders. The 1101s don't really count because they were bought in, but there were also the VOR locomotives.

Do you think there might have been an element of wanting to be seen to be adapting to welsh norms, bearing in mind that many of the staff of the Welsh lines were less than enthusiastic about being subsumed into the GWR?

OK, a little bit more thought. The 2251 was lightest on the trailing wheels, so I've cut the tank back and given it a further cut out low down to give access to the motion. I rejected the pannier tank to lighten the front end. The trailing wheels have been moved back a few inches and the bunker lengthened to match which should bring back some of the water capacity. There's also a few other minor tweaks.

Yes, as soon as I uploaded the first image and looked at it I worried about the access and went back and drew the pannier tank. A half and half like the 9700s or the 1101s might be a possibility too. One really needs to be a proper designer who can do a weights study to work out which would be best. There would also be the possibility of extending the bunker and back tank to juggle the weights.

Here's johnster's 2251 based passenger 0-6-2T. I imagine it with screw reverse and autogear blasting up and down the valleys between two pairs of 70ft auto trailers. Very quick to draw for me, and probably for Swindon. 2251, extend frames, add 5101 trailing axle assembly, drop on a 5101 cab/tank assembly and modify it for the straight footplate. [This image seems to have gone for ever, not on my PC, but see the modified version a couple of posts down] Or if the access to the motion is too impossible, then perhaps a pannier tank. Its basically a 94 tank, but the water capacity would be greater than a 94 with a bigger tank under the coal space.

And indeed was probably about as powerful as was sensible for an inside cylinder locomotive, so the only advantage of going for 8 driving wheels would be axle load, which wasn't really an issue on the welsh valley lines. It seems to be that an 0-6-2T configuration tends to be suited to inside cylinder locomotives, since the cylinders can be between the driving wheels to drive the centre axle, and a 2-6-2T is better for outside cylinders where the cylinders must perforce be in front of the wheels. The 2-6-2T can also go a bit heavier and longer in the boiler. There's also the problems we've noted above where the cylinder/piston arrangement impinges on the boiler and limits how large it can be. Its one of the minor mysteries to me that the GWR almost never put autogear on the pre group 4'7.5 wheel tank engines. I think a few of the outside framed locomotives were used with it, but I don't recall ever seeing a record of fitment to the later more powerful inside frame types. Which also meant that the autofitted fleet was 4'1 wheeled 2021s and relatives, plus Metro and 517 2-4-0s with 5'2 wheels.

Perhaps just that 2-6-2T, with guiding wheels at both ends, was usually better .

I just had a quick look. Its tricky. The piston valves on the 56 are as close to the bottom of the smoke box as they can be and still have a drumhead smokebox. So the boiler has to go up higher (or revert to slide valves). There is room for that, but there's also weight to consider, and the bigger driving wheels seem to make a surprising (to me) difference. So I suspect the tanks need to be appreciably smaller to keep the weight within limits. The water has to go somewhere though, so perhaps that forces you towards an 0-6-4T. But then you have to ask why this is a better solution than the Std 4 boiler Large Prairies.

And even worse, as the larger valves move away from the centreline the best location for the eccentrics start to impinge on the space for the big ends. *If* I read the less than ideal drawing I have correctly, the GWR dealt with this on 4-4-0s by having a solid arm on the end of the valve rod, which would put a lateral load on it which I should have thought wasn't ideal. Piston valves could go above the cylinders, in which case they impinged on boiler/smokebox, below the cylinders as per above, in which case the steam passages were long and convoluted, or perhaps weirdest of all, on some 2721 pannier tanks, both on the centreline one above the other which must have made for steam passages like a nest of snakes!

The expansion link rocks to and from, its not set at a given angle. There are various videos on youtube which show how it all works, but I didn't immediately spot one I liked. I probably should do an animation of the actual GWR gear, but I haven't the best tools for that. I'm actually struggling to work out how best to explain how I see the design factors. I'm not sure this is the best place either. Perhaps some of you folk who work at 305mm to the foot could post links to your Mutual improvement Class material or something of the sort?

Lets see if I can clarify this a bit better... Mind you I am by no means a valve gear expert and I hope for those who are to correct me... I'm not sure angular irregularities is the best phrase. In the case of valve gear discussions that's probably best reserved for discussing the peculiarity that means that 50% of the piston stroke is not 50% of a wheel revolution. What I was talking about was just the relative alignment of piston rods, valve rods and cylinders. Piston rods pretty much must be pointing at the wheel axle no matter what everything else is doing or the relationship between piston stroke and wheel movement gets extremely odd. In Stephensons in its simplest form you have the valve rod pointing straight at the wheel axle which has the eccentrics on. So with slide valves between the cylinders the piston and valve rods would all be in the same plane viewed from the side. If with Stephensons you can't have the slide valve between the cylinders and all the rods in the same plane (usually lack of space) there are two commonly seen arrangements. The Stroudley arrangement, which normally has the valves below the cylinders, has the valves level, and the cylinders inclined by (say) 6 degrees, so valve and piston rods are not parallel. But both are still pointing directly at the axle centre. The indirect arrangement, as used on the Churchward standards, has the valves above the cylinders and valve and piston rods all parallel and level, but although the piston rods are pointing straight at the axle the valve rods are pointing at a point a good bit above. So in order to have the valve gear working properly the link and eccentrics are effectively inclined (lets say 10 degrees for sake of argument), and drive a lever on a shaft, and the lever at the other end of the shaft can be rotated by 10 degrees to drive the valve rod. With Walschaerts there's not the inherent symmetry from two eccentrics. And because there's just one eccentric driving the bottom of the link, the valve rod doesn't need to be aligned to the axle centre, and the way its normally laid out the valve and piston rods cannot be in the same plane. Does that help? Maybe this sketch will help too. Most of them are drawn from GW locomotives, but I altered the 15xx Walschaerts for clarity, and the slide valve between cylinders drawing is made up, especially the top view!