Edward Thompson: Wartime C.M.E. Discussion

Thanks Simon.

 

Castle drawing:

http://www.greatwestern.org.uk/drawings/loco/loco241.jpg

There is no measurable difference between inside and outside rods.
This makes sense as there is only one common inside valve gear to each side.
Short conrods give difference between fore and aft possitions of pistons/cranks in valve movements.
This can somewhat be compensated on drawing board if rods are equal short.Else not so easy.

 

There seems to some considerable misunderstanding of valve gears here.

The GWR Stars (excepting one with scissors gear), The Great Bear, the Castles, and Kings all had inside gear for the 2 inside cylinders that then reversed via cranked rocker arms to operate the outside cylinders 180 degrees out of phase.

What has been described by some above is inequality of angularity of the connecting rods if not all the connecting rods are the same length. On the GWR 4 cylindered locos, this inequality of angularity was actually the worst possible scenario because the direction of travel was reversed.

We know W H Pearce (the GWR valve gear guru) came up with a simple solution to this, though the results were not perfect, though this was masked somewhat by the GWR valve setting so that the timing for the exhaust release was equalised for inside and outside cylinders.

Personally, I consider that a 3 cylinder loco with 3 independant sets of valve gear and the inside cylinder having a different connecting rod length, ought to produce valve gear events superior to the GWR Stars, The Great Bear, the Castles, and Kings.

You can these days play around with all this on a valve gear computer simulator, and see which companies understood valve gear correctly, and which ones did not!

(With all due acknowledgements to my old friend Don Ashton)

Cheers,

Julian

 

Julian,

Both from Locomotive Adventure Volume 1. Page 149 for the Civil Engineer's objection to the original scheme for the Z with the N1 prototype cylinders and valve gear, on grounds that it was more than permitted relative to the fouling bar clearances at facing points (I suspect it's that a Z to this scheme coupled to something else with a long overhang could straddle a fouling bar and allow it to release, which is really quite some overhang).

Page 152 for the comments on the Schools. Regarding T.S. Finlayson, Chief Draughtsman at Eastleigh, in relation to the design of the Schools: "We knew he disliked Belpaires, taper boilers, circular smokeboxes and conjugated valve gears. However, he got the better of Clayton..."

Regarding the actual design: "A third valve gear was adopted for the inside cylinder but it necessitated a second reversing shaft linked to the main one. A conjugated valve gear would have provided a lighter and cheaper alternative."

 

When did the fix to the big end strap proposed by Harvey and Spencer come into general use (I think Harvey had seen something similar on a German design)? I know Cook did a lot of work on the brasses, metalling and accuracy of setup in the '50s, but I've always reckoned that was a life improvement, rather than dealing with the fundamental lack of stiffness in of the original design. The high-speed run problems with relatively new big ends (and conjugated gear) are more indicative of the lack of stiffness leading to rapid failure under heavy loads, in my opinion, whereas the Cook modifications seem targeted at slowing gradual degradation in service.

 

It is perhaps notable that Britain's railway companies seem to have worked through just about every possible permutation of cylinder placement on 3-cylinder and 4-cylinder engines. The option that "almost got away" was 4-cylinder drive concentrated onto the second coupled axle. But Sir Nigel had rebuilt Ivatt Atlantic 279 into that arrangement in 1915, in preparation for his planned Pacific - which was delayed by WW1 and then evolved substantially before emerging as 1470.

Ref your Jarvis comment, a large number of Bulleid pacifics have been preserved. so we can probably spare one to go into apple green to represent the imaginary Peppercorn light pacific for the Great Eastern. Whoops!!! An alert is already flashing on the duty officer desk at Livery Police HQ.

 

The Schools with Belpaire firebox and conjugated valve gear did get built in the end, only in 1/16 scale. It was built at Ashford and named "Holcroft".

 

That one might even have fitted the Hastings Line Tunnels (though possibly not if scaled up to full size!)

Tom

 

I think it would have given a Bulleid pacific a run for its money on a Bournemouth fast. To add a little more information it was built in the 1960s by Roy Donaldson who was a senior engineer in Ashford work. I believe H A Holcroft had an input to the design as he lived in Hamstreet not far from Ashford at one time.

 

I think that everyone would agree that the P8 was Mr Garbe's outstanding contribution to steam locomotive history, in the same way as the B1 would later become Mr Thompson's major success.

You go on to refer to the Prussian S10 express passenger 4-6-0 which, with its 4 variants, was clearly less successful. I had only heard of them through brief references by British writers, but there is quite a good article about them on English Wikipedia.

The early British 4-6-0s, built prior to the 1923 Grouping, were a very mixed lot - some very good, some poor. Many engineers had much difficulty designing a satisfactory fire-box/grate/ash-pan layout around the rear coupled axle. Life could be difficult for the fireman.

One of the most successful of these 4-6-0s was also one of the earliest - the Highland Railway "Castle Class" of 1900. A picture is attached. This design had the rear axle set well back - the same solution used in both the P8 and the B1.

 

4-cylinder engines with divided drive and equal-length connecting rods were obviously very common - all the GWR ones plus the large numbers of 4-cylinder engines in France and elsewhere following the de Glehn model.

Perhaps we should highlight a particular feature of the GWR engines, namely their very wide cylinder spacing (even wider than the already wide spacing on the 2-cylinder engines). This was because the cylinders were alongside the bogie wheels, which needed space to move sideways. Given the limited width of the British loading gauge, even on the GWR, cylinder size was severely constrained and the Castle Class 16-inch width was just about the limit. Some clever work, including a weird bogie design, managed to squeeze 16¼-inch cylinders into the King.

For Thompson's Pacifics, with 19-inch or 20-inch cylinders, the GWR solution was not an option. They would have crashed into station platform edges. If Thompson wanted divided drive with equal connecting rods, he had no choice but to move the bogie forward, lengthen the wheelbase, and face the accusations that the layout was ugly.

In spite of being so common for 4-cylinder engines, this layout does indeed appear very rare for 3-cylinders. I think the Webb 3-cylinder passenger engines did have equal length con-rods, but perhaps an LNWR expert could check this out (But note that the Webb 0-8-0s had concentrated drive onto the second axle- cf. Gresley O2 2-8-0s). I am indebted to the book "Compound Locomotives" by JT van Riemsdijk for identifying one further example from Switzerland - a 3-cylinder compound 2-6-0. Details in the attachment.

 

Thank you those who have put me right about the Great Western 4-cylinder locos. I had assumed, wrongly, that even the De Glehn layout didn't keep the lengths exactly the same.

I believe that "simple" solution is the bent shape of the levers that transmit the valve motion from the inside to the outside. Is that represented in one of the simulators?

 

I don't believe that any fix did come into general use - I can see there were variations done under Gresley and Thompson but by Peppercorn's appointment the drawing office had moved onto new designs firmly.

Which is pretty much what the Cox report said was the issue. I don't know of any further Cook modifications beyond that - can anyone shed any light?

 

Why?

 

I'm a long way away from a copy right now, but Harvey's 60 Years in Steam has some of the details. The stiffening web on the big end straps of preserved LNER 3-cylinder engines wasn't originally present; they looked very much like an IC engine big end, which transmits vastly greater forces with the rod in compression than in tension. In a steam engine, this puts excessive forces in the strap, leading to flexing and breakup of the brasses.

Immediately post-war (I think) Harvey and Spencer were tasked with looking at simple modifications to improve availability. They identified this weakness from the way the brasses were failing and Harvey proposed a stiffening web on the back of the strap, to improve the section modulus and reduce bending, which was implemented prior to the Cook modifications. He'd seen something similar on a German design, which had a crescent shaped web for optimum stiffness distribution, but the drawing office did a simple constant section web. Probably good enough, but less elegant.

I can't tell you which CME this was under, but it should be easy enough to work it out from the book.

 

Cook, on the other hand, says that improvements in manufacture and lubrication were the key factors.

 

A little from column 'A' and a little from column 'B' perhaps? Keep in mind the many significant across the board war-driven developments in metallurgy and processes around this time.

 

Very tolerant thread.
Has moved in next to no time from Gresley/Thompson to King / Royal Scott.
Cox and me support Thompson and Scott

 

The inside big ends on Baureihe 44 did rev longer ,faster and harder than most on LNER.
Smart move to copy something that works.

 

I've not read Cook (although he is on the list), so I can't comment on his exact reasoning behind that thinking. Institutional memory can be surprisingly short about detail changes, especially if the item is still causing problems, so it's possible he was unaware of the earlier modifications. Clearly Cook's work was still necessary, or he wouldn't have done it, but the earlier change was in all probability also necessary. You eliminate the issues with a design until it works properly, but it doesn't mean the last one you eliminated was the only one that mattered.