GWR Saint cylinder design and other cylinder questions.

I wonder if I may ask some questions here regarding Saint class cylinder designs and efficiency? The original Saint cylinders had inside steam pipes of the design fitted to 2999. The performance of these original Saints is legendary. I have read that the outside steam pipe cylinders fitted to the Hall class were not so well designed, (possibly with smaller steam chests?) and not as efficient. From 1930 on, many Saints received new cylinders with outside steam pipes.

My question is. were these outside steam pipe cylinders fitted to many Saints simply Hall class cylinders, making the locos, in effect, large wheeled Halls? Alternatively, were they outside steam pipe cylinders manufactured to the more efficient original Saint design with the larger steam chests (but with outside steam pipes?)

Another question. The new cylinders manufactured for 4930 Hagley Hall appear to have the " bellied-out" centre of the steam chests. Were they always bellied-out or was this a later improved design of cylinder after the success of the new pattern fitted to the Grange class?

Finally. Were the Manor class loco cylinders of the same (improved) design as those fitted to the Granges?

 

I think its probably more complicated than that. There were for example some Halls and I think Saints with 9 in valves, not 10in, but later refitted with 10in valves. Manors and Granges had 9in valves too. Halls as built had longer valve travel than Saints according to RCTS. GWR standardisation was more about parts being interchangeable than identical. Its certainly possible that the first Hall cyclinders manufactured were different to the last. As for modern new cylinders, who knows which pattern or drawing was available to them?

 

I remember many, many years ago when working on 7822 Foxcote Manor at Oswestry, measuring the cylinder bores for new rings. One side was 18", the other 18.5". That was from BR days.

 

I think that the Hall and Saint cylinders may be different, after Saint Martin the boiler was raised on the later Halls, presumably because of the change of wheel diameter This would probably have meant a change in the cylinders as they form the smokebox saddle. Not a huge change.
Rob

 

That raises an interesting point. The Hall smoke box saddle was deeper and it was part of the cylinder casting. The photos of outside steam pipe Saints show that the smoke box saddle was the same as with the inside steam pipe casting, still lower than on the Halls.

Maybe the internal dimensions of the steam chests and ports were the same on both inside and outside steam pipe Saints? Another reason why this might be the case is because there are records of great performances with the later outside steam pipe cylinders.

 

The bore of the later cylinders was nominally 18.5" compared to the 18" of the originals, which may have contributed. (Bearing in mind that this goes out of the window at the first re-bore.)

2999 used an original drawing of an inside steam pipe block, although I can't remember what the date of the drawing was. I can pretty much guarantee that there will have been more than one drawing for the 'same' block! The bore on 2999 has actually been increased to 18.5".

The differing saddle height between a Saint and a Hall is one of the reasons why a new block was made for 2999. The original plan for the project, back in the early 80's, was to use Maindy Hall's original block, and employ some bodgineering in order to lower the saddle, while still keeping the outside steam pipes.

It's worth bearing in mind as well that the port/valve rings on a Saint are narrower than a Hall; while this can be resolved by changing the valve liners, which are really just consumable items, the fact is that the front end of a Saint has never been identical to a Hall. For what it's worth here's a photo of one half of the new block made for 2999:

http://www.thesaintproject.co.uk/Media/RgalEnlarge/CylinderNewlyCast.jpg

 

Yes the right hand side hand a liner and the left didn't. We fitted the left with a liner in 2019, after the bore size got to 18.75. Normal maximum diameter with a liner is 18.5, but you can go larger the first time round with no liner.

I had one particular 'engineer' attempt to tell me that both cylinders should be lined so they came out the same size. I asked him to show me what problem had been caused in over 30 years of running with different sized cylinder bored. That stumped him.

 

I was told that on a multi cylinder internal combustion engine the differing piston weights throws the balancing out, increasing crankshaft bearing wear. I'm not sure how well that would actually stand up if you did the mathematics (which is beyond me). On a steam locomotive that doesn't generally have full balancing of the reciprocating parts anyway it doesn't seem obvious that would be a major consideration, but I'm happy to be corrected.

 

I think that it concerns differential thrusts. The increase in bore size might be small, but it's on the circumference so has the maximum effect. The thrust is the product of steam pressure and the square of the bore, so could be significant, especially where the cylinders are large to start with, leading to torque on the crank axle and a tendency for, say, a small 0-4-0 to crab on the track.

That's what engineering logic says. In practice, reboring cylinders to different sizes has been going on for nigh on two centuries and it doesn't seem to happen that way!

 

It's common enough that one end of one cylinder does more work than the others because of limitations of the valve gear design and/or errors in valve setting. Within limits it doesn't seem to matter much. The same should go for a difference in diameter.

Hasn't 6024 been fitted with slightly smaller outside cylinders, while presumably keeping the original inside ones?

 

I think in theory, unequal cylinder diameters in a two cylinder loco might lead to frame twisting due to unequal forces developed each side. Set against that of course, the forces are out of phase anyway, so there must be a degree of cyclical twisting on a two cylinder locomotive; the impact of two cylinders of different sizes would therefore be just to make that existing twist slightly less even.

Also probably worth remembering that the NER (and a few other railways) successfully operated two cylinder Wordsell-Von Borries outside cylinder compounds for many years. However much the designers and valve setters tried, it is hard to imagine that each cylinder developed exactly equal forces. So I suspect a 1/2" or so difference in cylinder diameter on a two cylinder loco is probably more of theoretical than practical concern!

Tom

 

I believe that's correct yes.

We have never noticed an adverse effect from Foxcote's cylinders being slightly different diameters. Running the figures, we have a tractive effort of 28102lbf on the right side and 27338lbf on the left. So 2.8% higher on the right. However I'm quite sure you could get your valve setting out by the same amount, or the porting could be different by similar, slight misalignment of keyways on the eccentric axle etc..

 

And of course one side of the piston has a rod and mostly the other side doesn't. I think the idea that in a steam engine - an essentially crude device - dire results would occur from an imbalance between sides of the motion is entirely exaggerated.
The One TWO three four beat of the old LNWR G2's suggests some sort of imbalance but never seemed to bother them - and for all of the criticism of the Gresley three cylinder beat rarely did it stop them doing their work. It was a somewhat dubious big end design that was their Achilles heel.

 

The Super Ds' odd beat was a symptom of their Joy's valve gear. It was probably fine when first applied to the engines, but constant new versions and rebuilds increased their weight, which in turn made them sit (slightly) lower on their springs. Joy's gear has a vertical component driven off the connecting rod, and this small change in height would alter the valve events. Ted Talbot goes in depth into the subject in his book on the LNWR Eight-Coupled classes. But it had nothing to do with cylinder dimensions.

 

Something I hadn't really been aware of until helping set the valves on our H class is that the process is akin to trying to use two simultaneous equations to solve for four variables - it can't actually be done.

As an example: suppose you want to set the valves by equalising the port openings at the point the piston is fully forward and fully back. You want to do that with the valve gear in full forward and full back gear. So you have four numbers you are trying to equalise, but you only have two things you can vary: the length of the two eccentric rods. So in practice, adding a shim to change the length of one or other of the eccentric rods will change at least two of the four numbers you are trying to equalise. (And all that before you consider that you are doing the process while cold, but when the valve rod heats up, it will expand, which will affect the port openings more at the front of the valve than at the back ...)

Little wonder the process seems to me more art than science! You can in principle design an excellent valve gear on paper, but as soon as you add some wear, it is perhaps not surprising things start to sound off beat - never mind bouncing the crank axle up and down on uneven track!

Tom

 

Valve setting can get very confusing if you let it.
Not only does changing one thing, like an eccentric rod change another thing, but the amount of change of the other thing can be different depending on what you have done. I'm talking Stephensons's with launch links in particular.
There seems to be a tendency to over think valves and valve events/setting on steam engines amongst some enthusiasts which is unwarranted by the conditions under which they work as Tom says.
Another point is to the casual observer the only thing you can use to judge how good the engine is set up is the exhaust.
While it is obviously part of the valve events it is not as important to the good running of the engine as the other events in the cycle. The person setting the valves will have had accurate knowledge of the admission part of the cycle and the lead imparted to the valve than someone in a field with a video camera.
And for the record I have set the valves on several Stephenson's Link engines. It's not that difficult but you have to have a good idea when to stop chasing small errors in what you are trying to achieve.
I found it one of the most interesting things about working on steam engines.

 

One asymmetry that effects the working to a noticeable degree, and I don't think has been mentioned:
The connecting rod. When the cranks are in the top or bottom position, the piston is well behind mid stroke.
Thus the piston speed is higher, thus giving higher thrust at low speed when it is in the front half of the cylinder. But admitting less steam at high speed, due to shorter opening time.
And vice versa in the rear half.
The shorter the connecting rod, the more noticeable effect. Also this can partly be offset by the valve gear.
But just one more of the imperfection in the art of adjusting the valves and valve gear.

 

Ah, the black art of valve setting. With any conventional loco valve gear the cardinal points of admission, cut off, exhaust and compression are inextricably linked and you can't alter this. It is usual to adopt two different approaches, depending on the type of valve gear. For constant lead gears, such as Walschaerts, the usual way (but not the only way) is to set the valves to give equal lead at both ends of the stroke. This requires the valve to be moveable relative to the valve rod or for the blacksmith to become involved in shortening or lengthening the rod. with a variable lead gear , such as Stephenson's link, it is usual to set the valve to give equal openings for each port. This is usually done in full gear, using shims between the eccentric strap and eccentric rod, as Tom said. Once set for full gear in both forward and reverse, it is checked in the normal running position, which will depend on the locomotive and duty. in this position, both eccentrics are having an effect on valve movement so further further adjustment is made to effect a compromise.
With slide valve acting on a vertical face between the cylinders it is almost impossible to take these measurements directly and you either have to take accurate measurements from a fixed point or use small wooden wedges to establish the openings. Inserting the wedges into the port opening will leave a tell-tale mark on the wedge, which can then be measured.

As I believe that the H class has Stephenson's link valve gear, I'm surprised that it was set up to equalise the port openings at end of stroke, which is effectively setting for equal lead. However, as has been said, valve setting is inevitably a compromise and the different valve events are linked.

 

Ah, I may have got it the positions of the bits of motion wrong way round (it was was a while ago - equal port openings at front and back in full forward and back gear was what we were trying to achieve, I seem to recall). What I distinctly remember was having four numbers you were trying to control, but only two things you could vary to control them. (I also remember an awful lot of pinching the loco forward and back to get the valve in the right place; and the fact that moving from back to forward gear isn't the easiest thing in the world on a cold loco with a steam reverser!)

Tom

 

I was going to say that the worst thing about valve setting is the endless moving the loco back and forth with pinchbars. You also need space to do this. Even a 5 foot driver needs a good twenty foot of space to do it. I can well understand why the main works had powered rollers to enable this to be achieved without moving the loco.
Never thought about the problem of valve setting with a steam reverse.