LMS 2P 4-4-0

It's the kind of thing I can imagine having been set to draw when I did O-Level Tech Drawing; Excel makes it all so much easier!

Tom

 

I sometimes wistfully muse upon what some of the great designers would've done, given access to modern CAD software (and the hardware to go with it, of course!) .... then I think of the John Chester Cravens and Stephen Lewins of yore!

 

For engines without a tail-rod, the piston thrusts will be different in the two directions, due to the different effective piston areas. When running forward, the effect of this on the overall turning moment should be the opposite of the effect described above.

I've no idea whether the offset was intentional for this reason, or what the relative magnitudes of the two effects would be.

 

JCC: 'Where is the flogging members of your family tab?'

 

It's interesting to note when piston tail rods went out of favour. You have to suppose that, had there been any real issues, getting shot of 'em wouldn't have rolled out across all companies as actually happened. How much of that decision came down to any (presumably) marginal differences being outweighed by savings in maintenance of an extra set of glands I've no clue.

 

I am sure you had fun doing that. The only source I have found for valve gear fundamental principles is in Chinese textbooks, I don't know whether there are any decent English language books on this. One of the Chinese books (appropriately entitled Steam Loco Knowledge) has armfuls of equations. which are a bit off-putting. With a bit or persistence however, I think their equation for the movement x of the piston/crosshead moving rearwards from the front extreme, is:
x = R(1-cosӨ) + ((R squared/2L) x sin squared Ө)
where R is the crankpin throw, Ө is the angle of the crankpin away from the horizontal, and L is the length of the connecting rod. The second part of the equation is the "error" caused by angularity (and at Ө = 0 or 180 degrees, the value of course is 0).

Another book, "Walschaerts Valve Gear", which covers similar ground although not quite so equation heavy and covering also the dimensions for each component of the various classes (wouldn't it be wonderful if there was a similar book for UK locos so you could easily look up e.g. the lap and lead for various classes?), has the attached diagram:

where CG and DG is the connecting rod with the crank perpendicular/at maximum angularity, G is the crosshead gudgeon pin, shown just to the rear of the cylinder centre, the distance being equal to OH, the difference due to the angularity. It notes that this difference increases or reduces the cut-off depending which way the piston is moving.

It is not clear that in designing the valve gear, anything was done about this, possibly because it wasn't considered material, or it is outside the scope of the book, or there was not much that could be done. The attached extract from the Fowler 2-6-4T valve gear drawing showing inter alia the cut-off for forward and rear strokes (for forward gear) suggests that it may well be that nothing was done (and based on this data, as the cut-off shortens, the discrepancy tends to reduce).

 

Blimey, I have enough trouble translating from German! Let's not go into Chinese!

 

I'm sure that there are many people who think the whole theory of valve gears is Chinese. Or double Dutch.

 

Correct!

 

Eric Langridge mentions they had a valve gear model at Derby D.O. to work out the valve events. It would be interesting to know how it worked.

 

The inequalities of the piston travel also apply to valve travel which effectively doubles the problem. Valve gear designers come up with all sorts of clever ideas to minimise the effect of the fundamental problems of converting rotary motion into reciprocating motion and give the desired valve events, such as introducing back set on Walschaerts expansion links and non-symmetrical rocking levers on link motions. I have a drawing somewhere that shows the variation between port openings at front and rear for a simple eccentrically driven valve but I've hunted high and low and can't find it.

 

I believe that the valve gear model used at Derby is on display at the Derby industrial museum.

 

I gather that the late Don Ashton's book on Walschaerts and Stephensons' valve gear is still available from Camden books.

 

I recall Don Ashton discussing with me over our regular telephone conversations the details of the 47XX new build (The Night Owl) when he was working up some of the details for Llangollen - I recall we discussed the centre line of the motion - the above the centre line called for stepped eccentric keys on the axle on these types compared to 'inline' GWR types of the Churchward standard 2 cylinder locos with of course Stephensons gear.

Cheers,

Julian

 

If that's the Silk Mill Museum, I think I saw it there some years back. Here's hoping it's still there when they reopen.
Pat

 

Interesting discussion above about connecting rod angularity and related matters, albeit way off the thread topic and hitting the limits of my technical understanding!

In the book by David Hamilton on Caledonian Locomotives of earlier years, attention is drawn to the asymmetric cylinder/ valve designs of many Connor-era engines, with the mid-points of the valve chests being well behind the mid-points of the cylinders. See attachment. The author comments:

"This meant that the steam passage from the front end of the steam chest to the front end of the cylinder was considerably longer than the corresponding passage from the rear steam port to the rear end of the cylinder. Due to the angularity of the connecting rod, a larger amount of steam was normally cut-off during the backward stroke of the piston than during the forward stroke, so the greater length of the steam passage to the front end of the cylinder would be expected to increase the disparity. Dugald Drummond adopted the opposite course by putting the mid-length of the valve chest a little ahead of the mid-point of the cylinder."

I don't think that I've ever seen such cylinder asymmetry described for any other designs. Were these practices unique to the Caledonian, or found more widely?

 

This all leads Me to think about the 'riding quality' of the B17 class, a combination of monoblock cylinder, divided drive, unequal length connecting rods and Gresley/Holcroft valve gear should lead to the middle cylinder effectively having different valve events to the outside cylinders, (It would actually be different 'Piston Events' with respect to the valves) which at the very least would lead to the loco 'swaying about'. Maybe a 'Peppercorn Rebuild' with three sets of valve gear would have improved matters.

 

I believe that A. E. English of the LNER identified the root cause of the riding issues of the B17. English had the reputation of being a gifted mathematician, he looked at the centre of mass of the locomotive in conjunction with the axle load limits of the design. The centre given by looking at the axle loadings was not coincident with the true centre. There was a destabilising couple.

 

I agree that there are likely to be some issues with the component of valve movement arising from the return crank because of angularity. (I see the author of the attached paper refers to "pesky irregularities" - but ignores them http://dougkerr.net/Pumpkin/articles/Loco_Valves.pdf ). It would be a complicated analysis to determine the effect of this coupled with the discrepancy of the piston movement. Re backset of the expansion link drive pin, my understanding was that the position of the drive pin is determined by the requirements that (a) there is an equal fore and aft angular swing, and (b) the expansion link is perpendicular to the motion axis when the main crankpin is at fore and aft dead centre (which also determines the angle between the main crank and return crank). Can you interfere with these principles?

 

'A destabilising couple' which is essentialy an out of balance force, as one would get with divided drive and unequal length rods and 'incompatable' valve events, the centre of mass and axle load limits add to the problem.