Piston vs. Caprotti Valves 7

Piston vs. Caprotti Valves – The Final Discussion? Part 7

See previous page for Wardale’s responses to John Duncan’s Comment 25.

Note: The full text of this correspondence can be downloaded in PDF format under the title:
Wardale responses to Caprotti proposals 3 Sept 2009.

John Duncan’s Comment 26: Quotation from Locomotive Post (date not legible)


The articles on the British-Caprotti Valve Gear which have appeared in this paper have given valuable information on the parts and layout out of the gear, but possibly many are wondering how it performs under working conditions. As I have had the good fortune to work one of the L.M. Class 5 engines fitted with this gear, my experience may be of interest.

This was the first time I had been on one of these engines, but I found it simple and easy to handle. In fact, throughout the 115 miles run, on rising or falling gradients or on level track, the performance was highly satisfactory. Cut-off is variable between 70 per cent. and 3 per cent, and I found that it really works at 3 per cent.

Starting was carried out cautiously and a trifle apprehensively, but once clear of the station and with the throttle wide open, as per instructions, all doubts were dispelled by a display of remarkable acceleration. Due to the by-pass effect caused by the valves drop­ping off their seats when the throttle is closed, the engine coasts with complete freedom. Full advantage was taken of this to coast over 40 miles in stages between Glasgow and Carlisle. At all times absence of noise was noticeable, but when coasting, even at high speeds, the quiet, smooth running of the en­gine was impressive.  The reverser may be used in the same way as any other except that reversal from one gear to the other is not completed until it has been placed in full back or fore gear, according to which is re­quired. Having done this, any desired cut­off may be selected, and the throttle may be closed and reopened without resetting the gear, except as working conditions demand.

The results obtained with the British ­Caprotti valve gear provide much food for thought and, in my opinion, strengthen the case for working with wide open throttle and the shortest suitable cut-off. Reducing the ad­mission period does not adversely affect the other valve events. Reducing the admission period with other valve gears also reduces the amount by which the port is open. But in the British-Caprotti gear the valve lift re­mains substantially the same at all cut-offs. Consequently on level track, it is possible to accelerate a train using a cut-off as low as [illegible] per cent. This is what can be accomplished by a small quantity of steam admitted to the cylinders at maximum pressure to exert powerful thrust on the pistons at the commencement of the stroke. Another important factor. is that with this gear, exhaust is delayed to obtain the maximum work from the expanding steam.

When an engine is stationary, maximum power is exerted with the crank on top or bottom quarter, but at speed new factors come into operation. Then the steam exerts its force with the greatest effect in the rear part of the stroke, because as it nears mid-stroke, piston speed is reaching a maximum while steam pressure is falling to a minimum This, I think, explains the power and effectiveness of the British-Caprotti gear at such a low cut-off as 3 per cent. Steam passage into and out of the cylinders with comparative freedom and exerts its greatest pressure on the pistons, from the commencement the strobe, and continues to work up to the point of release.

Wardale’s response:  “Caprotti Performance” by P. Fraser adds nothing new. At best it might be taken as an indication that the Caprotti engine was better than a piston valve Class 5. This might be so, but we are concerned here not with 1930’s piston valve design but state-of-the-art Porta type valves, which are a very different proposition. Contrast Fraser with C. J. Allen on 71000 in British Pacific Locomotives: “Some engine crews have been able to make nothing of the engine, and have lost time heavily …” That’s hardly scientific, it is true, but neither are the great majority of comments on the 5AT.

John Duncan’s Comment 27: Excerpt from email from Angus Eickhoff published in Steam Railway issue No 276 Oct 2002 p.38 under the heading “David Wardale blueprint needs to be even more radical”.

“Mr Wardale is rightly concerned about hammer-blow on a two-cylinder machine. Astonishing therefore that the engine has been drawn with ordinary Walschaerts valve gear, which will contribute considerably to the reciprocating masses. Surely a more sensible alternative would be to use British Caprotti valve gear, which has no reciprocating mass at all, and gave good results?

Probably the strongest argument for rotary cam valve gear on such an experimental machine is that in order to change completely the characteristics of the valve events, all you have to do is insert a different set of cams.

Certainly with regard to valve gear, Chapelon himself was very clear. “Poppet valves are necessary…. on locomotives with very high superheat to avoid the necessity for lubrication of admission ports of piston valves and thereby protect the oil from cracking effects and retain its value as a lubricant.”

Chapelon also outlined the problems with inertia and the amount of space taken up by modern piston valves. He pointed out that in order to achieve the same valve opening as a piston valve with a travel of 200mm (8in), a poppet valve only had to be lifted off its seat by a mere 30mm (1” approx), with a corresponding reduction in inertia forces.”

Wardale’s response:  Quote from Angus Eickhoff (part relevant to valves only).   

  1. Walschaerts valve gear does not ‘contribute considerably’ to the reciprocating masses. Apart from the crosshead arm, union link, and reciprocating part of the combination lever, all very small components, no part of Walschaerts gear is considered when determining the engine’s balancing. The above-mentioned items, with their bearings, contribute 4.9% of the total reciprocating masses to be balanced (FDC 8 (24) and (28)) – hardly ‘considerable’.   
  2. “… such an experimental machine [as the 5AT]…” Is the 5AT experimental? I don’t think that’s likely to improve the possibility of funding. It is not experimental, and there will be no need to “change completely the characteristics of the valve events.”    
  3. Cracking of oil is prevented by (a) delivering oil direct to the valve liners, (b) valve liner cooling and (c) preventing air from entering the steam chests and cylinders when drifting (drifting steam).   
  4. Piston valve inertia and high temperature steam have already been answered.
  5. Space: there is adequate space for the 5AT steam chests and piston valves.

John Duncan’s Comment 28: Excerpt from 5AT Website: unpublished letter to Steam Railway from Bryan Attewell BEng CEng FI MarE in response to David Wardale’s two-part article “Steaming into the Future” published in SR 272 June 2002 and 273 July 2002.

“Moving on to another thorny issue, why has David chosen to use Walschaerts valve gear rather than one of the more modern poppet valve alternatives? Conventional gear links the operation of inlet and exhaust ports in a sub optimal way, while poppet valve gear enables the two events to be independently controlled.

While the Caprotti solution allows for efficient mechanical operation, electrical actuation would enable microprocessor control of the valve events, allowing the locomotive to operate at optimum efficiency over a wide range of conditions. Motorists will know that all modern cars have electronic engine control systems, while many readers will remember that electrically operated valve gear was successfully employed on a class of express Garratt locomotives between the wars.

Further the power required to operate poppet valve gear is considerably less than that needed to drive piston valves. Walschaerts gear is pretty to see and is reliable, but it really has no place on a 21st Century locomotive.”

Wardale’s response:  Re quote from Bryan Attewell (part relevant to valves only).   

  1. Why is a poppet valve alternative ‘more modern’ than what is specified for the 5AT?  
  2. Linked valve events vs. independent ones: what does this actually mean for the desired end result, cylinder performance? Given the essential equivalence of 3450’s and 71000’s indicator diagrams, I would suggest the answer is very little.  
  3. “… electrical operation and microprocessor control …”  This is losing the plot again. For the purpose for which the 5AT has been proposed, it is to be kept as a classical steam locomotive as far as possible, which includes having the driver control it, which he is paid to do.   
  4. Power to drive the valves: already answered.   
  5. “… Walschaerts gear is pretty to see and is reliable…” Thank you.

Wardale’s concluding paragraphs 

That answers all your and others’ comments. It is a fairly definitive answer, and adds to, or supersedes, as the case may be, what may have been said previously on this subject. I think it gives a pretty robust case for the piston valve and valve gear design for the 5AT.  In particular, your primary assertion, that only poppet valves are suitable for high speed, is demonstrably wrong. Full stop.

To summarise, the only clear advantages of Caprotti valves appear to be:

  1. Reduced cylinder oil consumption and 
  2. Larger valve openings for admission at short cut-offs.

The clear advantages of piston valves and Walschaerts gear, as specified for the 5AT, appear to be

  1. Larger valve openings for admission at long to medium cut-offs,
  2. Larger valve openings for exhaust at all cut-offs and
  3. Aesthetically superior.

There are two main reasons why I do not agree in principle with your comments. Firstly, they are based on a comparison between BR (or ‘first generation’) piston valves and Walschaerts gear on the one hand, and Caprotti valves and gear on the other. What you have to say may or may not be valid for such a comparison. But, as should be clear, the piston valves and Walschaerts gear for the 5AT are on an altogether higher level than those of First Generation Steam in general, and BR locomotives in particular, which itself invalidates your arguments when applied to the 5AT design.  Secondly, they are not supported by any engineering calculations.  In general, comments and suggestions are all too often generalized, vague (‘poppet valves are modern …’; ‘2 cylinders are not suitable for high speed…’), not based on a proper engineering understanding, and unsupported by data.  In view of this, it was agreed with Chris Newman some time ago that any comments and suggestions had to be supported by calculations to at least FDC level for them to be considered.  Perhaps not surprisingly, the stream of suggestions dried up.  I have taken some effort to counter your points in this case, but as with all other suggested alternatives to the 5AT specification, they must be supported by adequate engineering data, to FDC standard and specifically related to the 5AT, for a valid comparison to be made.

Well, despite all the evidence in favour of the current 5AT valve design, fixed ideas in the mind being what they are, I don’t expect you to change your opinion. But I don’t expect to hear any more on this subject either!

Best regards


See previous page for Wardale’s responses to John Duncan’s Comment 25.