Piston vs. Caprotti Valves 1

Piston vs. Caprotti Valves – The Final Discussion? Part 1


The following discussion is a response by David Wardale dated 3rd Sept 2009 to several observations and comments put to him by John Duncan in support of the adoption of British Caprotti valve and valve gear on the 5AT locomotive.

Wardale’s responses are published on this website at his request.  He presents 28 responses, each being preceded by the individual comment to which it relates.

Note: Both John Duncan’s original comments and David Wardale’s responses have been transcribed from image scans into digital format.  Every effort has been made to ensure correct transcription.  They are available on this and six subsequent pages or can be downloaded in PDF format under the title: Wardale responses to Caprotti proposals 3 Sept 2009.


Note: John Duncan’s comments 1 to 7 (below) relate to a statement by Wardale that appears on this website, saying:

“Caprotti valve gear is expensive specialist equipment, so piston valves and Walschaerts valve gear have been chosen because with Porta refinements they give a performance of the same standard and at much lower capital cost. Indicator diagrams from the “Red Devil” – which the 5AT will greatly improve on – have proved that. Can Mr. Attewell produce figures to show that poppet valve gear requires “considerably less” power to drive than Porta-type lightweight piston valves driven by Walschaerts gear running in needle roller bearings?”


John Duncan’s Comment 1

Cost:-  In the 1957 build BR Std. ‘5’ Walschaerts valve gear 73165-71 cost £25,606 per loco.  In the 1957 build BR Std. ‘5’ British Caprotti valve gear 73146-54 cost £28,469 per loco.  Difference in capital cost of £2,863 per locomotive, a small amount of capital cost (10%) compared to the cost of mileage exams on a piston-valve fitted locomotive.

Wardale’s Response:  The BR5 increase in capital cost for Caprotti valves is 11.2% from the cost figures you give, not 10%.  If this were replicated for an actual 5AT construction cost of, say, £3 million, it would amount to £336,000, not a “small amount.”  However it is not possible to say what the relative costs for the 5AT would be. In the case of the BR5, BR workshops were fully tooled up for producing piston valves and Walschaerts gear, whereas relative to this Caprotti gear was specialist (see point (3) below), and the engines so fitted had to have different cylinders, probably quite an additional expense. For the 5AT everything has to be designed and produced from scratch, the piston valves and Walschaerts gear are more sophisticated than in BR times, and everything associated with steam locomotives tends to be in the nature of specialist equipment now. Hence the relative cost is an open question. The only way to settle it would be to make 100% detailed designs for both and have them costed, not an exercise likely to be done.


John Duncan’s Comment 2

The British Caprotti valve gear ran from General Overhaul to the next.

Wardale’s Response:  “British Caprotti valve gear ran from [one] general overhaul to the next.” On average the BR5s on the LMR ran some 189,000 miles (302,000 km) between general overhauls from data given by Cox for 1957 (a figure, incidentally, easily beaten by the Southern West Countries which were then almost all unrebuilt – so much for commonly held perceptions!)  From page 241 of The Red Devil you will see that the piston valves, rings, spindles and valve liners of that locomotive required attention only at 250,000 km intervals. Other things being equal, the distance travelled by a piston valve, and hence its wear, per km is inversely proportional to the coupled wheel diameter. Adjusting the above figure for the larger wheel size of a BR5 vs. the 26 Class (74″ vs. 60″) gives 308,000 km. Therefore we have already shown that Porta type piston valves would also go from one general overhaul to the next without attention. But this does not allow for the 26 Class figures being depressed because of incorrect materials being used for certain valve components, intermittent lubrication starvation, and a persistent foaming/priming problem, the scourge of lubrication. All this is documented on page 240 of The Red Devil. Without these, the periods between attention would be longer still. The same period – or longer – applies to the valve gear, with all Walschaerts pivots on roller bearings and mechanical lubrication of the expansion link – dieblock rubbing surfaces (a standard fitting on the SAR 25 Class). Therefore your inference that any higher capital cost of Caprotti gear would be saved by reduced maintenance charges is incorrect. The reason for this is that you are basing your argument on the piston valve design prevalent on BR in the 1950’s. But, as you know, the 5AT will not have such valves, therefore any argument based on them is not valid


John Duncan’s Comment 3

British Caprotti valve gear was not specialist equipment.  Camboxes and the poppet valves were standard. The cylinders could fit on the left or right hand sides. The drive shafts were standard Hardy-Spicer carden shafts with standard universal joints

Wardale’s Response:  Compared to piston valves and Walschaerts gear, Caprotti valves were specialised equipment. Saying the equipment was “standard” is not the point. By “specialised” is meant requiring an engineering effort to produce it that was higher (i.e. more sophisticated) than that required for piston valves because of greater precision, special materials, specialist manufacture in the case of the cams, etc. In this sense ‘standard’ components could be specialised.


John Duncan’s Comment 4

BR standard ‘5’ with British Caprotti valve gear took 1.87 HP in full forward gear to drive the valve gear on both cylinders (BR test results on 73154).

Wardale’s Response:  ‘BR5 Caprotti gear took 1.87 hp to drive in full forward gear’. Power being a function of speed, this figure is meaningless without giving the speed at which it was measured: full forward gear suggests low speed, in which case the figure would not be valid for normal running speeds. See also point (5) following.


John Duncan’s Comment 5

I do not have the calculations for the proposed Porta-type piston valves for the 5AT locomotive.  As described, the piston valves might be lightweight, the drag on the valve must be enormous with eight valve heads at 175 mm (6.9″) diameter piston valves at 75 mph, 5.69 revolutions per second, on the 1880 mm (6′ 2″) diameter driving wheels.

Wardale’s Response:  “The drag on the [Porta-type piston valves] must be enormous.” False. The drag on this design of piston valve is actually less than on former designs: to understand this you have to read and understand Porta’s papers on the subject – FDC 4 Refs. (4) and (11). If the drag on the four valves of such classes as the Castles, Kings, Lord Nelsons and Stanier Pacifics were “enormous” how could these locomotives, all of which had less cylinder power at high speed than the 5AT, have attained high speeds, which all did? If the drag on Porta type valves were high it means friction would he high, and the extremely low wear rates leading to the extended attention intervals given in (2) above would not have been achieved. In fact the 5AT power loss due to valve ring friction has been calculated: at maximum drawbar power, which occurs at some 26% cut-off and 113 km/h (71 mph), it is 4.9 kW (6.6 HP) for all valves combined (FDC 4 (133)). This is only 0.2% of the cylinder power the 5AT would be developing at these conditions (2380 kW (FDC 1.3.F (17)) and is not “enormous”.


John Duncan’s Comment 6

The piston valves are not supported by a valve spindle with bushes at either end. It reminds me of the connection and floating lightweight piston valves on the unrebuilt Bullied Pacifics with wear in the connecting pins resulting in valve over travel.

Wardale’s Response:  The 5AT valve design should not be compared with Bulleid’s valves. They are not the same thing. The valves rest on the liners without tail rods to give virtually 100% steam tightness – impossible with a rigid spindle and bushes – yet at the same time achieve the very low wear rates already indicated (based on actual experience with 3450). The quasi-indefinite maintaining of steam-tightness (as measured on 3450) is the reason why these valves have such a long period between attention. The only pins not on roller bearings (which have effectively zero wear) are those coupling the valves to their spindles. This type of pin was used on the Rio Turbio 2-10-2s and on 3450, and on these locomotives gave negligible wear.


John Duncan’s Comment 7

The proposed 5AT piston valve, 4 per locomotive 13 piston valve rings per valve head. 104 piston valve rings per locomotive. How secure is the pin in the fork end that cannot be seen?

Wardale’s Response:  “How secure is the pin in the fork-end that cannot be seen?” Obtaining perfect security – such as that achieved with countless millions of i.e. engine gudgeon pins, a similar component – is a matter of correct detail design, which shows the importance of the latter. The issue here is not so much mechanical security, which in this case is relatively easy to obtain, but designing to ensure lubrication by oil in the exhaust steam and at the same time to negate the possibility / effects of carbon deposit between the pin and bush. This is one item where R & D needs to be done as part of detail design, using a simple test rig simulating valve spindle movement and loading and exhaust steam conditions.


See next page for Comment 8 to 10