Piston vs. Caprotti Valves 4

Piston vs. Caprotti Valves – The Final Discussion? Part 4


See previous page for Wardale’s responses to John Duncan’s Comments 11 to 16.
See next page for Wardale’s responses to John Duncan’s Comments 21 to 24.

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 Comments # 17

E.S. Cox states in his book ‘British Railways Standard Steam Locomotives’ page 112: ‘the ultimate British Caprotti system was thoroughly workman like, and its extended use would have been practically certain had steam continued’.

R.C. Bond states in his book “A Lifetime with Locomotive” page 246: “These thirty engines, one of which, 73154, was the last steam locomotive to be built at Derby, turned out very well. There is, I think, little doubt that had steam continued longer, poppet valves would have supplanted piston valves, at any rate, on higher mileage locomotives”.

Wardale’s Response:  Quotes from Cox and Bond: irrelevant, as they are based on piston valves as they knew them, not Porta valves.


John Duncan’s Comments # 18 – Quoting Wardale: “Less than 100% certainty about steam tightness and good flow coefficients past double beat poppet valves”.

British Caprotti poppet valves and their cages have two seats, one flat and one tapered, so the valve cage and double seat poppet valve expand at the same rate always maintaining their steam tightness. When open steam can pass round the outside and through the hollow inside of the poppet valve, giving greater opening for the steam flow than the piston valve and no ‘wiredrawing’ on valve closure.

  1. Caprotti poppet valve (Steam passes through the middle and outside when open).
  2. Caprotti valve cage (Expands equally with the poppet valve).
  3. Sulzer diesel engine poppet valve.
  4. Rover car petrol engine poppet valve.

Wardale’s Response:  Steam tightness. That Caprotti valves may have solved the suspect steam tightness issue by the design shown is tentatively accepted, although this needs confirmation by tests (as made on SAR 25NC and 26 classes, see The Red Devil pages 316-321: such a test could presumably be carried out on 71000 – why not investigate this?) This being the case, it no more than brings poppet valves up to the standard of Porta piston valves. Very long lap Porta-type piston valves as specified for the SAT will minimize ‘wiredrawing’, which is present to some extent with any valve, whether slide, piston, or poppet, see point (20) below. In this regard an independent observer has made the point that at high speed and short cut-off the indicator diagrams produced by 3450, as given in The Red Devil, are as good as those of 71000, showing the thermodynamic equivalence of the two, as noted by Chapelon in your quote. And the 5AT will be (much) better than 3450 in this regard.


John Duncan’s Comment 19:  Quoting Wardale: The fact that the piston valve for the 5AT will in all respects (e.g. steam flow, lightness, lubrication, wear, resistance to high steam temperatures, resistance to steam leakage) greatly superior to the general level of piston valves in the past with which Caprotti valves have been compared;

The list above under e.g. is all the advantages of British Caprotti poppet valve gear has over piston valves and more. Even with the advantages proposed by the 5AT piston valves might be lighter in weight but the lubrication required for 104 piston valve rings per locomotive must have a high oil consumption compared to no lubrication on British Caprotti poppet valves, which remain steam tight due to the unique system of the valve seats being in a cage and the poppet valve having two seats one flat and one coned. The valve and cage expand and contact together, remaining steam tight overhaul to overhaul.

Wardale’s Response:  Valve lubrication requirements do not depend on the number of valve rings. There would be no difference for 2 or 20 rings per valve. No lubrication for the valves may have a detrimental effect on cylinder liner and piston wear, see (15) earlier. The Caprotti valves themselves have no lubrication, but the camboxes do, and in fact Figs. 105 and 106 in my copy of Locomotive Management, illustrating the arrangement of Caprotti gear, show a number of lubrication points.


John Duncan’s Comment 20:  Other advantages are: –

  1. Resistant to very high steam temperatures.
  2. Allow steam to flow freely through the middle and outside of the poppet valve.
  3. Inlet and exhaust events are separate.
  4. Optimum valve events can be designed into the four cams in sealed cam boxes.
  5. Complete bypass when the regulator is shut, allowing free running.
  6. There is no alteration in valve timing by the movement of wheels and axleboxes in horn ways.
  7. Inlet valves are smaller than exhaust valves giving free flow and reduced back pressure which cannot be done with piston valves. (71000’s inlet valves are 6.25″ diameter & 0.796875″ lift, exhaust valves are 7″ diameter & 1.078125″ lift).

Wardale’s Response:  Other advantages:

  1. Resistant to very high steam temperatures: no advantage over piston valves with cooled liners. It is true that no cooling needs to be provided with poppet valves, but it is simple to manufacture with no moving parts and hence gives negligible capital and maintenance cost increase.
  2. Irrelevant as an advantage, it is merely a feature.
  3. Inlet and exhaust events are separate. True, but what does this really mean in terms of higher cylinder efficiency? This must be quantified for an advantage to be claimed. To repeat from (18) above, the indicator diagrams produced by 3450 have been judged to be as good as those of 71000, and the 5AT will be much better.
  4. Irrelevant as an advantage, it is merely a feature – we might just as well say that optimum valve events can be designed into the Walschaerts gear.
  5. As explained in point (11) earlier, this is no advantage for the 5AT.
  6. True, but what effect does this really have on Walschaerts gear – with poorly laid track at depots there may be some (barely) measurable effect, but with near-perfect high-speed track – which is where the 5AT would operate – it will be negligible. Therefore this would be an advantage on paper only, and in any case you must quantify the actual improvement in cylinder performance before you can claim any advantage for Caprotti gear.
  7. There is no net advantage of poppet valves in this regard. It is a major function of piston valve lap to give greatly increased flow past the valves to exhaust steam compared to inlet steam, compare The Red Devil Figs. 48 and 50. On poppet valves the larger exhaust valves with higher lift merely give the same effect, but to a lesser degree. The question of valve opening merits some investigation. From the data you give for 71000 (and it is interesting to see that valve lift can be specified to one millionth of an inch!), the nominal valve openings at full lift, for the upper and lower seats combined, are 201.9 cm2 per valve for the inlet valves and 305.9 cm2 for the exhaust valves, with no allowance for the flow obstruction caused by the valve cage bridge bars. But these are clearly not the true limiting openings concerned. To go to or from the upper seat steam must pass through the annular space at the centre of the valve, and that passing the lower seat must go through the annular gap between the valve body and the cage, both of which are smaller than the respective nominal openings: this is clear from the valve on the right of your figure. Scaling the various diameters from your figure, and allowing a 10% reduction in the central flow area due to the valve ribs, gives the following minimum combined flow areas to/from the top and bottom seats, per valve, which are the effective valve openings at full lift (nominal maximum opening): inlet valves ≈ 147 cm2, (27% less than nominal opening), exhaust valves ≈ 185 cm2, (40% less than nominal opening). Note that these are independent of valve lift, which is why the exhaust opening is reduced the most – the extra lift of the exhaust valves is of no advantage to the minimum flow area. The moral is that a large nominal opening is wasted if the steam flow to or from that opening is restricted, which is the case. Added to this is the possibility of a relatively poor overall flow coefficient past the valves due to somewhat indirect steam passages with some sharp-edges. These figures may be compared with those for the 5AT, calculated from the FDCs, particularly Appendix 1 of FDC 5 giving the valve events. Selected maximum admission openings through the valve liner ports per cylinder (i.e. for two valves, average of front port and back port) are: 75% cut-off (full forward gear) = 491 cm2, 30% cut-off ≈ 134 cm2, 20% cut-off ≈ 94 cm2, 10% cut-off ≈ 65 cm2. Maximum exhaust openings: in mid gear (4.8% cut-off) ≈ 442 cm2, rising to 491 cm2 by 25% cut-off and remaining constant thereafter up to full gear (as the valve fully uncovers the liner port). This analysis is exploratory and approximate as it does not consider all factors involved, which have pros and cons for both piston and poppet valves, but it does allow qualitative conclusions to be drawn. It shows that the maximum exhaust opening of the 5AT’s piston valves is (considerably) higher than that of 71000 at all cut-offs. The inlet valve opening is also higher down to somewhat above 30% cut-off.  Thereafter the advantage lies with the poppet valves. The extent of this advantage at very low cut-offs, where the cam profiles may not give full lift of the valves, making the nominal opening less than 147cm2, cannot be ascertained. All-round the advantage lies with the piston valves. The exhaust opening is most important, because it is harder to exhaust steam quickly than to admit it, due to its low pressure. (This does show, incidentally, that any tendency of Caprotti valves to give high peak draught will be accentuated on the 5AT. which would be very destructive without features such as the Kordina, oil firing, and special attention to brick arch security.) During inlet, it is hardest to minimize the admission triangular loss on the indicator diagram when cutting-off around the middle of the stroke, where piston velocity is highest, and here the advantage also lies with the piston valves, which will, for example, assist in good acceleration through the mid cut-off range. Although the piston valve admission area is less at low cut-offs, (1) by definition the quantity of steam to be admitted per stroke is then low and (2) the piston is moving relatively slowly, so that keeping full pressure on the piston head is easier than around mid-stroke. Which means that the lower valve opening area here is not such a disadvantage.

See previous page for Wardale’s responses to John Duncan’s Comments 11 to 16.
See next page for Wardale’s responses to John Duncan’s Comments 21 to 24.