Seepage in Small Steam Engines
Seepage in Small Steam Engines
by R.G. Neff
The Live Steamer, March-April 1951
The subject of this article is probably an old story to many of the readers of The Live Steamer, but it may prove helpful to others who are building steam models, especially the very small ones, such as O gauge size, and are having difficulty in getting them to run as well as they think they should.
I never tried to use lead on O gauge steamers because it would be only a few thousandths and very hard to set evenly at all four points of steam admission. Instead, I set the valves to admit steam to the cylinders as near as possible to dead center. The models ran nicely, but not quite as slowly as I thought they should and seemed to "buck" slightly at low speeds. By watching the valves carefully, with chamber heads off (note: piston valve type cylinders) I could see there was a slight seepage of steam around them, not enough to hear it, but enough to show small bubbles although the valves were lapped in with kerosene and had only enough clearance to move freely, which would be only 1 or 2 tenths.
I have discussed this matter with several men who know steam engines, including the division master mechanic of D&H RR, and all agree that in any engine there is a slight seepage of steam to the ports before the actual port opening occurs, regardless of how well the valves are fitted. The volume of steam reaching the cylinders is small and has no effect on an engine of any size, but when we get down to O gauge cylinders and the tiny clearance at the end of the stroke, just a little bit of steam getting into them too soon will have the same effect as pre-admission in larger engines.
After figuring this out, I built the next model with a longer steam lap, so that the port opening occurred after dead center, about one spoke of the drivers. As soon as this model limbered up, it showed the result of the change in smoother running at lower speed. Another result that I hadn't looked for was a notable increase in the efficiency of the engine. Previous models had run 2-1/2 to 3 minutes per ounce of water, but this model ran 4 minutes per ounce, under normal operating conditions, pulling 15 cars at a scale speed of 45 to 50 MPH. This in turn meant that a lower fire would keep the pressure on the head. The latter result is due to the absence of bucking, in which the engine has to use up part of its power to overcome the effect of pre-admission.
To me the above results proved that seepage could affect the running of an O gauge steamer, and that an allowance should be made for valve design. It doesn't seem to affect 3/4 inch scale models. I drove Ed Bergh's 3/4 inch Hudson and it ran very smoothly at very low speed. Ed said that he had made the lead in exact proportion to that of the B&A prototype. I wonder if it has any effect on 1/2 inch scale models, but have had no chance to ask any of the builders in this size.
The Live Steamer, May-June 1951
With all due respect to Brother Neff, whose work I very much admire, my experience tells me he is entirely on the wrong track with his piston valve leakage. All he wants to do is to fit an efficient mechanical lubricator which will keep a constant film of oil between the valve bobbins and the liners. This will form a perfect seal - kind of a liquid packet - and there will be no more need to cut the clearances so fine, reducing friction to a minimum. He will also then be able to give his valves the correct amount of lead, which is necessary to get the full pressure of steam on the piston heads at the instant the crank has passed the dead centre point.
The reason his engine with longer laps showed increased efficiency, was not because of the later admission, but the earlier cutoff. If that locomotive had my arrangement of ports, valves and valve setting, fifteen cars would be just so much chicken-feed to it. It would take Brother Neff for a ride quite easily. I have here at the present minute, a coal fired O gauge 4-6-2 which I built in 1925, when everybody said it couldn't be done. At the MRC Exhibition at the Kingsway Hall, London, in that year, she hauled an adult weighing just under 160 pounds, and we kept her in steam for an hour or more and gave a ride to about every kid in the show. She still can do it after 26 years, and the valves have proper leads. She doesn't "buck" when starting, for the simple reason that when the port on one side opens to lead, the port on the other side is wide open and the crank in the position of maximum thrust, which is sufficient to neutralize any tendency to "backfire", even if the steam got to the piston-head in time to exert enough effect, which it doesn't. Everything on this earth takes time, even a flash of lightning, and steam doesn't move quite as fast as that.
See Valves for definitions of lap, lead and cutoff.
The Live Steamer, July-August 1951
I am afraid that LBSC got some wrong impressions from my article on "Seepage" in the March-April Live Steamer. Any competent machinist can fit a round plunger, such as a piston valve, into a round bore closely enough to hold low steam pressure without leaking and yet slide freely. My valves do NOT leak, but there is definitely a seepage into the cylinder that begins when the valve still has several thousandths to travel to the actual opening, which can be easily demonstrated with air pressure and a stethoscope leading into the cylinder. The lead on many slow speed engines in Big Practice is 1/8 inch, which scales down to about 2.5 thousands, so if the seepage begins at 0.005 inch or 0.006 inch we really have pre-admission leakage, it isn't necessary to flood the valves with oil to seal them, as the water seal supplied by the steam serves the purpose and a light oil feed is enough to prevent scratching, etc. Most steam builders are interested only in the locomotive itself, whereas the fellows who buy my models are model railroaders having operating layouts, but preferring steam power to electric dummies. I am sure that most of my models spend a lot of time in slow speed running, doing yard work, climbing grades, etc. and my objecd is to make them run smoothly at low speeds, which they won't do if there is any pre-admission of steam to the cylinders. It is perfectly true as LBSC says, that lead is used to compensate for the lag of steam entering the cylinders, but at low speed this lag becomes negligible, hence the small lead used on a switcher. A high speed passenger engine with a long lead, bucks decidedly when starting its load, as I found when riding on D&H Pacific No. 607 pulling local. But of course it smooths out beautifully when approaching its running speed. If I used pressures ranging from 60 to 100 pounds as most builders do, bucking would not be so evident. But my models develop all the power their weight will hold down with pressures of only 40 to 50 pounds, and it seems more sensible to smooth them up by adding a few thousandths to the steam lap, than by carrying higher pressure not otherwise needed. The few thousandths added to steam lap to get zero or slightly negative lead, adds only a few points to the cutoff, not enough to account for the increased efficiency. So for the above reasons, I'll stand on the facts stated in the article.
There was another slight misunderstanding in LBSC's interpretation of the article. I spoke of a model as pulling a train of 15 cars, which referred to as "chicken feed". Trains of that length are the practical limit on my small testing layout. The owner of that model, a NYC Mohawk, reported pulling 35 cars over a 1% grade. I am sure that several of my heavier models would pull me, 160 pounds, if I wanted to rig up the necessary apparatus. I am building these models to perform the definite type of service with a minimum of attention and trouble and my customers must like them, because a few have bought a second one, and one is about to order his third.