Re: Design of Reciprocating Piston Expanders
March 21, 2015 12:58PM
Read the article Ken and I did for our SACA Newsletter, we went into this exact subject.
Re: Design of Reciprocating Piston Expanders
March 21, 2015 09:10PM
Hi GeorgeN,

Do you have more information on the hydraulic valve system? I've built, but yet to run mine. I have some check valves in there so that for each cycle, all new fluid is passed to the lifter. It does create a runaway potential in that the faster it runs, the faster and higher it lifts, requiring a feedback control. May just have to run it as is, but it would be great to know more of what Hal Fuller discovered and developed.
Thanks, Keith

Edited 1 time(s). Last edit at 03/21/2015 09:18PM by kdc2.
Re: Design of Reciprocating Piston Expanders
March 22, 2015 09:42AM
Here's a comparison of several mechanical systems. I would like to use the piston as the only mechanical actuator. I think one shot air valves should supply some design inspiration. The basic principle involves a delay through an adjustible bleed to the pilot just like the valve in the Z cycle engine. Perhaps that principle can be applied to something like the Hemick/Crank valve.

Lohring Miller
Re: Design of Reciprocating Piston Expanders
March 22, 2015 11:42AM
I'm back again:

All this talk about steam actuation of valves may have disregarded the potential of an existing (and certainly fascinating) engine designed by a refrigeration engineer by the name of Stover and highlighted on Tom Kimmel's excellent website. Here's a link to it: [] BTW Tom has updated the original info on the engine with much more extensive imaging (thanks Tom!).

I became fascinated with this project and applied my primitive thought processes and here's some observations that have resulted:

The sliding cam is easily created even by garage-level enthusiasts; it takes advantage of the fact that relatively small movements and forces are necessary to move the pilot valves.

The engine as depicted is obviously intended for relatively slow speeds; the pilot valves are wildly isolated by long feed lines to the actual cylinder feed valves.

The actual operation of this engine is restricted to relatively low pressures/temps and would require necessary lubrication--so may not be relevant to many participants on this thread. There are also severe limitations on maximum rpm's but careful design might make the concept viable.

For me, the obvious direction to take has been to integrate the pilot valve into the primary intake valve assembly, thus eliminating the length (and resulting delay) of the connecting tube. It is also possible to create both opening and closing forces with steam actuation using an intelligently designed pilot valve. I spent quite some time and exercised more than a few brain cells creating my own concept for an "updated" version.

In any case the ingenuity displayed and the quality of the fabrication of this engine deserve respect and observation IMHO

Re: Design of Reciprocating Piston Expanders
March 22, 2015 03:37PM
Hi Folks

There is a 1935 GB patent GB430425A by Verkspoor N.V. of Holland for a hydraulically operated inlet valve with the production version using a Bosch style inline pump. Valve cutoff was controlled using a spill port just like a fuel injection pump.
Here are two links, one in original language and hopefully the second translated. The article about Verkspoor Meier Mattern is half way down the page.

Re: Design of Reciprocating Piston Expanders
March 22, 2015 09:54PM
Good research Brian. It's the Spilling engine, on of the few in production piston engine steam companies in business currently. Thanks, Keith

I missed the Meier Matter engine, still studying it, youtube. the video may lead to another.

The Ascherslebener link is pretty facisnating goog translate and looks to have some linked movies too. It looks like we are blocked here in the states from viewing the movie. I got flagged by youtube's TOS trying to circumvent sad smiley

Edited 2 time(s). Last edit at 03/22/2015 11:45PM by kdc2.
Re: Design of Reciprocating Piston Expanders
March 22, 2015 11:21PM
Jim Crank Wrote:
> Lohring,
> Read the article Ken and I did for our SACA
> Newsletter, we went into this exact subject.
> Jim
I did reread the article and see the reference to a controlled bleed into the bottom of the valve actuating piston. It's not clear to me that it woud work as a dependable delay. I'm still thinking about it. If it works, it would be the perfect answer to a simple uniflow expander.

Lohring Miller
Re: Design of Reciprocating Piston Expanders
March 23, 2015 01:25PM
Lohring wrote"", Jim wrote"", I was searching hi/low in Jim/Ken's white paper for the controlled bleed valve and realized it is likely in the bulletin. Sigh
Re: Design of Reciprocating Piston Expanders
March 24, 2015 09:35PM
The electronic valves I have been exploring were first developed for an ic engine. And wereally to slow.

I have a version that is much faster. It is a type of series valve. But not really. I am thinking two cylindrical valves that seal against each other. The are sealed aginst each other. One is shifted opening a gap between them. Steam flows out the gap into the cylander. These valves shuttle back and forth. Switching roles. One opening the other closing. The on the next cycle the roles are reversed the one that closed the gap opens the gap and the other follows closing the gap. The steam flow is thru the hollow cylindrical valves and into the engine cylander thru the open gap when that are not butted aginst each other. These are like the IC valves that were shuttled back and forth by springs. When latched one spring is compressed held by the electro magnets in context with a plate attached to the valve shaft. Turing off the holding magnets allowa the compressed springs to propel the valve towards it other position where electro magnets are engaged to latche when it gets close. A cording to the article I read on the IC valve they got a 4 ms open time with a single puppet valve. This has the added feature that closing can timed to opening with the closing valve following the opening valve presenting a reduced steam flow area. Throttling by controlling the valve opening gap. These would be computer controled. This would have the throttle an analog input to the computer. With a mechanical shut of valve that closes and opens by mechanical conceptions to the throttle control. A IC gas peddle like foot control. At low speak the valves can be shuttling back and forth effectively a throttled long cutoff getting off the line. It means I need a large displacement to keep rpm down. That is the reason I am figuring on a compound. High expansion at low rpm with overlapping cutoff the line to some speed were initia keeps things smothe. That is a big design prpblem. Assuming the valves work the engine needs to be the right displacement for the steam conditions and vehicle weight and airodiynamics. With the valve timing limitations I need it to be a three stage compound. But unlike the old simple compounds. The stages have inter stage receivers large enough to smothe out pressure fluctions. Each stage is operatoring between fairly constant pressures. For example say each stage has a 3 to expansion. The first stage expands to some pressure above the interstate reciever it exhausts in to. The computer control varies cutoff to maintain interstate pressures. The valves can be rapidly shuttled for throttling so each stage has throttling. The interstate receivers have valves allowing steam to be bypassed fro. The previous reciever. That is.used at start to precharge them to operatoring pressure. And insured a minimal pressure at all times. Varing the clearance and cutoff mataining a near expansion ratio with the varied by cutoff. This is a counter flow with separate computer controled exhaust timming. So compression to near inlet pressure, controled by the same computer controlef shuttle valves eliminates clearance losses. This idea came from Jerry's original compression paper. For example say we have a 3 to 1 expansion ratio. By varing the clearance we in theory get to a point were we have 50% clrarance. That would be a 3 to 1 expansion with 0 steam afmission. 1.50/0.50. At.the other extremely 0 clearance and 1/3 cutoff. So with this the amount of inlet steam can vary from 0 to 1/3 the cylander volume. What are the real limits. I don't know. But let's say that we have some way to vary the engine stroke under power. The same variable clearance control power range is multiplied by the displacement varing. Now again say we have zero clearance at full stoke. If we reduce the stroke by 50% about the center of stroke. We have 50% displacement and 25% of the total full clearance. A 3 to 1 expansion 0 inlet. We.wouldn't really have a 0 clearance at full stroke. This is just simplified for explenation. One implementation idea is a two crank engine with two pistons per cylander. The effective stroke can be changed by varing the crank relationship. Shifting one forward and the other reverse a number of degree. This works as the effective TDC is when the leading cylander is past it's TDC as the other is before its. That's thrin displacement point. You can test this on your spread sheets. The displacement is the sum of the two cylanders. Set up a Sim that calculates displacement from crank angle and do for two cylanders. You can assume a scotch yoke for simplicity. The cylander position is then a simple.trig function. Doing that you should get a sin wave displacement exactly centered between the cylander pisitions. Maybe an extremely heavy-duty resolver could by used to shift crank angles from a common output shaft. Don't know. This is the main problem. An effective way of varying the stroke. Off the line we have full stroke with duty cycle valve throttling. Depending on foot feed we may be running a lower expansion ratio for hard acceleration or pulling. What ever. But in any case as.speed builds the stroke is reduced (Again depending on foot feed input The design goal is that in normal driving the engine would be running at its highest expansion expansion ratio. That would be the range were we can control power with cutoff an maintain a constant expansion varing the stroke. For a family vehicle that might be 25 to 80 MPH level no head wInd. Won't that no head wind high enough to maintain normal open high way speed with common head winds. To get high aversge MPG we need high speed efficiency.

Maybe a simpler engine could do the job. The problem is that to get efficient highway MPG. And not have a monster control problems. We maybe can get a higher rpm using tuned ports as S.E.S. did. To get the MPG up we need best efficiency at highway speefs. Power is a function of speed. Without arodynamic body tuning that could be a cubic relation. Low efficiency at 15 MPH has minimal effect compared to 60 MPH.

We all know how well the ideal Rankin cycle works. Even adding compression and clearance calculations it is by no means real world results. I have my steam properties plugin for VisSim working vary well for single state regions. The saturated mixture region is a problem in calculating constant property paths using integration. Not a problem calculating a mixture state point. Its the differentials jump across boundries. I think an engine should expand a wee bit into wet steam. There is a time factor involved for a subcooled gas to liquid transition. But yet another theory to be tested. For best efficiency we need to push that limit. Knowing how far into the wet steam region we can expand with out problems and parameters effecting that. For example air most recent and temperature might have an effect. Something that can be worked on over time. Also with a computer control engine it just a program mod.

Anyway thats about my complete crazy idea.

Re: Design of Reciprocating Piston Expanders
March 25, 2015 06:48PM
Hi Andy,

A resolver shouldn't be that hard to do as long as it doesn't rely too much on outside control. Say that the phase presented by the resolver is directly in relation to engine torque. I.E. as torque starts to increase due to more throttle then the resolver moves against a spring or hydraulic pressure, or both.

The "resolvers" to change the cam advance or retard in my compound design depend on something similar. At low rpms the oil pressure is lower creating a longer cutoff. Full, long cutoff is always there at stop through low rpm regardless of the other controls considerations. Cold oil would also keep it at longer cutoff until the engine is warmed up but cutoff is controlled by oil pressure to the "resolver".

Remember that according to recent thinking on this Phorum that a compound is death to a modern steam car! I don't believe any of it. I gave my reasons, nobody said much. There are just too very many variables between a new design of compound and the old ones, or any new engine and existing similar ones. We are all trying to learn from the past examples though so as to avoid expensive mistakes should any of us get to the point of cutting metal. For most of us that generally means money or dedicated time.

For me that means it all has to start on the test stand with a solidly built test engine, a dynomometer and full instrumentation which includes tribological analysis.

Several well thought out designs are needed to lay out a path of exploration and much flexibility of mind to take advantage of new and improved information from those test results. I now have three or four good designs ,I think, to get started with, but I am willing to S.K. favorite parts of all of them if testing doesn't prove their effectiveness.

Jerry Peoples has a new design for intake valves as well as a different cycle. I have a variation of his "Z" cycle with the zero clearance, high compression back into the inlet port cycle. Both cycles yield higher expansion ratios due to the zero clearance part. Mine should be more thermodynamically efficient (h1-u2) while his (h1-h2) should be more mechanically efficient. Only identical cylinders on a dyno and comparative charts will say for certain which way to go. Jim and Ken have a new design of bash valve which I don't know much about since I don't have that Bulletin that explained it. That needs to be tested.

The two competing designs for a viable condensing system need to be tested. To get any higher horsepower will require this. Freeze proofing is a big issue here as well as with the whole system. Freeze proofing the various pumps in the system needs addressing.

So many issues just don't show up until one has the broken pieces in their hands.

Best Regards,

Bill G.
Re: Design of Reciprocating Piston Expanders
April 01, 2015 01:05AM
That oil pressure stuff sounds like a Woodward analog steam governor before they went electronic....they worked really well both for variable speed control and fixed speed control like on an electric generator.And like most things commercial steam bigger and more expensive then a hobby steamer would use.
Re: Design of Reciprocating Piston Expanders
April 27, 2015 04:21PM
Hi all, good read! Here are comments directed at Dullfig, Frustrated, George-N, Andy and Bill Gatlin.


Hello all:

been thinking about hot to put together an expander for a car. I know that car engines and steam engines have very different requirements, but quite frankly very few have the funding to build an engine from scratch. So here is my current thinking for converting an engine to steam:

take a 1.2L 3 cylinder diesel (kubota makes one for tractors and such) and turn it into a steeple compound. I was thinking that I could remove the current cylinder liners, and machine exhaust ports in them. The exhaust steam would then go through what used to be the water jacket. In other words, no water cooling, just use the water jacket to exhaust the steam.

the HP cylinders would be a sepparate unit machined from aluminum that would replace the engine head. So where the head used to be bolted on, there would now be a block with HP uniflow cylinders. Kinda sorta like Doble's Ultimax.

Has this been tried? any obvious reasons why it won't work? Would the diesel crank and rods be strong enough?

BTW, the engine is rated at 70HP. Since steam doubles the power pulses, does that mean that (given enough steam) it could put out 140HP?


1. Loads: You could try to figure out what loads the rods and bearings are seeing in the diesel. Diesel combustion pressure is very high, but I am not sure diesel peak pressure puts the same loads on the critical components as steam peak pressure - time being a factor?

2. Power: It depends. To mimic the Diesel but with twice the pulses per time, you must assume extremely short cutoff and that you are able to breathe steam at the same rpm as the Diesel. In the steam engine you can have a higher MEP for the same peak bearing and rod loads, by increasing cutoff. A poignant question then becomes, what sort of valve train are you thinking of?

3: Exhaust to the water jacket? First, is it not going to be way too restrictive? Second, it does not make any sense to me to add heat to the exhaust, only to remove it in the feedwater heater. Why not run water through it?


Ken, you mean US2102389, not 2103389.


I didn't know Mr. Fuller developed an hydraulic valve gear. Can you please provide some leads or description? I am trying to convert the Skinner/Fuller gear that I knew of, the mechanical one with an opening and a closing cam, into an hydraulic one. There was a thread a while back.


Why do you want a fixed expansion ratio? With the same amount of complexity, could you not exploit a lower condenser pressure when you have it, like on cold days, or at low power?


Bill Gatlin,

Remember that according to recent thinking on this Phorum that a compound is death to a modern steam car! I don't believe any of it. I gave my reasons, nobody said much.

I am stoked about much of what you have told us about your engine, though it seems to have changed a bit over time, I don't know how it is now.
Mine is for a small motorcycle.
It is a unaflow compound with one HP and two bigger LP cylinders. It uses hp clearance heating, bleed and reheat and has a deep, overdriven steam horn.

You were talking about blowby. Has anyone tried to replace the hot steam blowby with cooler water leakage? A trunk piston in a unaflow has two sets of rings, one of them always below the exhaust port. I am thinking about pushing water into the cylinder above the lowest ring. Steam pressure would push water across the ring, but the water is more viscous, so less volume is leaked to the crank case. It also doesn't have very much heat to give up to the presumably cool crank case.

Guys, how do you intend to make this zero clearance in real hardware? And where is the info on Jerry People's Z cycle?
Re: Design of Reciprocating Piston Expanders
April 27, 2015 10:23PM

I won't to run near to full expansion.

My idea is vary complex compared to a typical steam engine.

There is no throttle. A safety shutoff. The boiler runs basically direct to the hp inlet valves steam chest. I plan on using magneticly latched duel valves. The are microprocessor controled. I have explained these valves elsewhere. The engine will vary the clearance in some way that is also controled by a micro processor. As for Varing the clearance I have looked at several ways. Duel crank, opossed piston etc. The effective stroke and clearance can be varied by shifting the crank phasing.

There is the wobble plate engines that varying the angle of the plate changes the stroke. And again in opposed piston wobble plate that baring stroke with the plate would remain in balance. We have the basic expansion formula:

expansions = (stroke+clearance)รท(cutoff+clearance)

At full stroke we have min clearance. At full expansion and full stroke we have some cutoff that is the max. Reducing the stroke and increasing clearance maintaining full expansion we have shorter cutoff. By adjusting cutoff and stroke to maintain full expansion we have fairly constant expansion ratio but with varying amounts of inlet steam. The micro also controls separate independent exhaust valves to maintain compression to inlet pressure.

The primary mode of control being the above the engine is running as near a full expansion compression cycle as pritical. We can extend the upward power range by increasing cutoff. That would then be a less efficient partial expansion mode. Low power would be by the same valves vary quickly opened and closed to effect a throttling. The valves in a way are like a series valve areangement were one controles opening and the other closing. Except they operate in the same space and alternate open close functions and actually closing following the opening making the flow path restricted. In effect throttling by controlling the gap between the opening and closing valves.

The valves have a limit as to the time to transition that forces a multistage design to accommodate the slower valves. In theory a triple expansion can have a 125 to 1 power range running full expansion of 24 to 1 expansion ratio. That would be 2.884 expansions per stage. But stage expansion would be done so as to produce equal per power per stage.

It is not the topical compound either. It would have interstage receivers at a minimumpressure. A minimum bypass pressure device that bypasses the first stage steam (at boiler pressure). At start up and when a stage might draw down it's receiver it gets make up steam. All stages are operatoring in all conditions. The electric controled maglatch valves allow independent timed inlet and exhaust events and throttle for each stage and cylander.

If you think my engine is complicated check out writing a compiler. That's a computer program that takes a computer program written in a high level programming language and translates it into binary machine code.

Then think about writing a compiler for writing compilers.

A quick very simple exmple:

  DECLARATI0N = ID "=" NUM ";" :EQU!2 DECL[*1];
  ST = ID ":=" EXP ";" :STORE !2 C0MP1LE[ *1];
  EXP = TERM $(("+" :ADD|"-" :SUB ) TERM !2);
  TERM = FACTOR $(("*" :MPY |"/" :DIV ) FACTOR !2);
  FACTOR = ID | "(" EXP ")" | NUM;
  LET: 'A'| 'B'|'C'|'D'|'E'|'F'|'G'|'H'|'I'|'J'|'K'|'L'|'M'|
  DGT: '0'/ '1'/ '2'/ '3'/ '4'|'5'|'6'|'7'|'8'|'9';

DECL(EQU[X, Y]) => DEF:(X) := Y
EVAL(IDP(X)) => DEF:(X):
(NUMBER(X)) => X;
(#V1[EVAL(X), EVAL(Y)]) => #U1;
    #V =ADD, SUB, MPY, DIV;
    #U = X + Y,X -Y, X * Y, X / Y;

The above is a simple interpreter that takes simple arithmetic expressions. A value may be assigned by typing:

X = 5;

Storing 5 in variable X.

Y := X + 8;

Would add 8 and X storing it in Y and print the result 13.
I implemented SLIC which is much like the language the interpreter above is written. SLIC was used to write a COBOL compiler.

The point I am trying to make before someone says it's to complicated, is I have did much more complicated things. And have actually made them simple. The above example of a language designed for defining and implementation of a programming makes writing a translator for a language simple. I wrote the compiler that translates the compiler writing language.

Such languages are called metalanguages. Special languages can be designed for most any subject. Even steam engins. We use a specialized subset of English in talking about steam engines and cars.

Re: Design of Reciprocating Piston Expanders
April 28, 2015 06:44PM
Guys, how do you intend to make this zero clearance in real hardware? And where is the info on Jerry People's Z cycle?

the information is here- Steam Automobile Bulletin Volume 29, Number 2 March - April, 2015 page 16.

I joined the SACA so I could get an eye on this type information its worth it to join.

I have only glossed over the article lots of enthusiasm on Jerry Peoples and Jim Harmon. Ironically it does embody a electro magnetic valve.

Re: Design of Reciprocating Piston Expanders
April 28, 2015 07:36PM
Thanks, I will, soon as I remember it at the same time as I have money.
Re: Design of Reciprocating Piston Expanders
April 28, 2015 08:57PM
Hi Sidrug,

Its only 33 bucks a year, I usually I opt to the 3 years for 99 dollars. Oh ya they will send back issues of the bulletin once your a member and you get them for free if you ask.

I think you can still get copy's of the bulletin in the store-room. If you opt out. .

Re: Design of Reciprocating Piston Expanders
April 29, 2015 05:48PM
Ive been 3 months overdue and still got the Bulletin the whole time. I just paid for one year. The join the club link and to paypal all worked with no problems. I would have done 3 years but I figured lets just get this taken care of now. With all the money Ive spent buying mosfets just taking care of what I can.

Jeremy Holmes
Sorry, only registered users may post in this forum.

Click here to login