||There are steam enthusiasts that would like to have a little steam engine to drive a machine, generator or
go-cart but do not have the time, money or technical skills to make one. This article describes a neat steam
expander that utilizes readily available parts from a scrap bin. Modifications, if any, require no precision
machining or welding.
January 20th 2007, Stan Jakuba.
The Simplest, Cleanest and Cheapest Steam Engine
The first time I saw the innards of the air compressor of the wobble-piston type, it gave me
an inspiration for possibly the simplest, cleanest and cheapest steam engine. About six years ago, I had the
opportunity to fulfill that inspiration and build one. This is the story of the project: The description of
For those who are not familiar with the design of the small, ubiquitous air compressors sold
at Home Depot, Sears and similar outlets, the first picture illustrates the basic hardware. Thousands of these
compressors are discarded annually due to some electrical problem, and many of those have the compressor section
still in excellent condition and can be had for nothing. Fig 1. shows the compressor parts crucial to the
steam conversion, in this case to a single-acting, uniflow, non-crosshead, bash-valve expander.
How to make a steam expander from these parts:
All one has to do to make this a steam engine is to attach a basher (striker) to the piston.
Then make a plate that covers the cylinder and has a hole for the striker to go through.  Then attach the
compressor’s reed blade over that hole. The length of the striker must be such that when the piston reaches
close to the TDC, the striker lifts the reed a bit to admit the steam. An example of the striker and the valve
plate with its inlet hole is in Fig 3.
The electric motor provides the expander main shaft, main bearings and flywheel (the armature).
The crankshaft throw is formed by the small piece shown on the picture in front of the main shaft, and it is
portrayed with the big-end con-rod bearing on. There is no small-end bearing; the con-rod and the piston are
one piece. Thus the piston wobbles during its up and down strokes – hence the name wobble piston expander (WPE).
This feature necessitates these machines to be of a large bore & short stroke (2:1 typically).
The piston “ring” is formed by an inverted cup made of a proprietary plastic that
stretches to the cylinder compensating for the angle of the piston rocking. The cup seal is held within the
piston by a flat plate and a screw; some manufacturers swage the plate instead. The cylinder is made of an
aluminum-alloy sheet metal, anodized or coated with a Teflon-like substance.
A photo of the piston and cylinder is on Fig 2. The flange on the cylinder is clamped
between the head and the “crankcase” (the front part of the el. motor), and that is how the cylinder is
held in place.
The valve plate is shown up-side-down to illustrate the increased clearance volume necessary with
(uniflow) steam expanders. (As a compressor, the clearance volume was of course, minimized.) A large clearance
volume enables greater power of the expander and allows for a lesser lift of the inlet valve. Shorter lift
leads to the valve’s longer life.
Now, here is the trick that makes the WPE unique among all the bash valve engines: The valve
opening need not be symmetrical about the TDC. Accepting the rocking (wobbling) motion of the piston, it is
advantageous to place the valve at the piston periphery rather than the usual center. In this way, the striker
can reach the reed valve late before TDC but keep it open longer after the TDC. This assymetricity of the intake event
makes this uniflow expander easy to start and run at a low speed. It produces a decent torque right from idle, not
until the relatively high speed usual with the “symmetrical” bash-valve engines.
The machine we eventually built had, in addition, the cylinder axis offset from the centerline,
making the asymetricity so extreme that steam could pass by the piston a part way up the stroke. With only a
partial recompression, the engine ran as smooth as any counterflow machine would. More about this later.
To avoid the complexity of the exhaust valve, the expander is uniflow as said earlier. The
exhaust ports are formed by a few holes drilled at the suitable location near BDC of the piston travel. They do
not form a ring of holes about the cylinder. Instead, again, to take advantage of the rocking, the ports are
clustered in about 90 degrees of the circumference only, making the exhaust event also asymmetric. Although Fig 5. shows rather fancy, milled exhaust ports, there is no particular advantage for the slotted holes
except for assuring that the holes are not so wide that the flexible piston ring gets damaged by protruding into
The conversion to an expander is completed by modifying the existing head
or making a new, very simple one that serves just for bringing steam to the inlet valve. An example of a new
one with the inlet fitting installed is shown in Fig 4.
Thus equipped, the machine can be tested running on compressed gas (air) or steam. The
assembled moving parts and the cylinder are shown in Fig 5. The main shaft shown here is the one used in the
closed crankcase, not the motor shaft - the additional main bearing is visible to the right of the motor one.
Clean, no oil mess, operation:
There is no oil lubrication needed anywhere in the engine. The tribological properties
of the cup-seal and cylinder remove the need for lubrication there; the piston, the metal part of it, never touches
the cylinder. And all the bearings are lube sealed for life. Because the exhaust steam runs past the
con-rod bearing, a stainless steel version of that bearing is advisable if the machine is to run on steam frequently
cold. The stainless-steel bearing is available from McMaster and similar suppliers.
The cylinder surfaces are usually just hard anodized. They are mostly alike among the
compressor manufacturers. The cup seals, on the other hand, do differ among the manufacturers. For
running on steam, more suitable materials for the cylinders and piston probably exist, but I found the life of the
air-intended materials sufficient. Considering that the steam machine runs in a far cleaner environment than
many a compressor (no paint dust, sand, etc. in the “crankcase” below the piston), there is no need to hesitate
using the parts as they come with the compressor or are bought as spare parts. Intended for air, they will
love superheated steam.
As to the spec for the expander, the wobble-piston compressor operates typically at either
1800 rpm or 3600 rpm (nominal) pumping air into a reservoir rated at 125 psig. The life of these compressors
exceeds 2000 hrs. The same piston/cylinder assembly is sold in many different air-flow ratings; the flow is
determined by the length of the crankshaft throw. There are several manufacturers of these compressors, but
the parts and power ratings are similar among them. Generally, the air delivery is limited by the household
ampere-draw and the efficiency of the electric motor. The cheaper models tend to have therefore shorter
strokes because their inefficient motors draw too much current.
How to make the machine useful:
At this stage, the machine allows running for fun at no load. To use the expander for
powering something, either of the shaft ends needs a pulley or a coupling. There is a threaded hole and a
screw for that in the crank pin; it held originally the compressor fan. One can leave the fan in and make
wind. Or make and fit a pulley to the crank copying the fan hub. Or utilize the other end of the
motor. Or make electricity if the motor is still functioning because it can run as a generator if connected
to the appropriate voltage for the initial induction.
Concerning the steam conditions, obviously the rated pressure of 125 psig is perfectly safe.
The adiabatic compression temperature for that is above 500 F. A hobby steam boiler will produce steam
conditions benign by comparison. I had run the further below described WPE at over twice the rated pressure
(260 psig) without any adverse effect. The life under that load is, however, unknown. One of the
reasons why the hardware tolerates that much higher pressure lies in the fact that, at 3600 rpm, the pressure
inside the compressor is much higher than the 125 psig in the reservoir in order to push the air thru the exhaust
valves that fast. Of all the components, the reed valve probably hated that higher pressure most.
Collecting the exhaust steam:
The thus far described conversion works well and produces truly spectacular show for almost
no money. Exhaust steam is, of course, puffing into the open crankcase, and if the steam is wet and the engine
cold, much condensation results. The project I mentioned earlier required an expander in a closed circuit,
meaning that the exhaust steam had to be collected to bring it to the condenser. For that, we built a closed
crankcase, and, accepting the fact that there will always be leakage into the crankcase past the piston ring, we let
all the steam go into that space to keep the design simple. Fig 6. portraits a WPE suitable for mounting on
a dynamometer test stand.
It was designed to eventually power an el. generator. That generator was not capable of motoring, and therefore
a starter motor shown in the picture on the left was incorporated and tested. That necessitated the extra
bearings and an over-running clutch. With its highly asymmetric events, this single cylinder required only
a 1/2 revolution “push,” in the right direction, for getting it to rotate on its own. A miniature 8 W, geared,
el. motor (extreme left) sufficed nicely. With the production version, the starter was attached to the end of
the generator, away from any steam, and without the extra bearings. The final product, coupled to a modified
boiler in a home-heating device to provide emergency electricity for the home, is shown in Fig 7. The generator
is visible behind the crankcase block. Notice the offset of the cylinder axis from the crankshaft plane.
In summary, there are four feature of this expander unique with respect to all prior expanders
(and also with respect to the wobble piston air compressors):
(1) The piston axis is located substantially off the plane of the crankshaft,
(2) The steam admission is via a piston actuated inlet valve, an arrangement whereby the valve
is formed by a striker attached to the piston lifting a reed blade that covers the inlet port.
(3) The striker is located as close to the periphery of the piston as possible, the limiting
distance being the need to avoid the striker touching the cylinder at its max. rocking angle.
(4) The exhaust ports are arranged in a section of the cylinder only so as to provide an
asymmetric event similar to the inlet event. In the extreme offset case, there are no exhaust ports at all.
Features (1) and (3) imply opening the inlet port with only the minimum lead and cutting the
steam off at any desired angle. The design accomplishes this feat while retaining the utmost simplicity of
the valve train – no more parts than in a “symmetrical” bash-valve expander.
The exhaust from the cylinder can be accomplished via the usual ports as said earlier.
However, above a certain offset, a test showed that the expander will run equally well with or without the exhaust
ports. There is a limit to the offset – too much of it will wear out the seal cup on one side.
A multitude of tests were run with the described hardware. Power levels between 2 kW to
3 kW were obtained at high speeds. Overall efficiency was relatively low on account of the large surface area
of the grossly over-square bore/stroke ratio, the aluminum walls, and, of course, the tiny size of the expander.
Despite, efficiency could be “on par” by feeding in superheated steam. That would avoid much of the inevitable
and ample condensation of the saturated steam that was specified for this engine. Speed was also limited to
achieve quiet operation – the expander was actually not audible in the home heating system background noises of fans,
the burner and pumps.
As an example of a performance with steam at about the same pressure as the air compressor had,
and at 1800 rpm and 46 degrees cut-off, the shaft power was just under 1 kW. For those interested in the geometry,
the lead (before TDC) for this cut-off and 10 mm offset was 22 degrees.
In the case that any reader likes this design but needs more power and has plenty of parts, the
next picture shows how to obtain 12 times more power.