Oscillating Steam Engine

The Story

b&w retro picture I've been long fascinated with steam power. Using some spare time I'd been lucky to have, I started building my model. There are plenty of succesful examples of homebuilt steam models on the net, however all of them require a lathe to build. I only have a makeshift micro drill/mill press so I had to invent something a little bit different.

This machine is built from parts obtainable in hardware stores and computer junkyards. This particular construction requires 20x6mm aluminum bar, 15x2mm brass bar, 10mm brass rod, 12mm OD straight copper pipe, dead computer hard drive, some hardware/scrap metal parts/plywood and at least a small drill press.


manifold base manifold assembly I started with steam manifold. It was the hardest static part of my construction because the tool I have is not really a milling machine. Nevertheless, a somewhat crooked yet functional part came out after a lot of shrieking noises.

Things learned:

Brass is so much easier to machine. Making the manifold plate was easy: just drill, then mill around the mounting screws. Those should be level with the surface, or lower.

flywheel The flywheel is probably the most creative part of this construction. I used a spindle of dead HDD motor. Using a dividing head I drilled 6 radial holes for the M3 brass-rod spokes. I tried tapping the holes but it didn't really work very well because the spindle turned out to be hollow under the surface. Then I used a circular drill to cut out a circle from 7mm plywood. flywheel standard Using the chisel tip of a wood drill bit, the circle was thus converted into a ring (gloves, glasses, there can't be enough protection if you attempt that!). The ring was then fixed into dividing head (it's like a lathe chuck) and milled around and inside to round shape. The holes were made, spokes inserted and thus the flywheel was born. It felt too light, so I added a copper rim later. Probably some places sell small amounts of sheet copper, but I used plumbing pipes.

The flywheel standard was cut roughly from HDD case first, then finished by a milling bit. The round part was finished by fixing the job in a dividing head. Using the HDD case gives the benefit of perfectly matched fitting for the spindle axis. The bottom mount is not pictured, but it was cut from the same HDD case, then a 2mm slot was cut out with cutting disc (roughly), finished with a 2mm mill.

Next come the reciprocating parts. For the cylinder I used a 12mm OD/10mm ID copper plumbing pipe. Cylinder caps were made from the brass stock. Round dents were milled out using dividing head. The M5-tapped hole for the drive shaft was milled out, not drilled, to ensure perfect alignment, then tapped. Drive shaft bushing was similarly milled in a piece of M5 brass rod. Bushing is fixed to the cap with a regular nut. This provides some cavity for a gland: i stuffed it with some cloth soaked in lubricant later. The other cap is just a cap, it just seals the cylinder from the other end. exploded view of all engine parts

At first I wanted to get away with a cast piston. I used some kind of epoxy clay rated for 240 degrees Celsius. It worked perfectly with compressed air, but it didn't work with steam at all. Epoxy is porous, it expands and its friction increases dramatically. It doesn't degrade, it just gets stuck. So unless you're going to be happy with air-only engine, epoxy as piston material is not an option. BTW, dead CD-ROM drives are an excellent source of all kinds of shafts.

manually milled piston The 10mm brass rod seemed to be a logical choice, besides I already had a piston made from it some time ago. A quick test confirmed that it worked, but its diameter was some 0.3-0.4mm smaller than the cylinder ID.. I had enough efficiency problems already so I figured it might be worth trying to make rings. I bored two radial channels in the side of cylinder, wound copper to fit, soldered it. Then came really tedious work of filing it to fit.

file-turining the piston rings

Curiously enough, after filing, it turned out that imprefections in my plumbing-pipe cylinder, alignment errors etc etc add up so that the piston with rings was in fact only by maybe 0.1mm thicker than the original rod. Probably it was not worth it, but I had my share of fun with file-turning. Besides I like it how it looks.

High-temp epoxy clay is used for cylinder caps and to fix the attachment of the drive shaft to the piston. Not pretty, but easy to fix in case of a screwup.

The moving slide-side of the manifold is made from a matching piece of brass. The most complicated part was connecting the round cylinder to the flat brass plate. First, both plate and cylinder were drilled to 3.2mm (that's the limitation piston, complete of my generalized-trade-name-drill). Since I have some 4mm brass pipe, I then enlarged the holes in the cylinder. Plumbing small tubes into the cylinder holes wasn't very hard, grinding them level wasn't hard either.. But then a problem came: it wasn't easy to hold 4 tiny parts together to solder them down. Besides, I wanted to reduce the diameter of holes in the moving plate..

I took short piece of 4mm brass pipe, made a longitudal cut, squeezed it down until it's round and fits 3.2mm. Sounds easy, it wasn't ;) Then I could solder pieces of this thinned pipe into the brass plate. And after that's done, the cylinder-pipes were fitting the brass-plate-pipes perfectly. I only needed to heat them up and voila!

first test assembly prototype running at full steam And here's the first prototype assembly. The excitement of your first machine running can't be compared to anything. Even the first BASIC program on a PDP-11, first blinking led.. They all are bleak in comparison.

marksplate My PCB-etching experience left no doubts about method of making a marksplate. I tried two finishes though, one filled with paint, another plain. After some hesitation I decided in favour of plain finish.

a close up of a complete model

After that I made a base out of same piece of 7mm plywood, shaped it with side mills, added a fiberglass bottom to protect it from extensive heat, mounted the marksplate on soldered copper nails and voila, the machine is complete!

A note on lubrication

Lubrication makes the machine run very smoothly, improves seals, reduces wear. All videos shown below use no lubrication at all and you can hear that. I learned it hard way that regular motor oil is absolutely unsuitable for steam engines: when mixed with hot water it becomes sticky and glueish; the lube must be water based, just like, you know... Steam model sites advertise using some special steam oil. A short research showed that all steam oils use tallow or lard as a base and contain at least 4% of animal fat. The purpose of other components is in fact to reduce production cost. Lard works like a charm, it's an excellent steam lubricant.


Here are some videos of the working machine. All of them only show the prototype build, although you hardly can tell in compressed video. Youtubified clip on the right is made from downloadable clips below. Convenience.


fancy looking blueprints

In the course of building I made some plans. They are not really complete, I mostly made them for myself to clarify the construction somewhat. Here they are, proceed at your own risk.


My steam engine took the honour of being mentioned in the following blogs or sites:

Thank you guys!

Viacheslav Slavinsky 2006 svofski on gmail
Updated: Tue Sep 4 20:04:18 MSD 2007