Barry Dunman

New 21 January, 2001

Barry resides in South Africa, and is a software engineer. he claims to have little training or skill as a model engineer, but I think you will agree that the obvious quality of his model belies that assertion. I will let him tell the story his way:

Hi Roy,
The engine is as finished as it will ever be, so here are some pictures!

I'll briefly relate how all this came about. I write software for a living, but like to do more "real" things when I can. I've had the ambition to try some model engineering for many years, but never pursued it. About 15 years ago, I bought a little lathe (Unimat) from a hard-up fellow I met in a bar in a mining town, where I was doing business. It has been sitting in its box all these years, waiting for the moment.

Then in 1999, I got the idea to make some sort of "coffee table" engine - something that was visually interesting and unusual.. I did a lot of surfing, looking for ideas, and found your site, along with some other Stirling sites. I was captivated by your engines, but doubted that my non-existent machining skills plus the Unimat were up to the job. A few experiments confirmed this to be the truth!

Then I saw the "Test Tube Stirling" site, and made one up in an hour or two. I was amazed by how strongly it kicked - much more so than I would have guessed. The next evening the test tube got mounted in lashed up gimbals, and equipped with a con-rod and flywheel. It sort of ran, but balance was a big problem, and it was very jerky. The next night, it became a "two tube" engine, which solved most of the problems. It then crossed my mind that I should use this as the basis for my coffee table engine, because there is no super-critical
machining involved, and I could get some valuable experience in the gentle art of engine building. Obviously, the Unimat is a bit limiting in terms of the size of a piece that can be turned, hence some unusual design features,
such as the flywheel (which took weeks to make).

The engine starts after about 30 seconds, and runs very smoothly at about 40-60 rpm, which means that it can be easily tracked visually. The revs are self limiting, because as the engine speeds up, a point is reached where the marbles don't have time to reach the ends the test tubes - they just oscillate around the center point. An early problem was that heat tended to slowly travel along the tubes, causing the balloon "diaphragms" to distend, and eventually stopping the engine. This was addressed initially by making the pads that that the diaphragms work against adjustable (via the thumb wheels at the rear of the engine). The cure actually came with the addition of the fins behind the front gimbals. I tried it on one side first, unsure of whether it would get heat out of the tube, or just help conduct it further down the tube. The former was true, and the engine now runs indefinitely, the adjustable pads being almost redundant.

I have to say that building the engine has been a lot of fun, and a very valuable learning experience (witness my reject part box!). I get most of my materials from a scrap dealer, so the cost has been very small. The engine
certainly is a talking point - most people can't figure out what it is, much less that it's an engine that runs. Also, its motion is intriguing, with the marbles shuffling gently back and forth and the flywheel turning. 

So what next? I think I could use a bigger lathe - somehow I don't think I am going to pick up the next one in a bar! Then maybe an LTD ... or a Ringbom engine... or!!!!

Best regards

Barry Dunman

And here are the photos of an engine built by someone with "non-existent machining skills"!

Test tube Stirling, view 1

Back to Other's Engines  -  Home