Introduction

A roller mill is a simple machine consisting of a pair of rollers driven by a motor, and on which sits a drum containing the parts to be polished. The action of rotating the drum causes the mixture of polishing medium and parts to tumble around inside (hence the name...) with the resulting friction doing the work of cleaning or polishing. There are many different types of abrasives and polishing mediums that can be placed in the drum together with the parts to be cleaned or polished. At one end of the spectrum is carborundum type grits which cut quickly and are commonly used for smoothing surfaces, at the other are mixtures of crushed walnut husks and jeweler's rouge for imparting a high polish. The advantage of using the tumbler is that complex shapes can be polished with no effort - it just takes a little time (the tumbler is often run over-night.

So what particular use is it to the model engineer? Well, I had in mind polishing those many small silver soldered brass fittings such as valves, backhead fittings and so on, items which are otherwise difficult to polish because of their small size and many nooks and crannies. In fact, tumbling with the right medium will also remove fine machining marks on machined bronze and brass castings, though I have yet to test this out (I vaguely remember a ME article on this, I must look it up).

The tumbler doesn't have to be very big, my rollers are about 10" long, and the drum is a 6" length of 5" diameter platic piping (cut from a length of soil pipe left over from a bathroom installation), with two plastic caps fitted to the end. The motor is a variable speed model and this is an important feature as the speed needs to be set just right so that 'tumbling' in fact does take place. Slow speed will cause the mixture to slide around the drum, faster speed can cause the mixture to distribute evenly around the inside of drum preventing good tumbling.

I made my tumbler from parts I already had to hand in the workshop, although there is a drawing listed for this project (it's probably not there just yet... in a week or so) there is no point in making it exactly the same. You will get the general idea from the drawing but go right ahead and adapt to suit your own supply of odds 'n ends.

My rollers were made from 1 1/4" round steel bar, tube would have been more economical but I didn't have any long enough of the right size. The base was made from a length of 1/4" cold-rolled steel plate - again because it doing nothing else useful at the time. The end pieces holding the bearings (simple bronze bushes) were made from 2" x 1/8" angle. The three gears had been dismantled from a reduction gearbox that was attached to some junked equipment, but these could easily be replaced by a series of pulleys and plastic belting.

The only problems I came across were that my rollers were a little too close together resulting in some instability - the drum can fall off if it contains a single heavy part (I've shown them wider apart in the drawing). Also, the smooth steel rollers don't offer enough grip which can result in the drum slipping rather than turning. In the end I stuck some fine grade emery paper around one roller to give it more grip, I had discarded the idea of knurling the rollers on the grounds that continuous use would probably grind the drum away! Lastly, the motor I'm using isn't the quietest in the world, and as the tumbler is often left running over-night this can be annoying.

So far I have only tried two polishing mediums, both donated by a friend. One is crushed walnut husks and the other is very fine (I mean really tiny) glass beads. I'm a bit wary of the glass but the walnut seems to work very well on brass (although it manages to get itself jammed into the smallest hole). I'll investigate the availability of local supply of polishing materials and post a list here, there might be more information in that ME article if I can find it. Having made your tumbler you might even get the bug of polishing rocks and semi-precious stones, an even more common usage for the device.

(c) Chris Heapy 1996.

Back to previous page