An inertia starter for a model aero engine.
As soon as I started my efforts in model i.c. engine building I became fascinated with the questions it raised about how engine design and performance is affected by scale.
Following on from that I started thinking about scaling issues in general and about whether engine accessories such as starters and magnetos would work if scaled down to suit a miniature engine. I have not thought the issues through fully for magnetos – there is at least one non-scaleable quantity involved which is the ignition energy of the fuel/air mixture. For a starter however, and specifically the inertia starter, the mechanical aspects do I believe, scale appropriately with one important proviso. The proviso is that to store enough energy to work, a scale inertia starter appears to need a flywheel speed greater by the inverse of the scale factor. That is consistent by the way with the observation that a true scale engine should be able to run at the same multiple of speed compared to the prototype – 4x faster for a ¼ scale engine for example.
I happen to be particularly interested in the radial aero engines that powered so many aircraft in the 1920s and later. One of the common starting means for these classic engines was the inertia starter, in which a small electric motor or hand crank was used to spin up a flywheel to high speed. When the starter was engaged the flywheel drove the engine through reduction gearing, giving up its stored energy as it turned the engine over a rev or two, and with a bit of luck a pot or two would fire and the engine would start.
Having built a 25cc 5-cylinder radial aero engine model (see “A freelance radial aero engine”) I decided it would be really neat to add an inertia starter, mainly for the iconic sound of the rising whine of the flywheel spinning up followed by the descending groans of protest from the reduction gearing. And a running engine would be the icing on the cake. A secondary consideration was to see if my understanding of the scaling laws held up.
At this point I should say clearly that I have not built a ‘scale’ inertia starter; at the time that was beyond my reach, and indeed I am not sure it would be in my reach now. What I have built is a proof-of-principle starter that is somewhat over-scale for my engine, quite crude in design, but does start the engine. It is a step in the right direction.
In fact I designed the starter around two items. The first item was a gearhead from a defunct battery screwdriver; a neat double-planetary unit of reasonable overall size and having about the right ratio. The second item was information; I had measured the power needed to drive the engine at cranking speeds and figured out roughly the energy needed to rotate it a couple of turns. That allowed me to size the flywheel and find the necessary flywheel speed, which turned out to be 27,000 rpm. I went for an electric motor drive and found that a standard RC servo motor over-run at 12V could reach 30,000 rpm unloaded and about 25,000 rpm when driving the flywheel and reduction gearing - close enough. I added a torque limiting clutch, as standard on the real things, and a starter dog for engagement. These components were located within a simple cylindrical housing.
The photos at the end of this post show the unit mounted on the engine test stand (clear of the engine for easy access) and the components removed from the housing. I forgot to include a scale in the photos but to give an idea of the size, the motor at the right is 17mm diameter by 18mm long, the flywheel about 30mm diameter and 6mm thick and the gearcase about 35mm diameter. The torque-limiting clutch is to the left of the gearcase. It uses fibre washers for friction plates and ‘spring’ pressure is provided by soft rubber discs. The two halves of the casing are joined by a fine thread so the rubber spring pressure can be adjusted by screwing the two halves closer together or further apart. I am quite pleased with that clutch; it is compact and simple but works well.
There was a certain amount of adjustment needed as I began to try the starter, working gradually up in flywheel speed. The initial take-up of the drive is quite abrupt as one would expect, and various grub screws etc needed to be nipped up as things started to slip under the increasing loads. The clutch took a little while to bed in and also needed adjustment. After a while though I was able to run the flywheel at full speed and the starter would usually take the engine over a rotation and a bit with the ignition off. Not quite as much impulse as I had hoped for, but with the engine warm, ignition retarded and everything in good order it will start more often than not.
I have recently been picking up the threads of my model engine work; oddly, since retirement a few years ago I have seemed to have had less time for it not more, until recently. Joining Model Engine Maker has been a big encouraging factor. Thank you to all the members who have already shown an interest in my work, and offered helpful comment and information.
Is anyone else out there working along similar lines? I would be very interested to know whether any true scale or close to true scale accessories have been successful, and more generally to hear other people’s comments and observations on the scaling issues I have touched upon.