The rotor assembly needs to be reasonably well balanced for the engine not to tear itself to pieces.
The rotor assembly was statically balanced, but this will hopefully be sufficient as:
- The compressor wheel is a commercial turbo part and should be very well balanced already
- It is a shrink fit on the shaft so it should be concentric
- I did my best to make sure that the shaft was true (machining between centres, etc.)
So if everything is assembled carefully (and all the spacers, etc. are dead parallel, and the bore in the turbine disc is true), the only thing that should be out of balance is the turbine wheel. So by statically balancing the turbine wheel, the whole assembly should be in reasonable dynamic balance.
One does need to be *very* careful not to introduce a bend in the shaft during assembly. As an example:
After I had finished the shaft, I fitted a bearing, the rear spacer and turbine wheel, tightened the nut up. I put the assembly back between centres intending to skim the tips of the turbine blades but at the last minute, thought I should quickly check the runout before starting to machine. It was miles out! (about 0.05mm). I slackened the turbine nut and it went back to ~0.004mm. I dismantled it again, gently stoned all the mating faces, gave it a good clean in fresh solvent and it went back together and ran true to ~0.004mm TIR again.
Initially I used the ‘rocking’ method – the turbine and shaft on nice, clean and dry bearings are placed in a loose fitting tube which is then rocked gently from side to side and see where the heavy spot on the turbine wheel settles. You add a bit of clay, blu-tac, etc. to the high point and adjust its position to get the assembly to settle randomly. You then then grind a bit off the heavy side of the wheel and repeat the exercise – hopefully the blu-tac needs to be in the same place, but a smaller lump. The size of the lump of blu-tac needed is a good gauge of how much progress is being made.
I’d got as far as I could with that – progressing in a few hours from a lump of blu-tac the size of a couple of peas and a very pronounced heavy spot to a lump about the size of a lentil and no real resting point. Most of the metal I removed was by looking for thicker blades around the heavy spot and thinning them out a little with the narrow belt sanding attachment I used to form the blades originally.
I thought that I might be done, and put it to one side for a while. I came back to it later, but couldn’t really make any improvements using this method. I was going to stop there, but on the advice of another builder, tried the ‘fingertip’ method described below.
The balancing method that worked best for me was the ‘fingertip’ method as described in Kurt Shreckling’s original book: In this method, you hold the rear bearing of the assembled shaft / turbine between finger and thumb, gently spin the turbine up using compressed air and feel for vibration as it slows down. You then repeat the test with a small piece of tape stuck to the turbine, moving the tape around the perimeter until you find the spot where the vibration is least. If this is better than the bare disk, then you grind a little bit off at the opposite side to the tape, then go again.
You don’t need to get the shaft spinning very fast for this – the vibration is felt as a gentle buzz at a relatively low speed. Once you’ve felt that, I don’t think there is any point going faster. When you get close, the lack of vibration is really noticeable.
It sounds long-winded, and I wasn’t really looking forward to it, but in the event, it went quite quickly – another few hours and I got to the point where I could tell the difference when a double layer of strapping tape (~120g/m^2) about 8mm square was added at the root of the blades (R=22mm). I’ve just worked out that this is ~0.03 gram.cm = 0.013 gram.inches which sounds surprisingly good – I suspect some luck was involved.
The turbine wheel, shaft and rear bearing were balanced as an assembly, and I have avoided dismantling them since.