Quick Update. In order to mount the flywheel to the bike I needed a 1/4-20 setscrew 0.36" back from the face of the little nub in the face of the flywheel. This would have required drilling in from the OD and tapping. My drill was barely long enough and my taps were not even close. A nut tap would have cost me $60 at least. I opted to tackle this feature with an angled thread from the side of the part. I set up the part on the mill using a drill press vise and an angle block to tip the part back at 60 degrees to the table.
Some quick trig found the distance along the X axis that I would need to offset in order to have the setscrew breakthrough the bore in the correct location to line up with the flat on the shaft.
Unfortunately while I was locating the part, I jogged the MPG in Y when I meant to move in X and crashed my Tormach TTS edgefinder. They don't make these anymore and there is not a great replacement option. If anyone has one to sell, please let me know. I was able to tap it back to running true within 4 tenths, but I still am not happy at all.
The corner of the nub was visually located with a chamfer tool and the 0.1851" offset in X was applied.
And here is the hole, counterbored and tapped. The location of the hole ended up perfect.
At this point I tried it out on the bike and it sounded like a helicopter was taking off on the roof of the house. Things were falling off shelves! Did the math and if you were to bike at a speed equivalent to 20mph outside, the flywheel would be turning at 5500rpm. At that speed, 1 gram of unbalanced weight (equivalent to a paper clip taped to the OD) exerts 5 pounds of force radially! So even though all features were indicated true within .0002", the unbalanced weight was too much.
I revisited the machining by drilling a second setscrew hole 180 degrees opposite to the original and mounting the flywheel on a 10mm arbor I turned from 12L14 steel.
From here, all the surfaces were skimmed to clean them up. The weak workholding (only 2 setscrews and a tight sliding fit to the arbor) made for awful rigidity and the cut just wanted to chatter like crazy. I had to finish all the surfaces at 150 rpm at 0.004"/rev feed and the smallest depth of cut possible. Even with that, the finish was not nearly as nice as it was before.
I remounted the flywheel to the bike and it was way better than before. The house did not shake, but it was just loud. Too annoying to use. At this point, I need to consider dynamic balancing the flywheel for these speeds, and probably upgrading the cheap bearings on the shaft of the trainer. It might be that the flywheel is perfectly balanced but the bearings and housing have enough slop that the shaft is able to wiggle radially and throw the flywheel out of true.
For now, the flywheel sits on the shelf of shame and will beg me to revisit it. I do have to say that even though it was too loud, it felt so much closer to riding outside. I really hope I can get this to work someday.
I have a lot more respect for the precision that is required in making those exercise spin bikes.
). Spin it and mark the low point when it stops. Spin it again and mark it. Do that a few times. Wherever you have alot of marks, drill a 1/2" divot. Clean off the marks and repeat the process, drilling new divots or deepening existing ones. Takes a while but you can get things nicely balanced if you're meticulous. I've done it a bunch on treadmill motor flywheels.
