Search all of my sites with Google

Friday, January 30, 2009

Motor Demo Video; Shaft failure

Today I spent entirely working on the drivetrain for the motor, since that part has to be finished before the pedal drivetrain can be setup, both for clearance reasons, and because the pedal's freewheel/receiver-chainring will attach to the outside of the motor's receiver chainring.

I did get it working, as this video shows. The bike is upside down so I could run the wheel without touching the ground, and so I could see and adjust things better. I need to make a teststand or kickstand that will keep the rear wheel off the ground and hold the bike vertically straight.
There's an extra roller in this video I had to use for now, because the small 1/2" ScootNGo chain used for the motor's final reduction stage is ten links too long for this application, and I don't have a safe method to shorten it, since the bike-chain tool I have doesn't work on links this small. The roller is the serpentine-belt tensioner from the dead '85 Ford LTD, as it is metal and won't wear the way any of the other potential tensioners I could make would during these tests. It's also already on a sturdy right-angle bracket, which made it extremely easy to securely clamp to the white ex-fork used for the rear interframe support bracket. Tensioning just means loosening the two radiator-hose clamps (also from the LTD) and scooting it up the tube, then retightening the clamps. Same way I tension the belt for the motor, by sliding the motor along the bottom tubes of the Magna's rear triangle.

It's running on only 24volts, from two of the car batteries, with no controller (just direct power). The reason it's in the gear it's in is that I have no cables connected to the derailers yet, so the rear one is in the default highest gear, and the unadjusted front derailer falls on the middle chainring there (I think the bike that derailer came from originally only had two front chainrings, but I can't really remember). Since each front gear gives about the same change from each other as two rear gears do, then if I were to shift into the largest chainring on the front, it would be about like shifting two more gears up in back, if they existed.

Unfortunatley, the drivetrain broke not long after this, as I was testing the bike's rear brakes with the motor running (to see if they would stop it even if the motor was going, in case of emergency on the road if the motor controller fails to "full on"--unlikely but possible). I was so absorbed in thinking how to fix it and improve it that I forgot to take any further pics today. :-(

The part that broke was the threading for the orange chainring's spotwelded-on nut, which both holds it onto the former pedal-axle and transmits it's energy into that axle. The nut's threading stripped out, as it was too soft a metal. The threading on the axle was fine. I had only gone with the hex nut because I found that using the part I'd intended to before would end up with the chainring too far out of line with the driving chainring in front of it.

Since all the thin hex nuts I have for this reverse-threaded shaft are essentially identical, all coming from scrapped cheap kid's bikes and the like, I decided to change how I would do this. The part I was going to use before is the much harder bearing ring, which is threaded onto that shaft just inside of the hex nut on a normal bike with one-piece cranks. I didn't use it until now because it's about 3 times as thick as the hex nut, causing the chainline problem mentioned above. But it also has 3 times the number of threads on it, and is very much harder metal (since it has to stand up to bearings rolling across one of it's surfaces under the pressure of pedalling's downstrokes).

In order to use it, I would have to bore out the hole in the center of the chainring's mounting plate so that it would allow the core of the bearing ring to fit thru it, so I could then weld them together. The lathe comes to the rescue again, since I don't have an actual drill press. I chucked up the mounting plate in the lathe's head, then put my largest Unibit into a lathe-mountable drill chuck, in the tailstock of the lathe. A few minutes' work later, I had a hole large enough for the core of the bearing ring to tightly fit into.

I clamped the two parts together and welded them. When I was done, I found that something in the heating/clamping process had severely warped the plate, to the point I could *see* the curve in it. It was not curved at the center, where it was welded to the ring, only as it got towards the outer edges. Rather like an old vinyl record left in sunlight, but without the small wavy warps in it. The only fix I had was to hammer it as flat as possible, which unfortunately isn't very--perhaps at least 1/16" in either direction from a flat plane.

Now, it was already warped a little bit even when it was on the exercycle, to the point of causing part of the wobble you see in the orange chainring in the video above, but I didn't think it would get so bad that it could vary more than 1/8" in either direction from a flat plane, just from the welding/clamping process. It didn't change noticeably from the spotwelding of the hex nut, but it was not heated nearly as much for that, either.

Even with the warping, the chain doesn't come off, which is good. I will still need to replace the mounting plate with something much better for actual use, though, since I don't like the way it feels or sounds this way. Not only that, but when I get the excess links out of the chain, it will then be problematic, as it won't have the tensioner in it anymore, and the chain is likely to come off if a bump is hit while it's in the looser positions during the wobble.

The bad news is that the threads broke again--this time not on the nut, but on the shaft itself. The particular bearing ring I picked was chosen for it's apparent hardness being greater than that of any of the others (using the lathe's toolbit to try to scratch the non-bearing surfaces). It is apparently harder than the actual crankshaft it would have been mounted on, as when I stalled the system again by braking with the rear wheel rim brakes, it sheared thru the threads on the shaft.

This is a real problem, as I don't have any shafts that appear to be harder than this one, and I have no way to put threads on anything this size, much less of a metal hard enough to be useful.

I can weld the nut to the shaft for more testing, but I will need a repairable drivetrain with removable parts eventually, so I now have to think of another way to build this part which is sturdy and disassemblable. I already have a bunch of ideas from before I did this one, but don't have all the parts needed to build any of them (which is why I went with this version), so I'll have to think of a different idea using only the parts I do have.

One possibility is a rear hub, using the freewheel cassette as the front gearset (in place of the simple triple-ring it has now). This complicates things some in that I would have to weld dropout tabs onto the bottom bracket area of the Roadmaster frame, to give a place for the hub axle to bolt to. It also means boring out the holes in at least the larger two chainrings so they will fit on the freewheel, and bolting them to existing sprockets there. The very largest sprocket, which is on the inside anyway, would become the granny ring, as it is of the same or similar size.

Hopefully I can still use the standard front derailer, as long as the hub is mounted in-line with the seattube, and I can slide the derailer down far enough to still reach the chainrings correctly. If not, I'll need to either add another section of seattube somewhere for it, or I will have to try to adapt a rear derailer to do this job (I'm not sure it can, because of the very large difference in sizes between the three rings).

I'm still left with the problem of adding the leftside pedal-drivetrain freewheel, which should go on this section--I had a nice workable idea for the one-piece crankshaft setup, but that won't work at all for a hub-based one. I still have some other ideas left but they're not yet explored in any detail, and may be unworkable.

It is a little disappointing that such a major part failed so simple a test in this way, but I'd certainly rather have it all happen now than later on the road. I suppose this is one of the problems with essentially an improvise-as-you-go design. :-) It's still fun coming up with solutions, especially adapting things never meant for this into something very functional and useful.

One other issue I need to deal with is the motor's belt system. Despite using the lathe to do it, the plastic receiver pulley's inner hole is not perfectly centered. I'm not sure how I managed that, but I did, so I need to re-lathe it out a bit to make it perfectly centered, as it is causing a significant wobble in on the freewheel, which at high motor speeds shakes the whole bike a little bit, as if I was going over a roughly paved road. That also means the belt is being stretched and loosened a little bit each time it goes around, and will wear it out faster, and cause slippage, losing some motor power when that happens.

It may also cause other damage, such as bearing problems or even shaft problems in the motor itself, long-term. A similar issue with my friction drive, where the rear wheel was not exactly the same thickness all the way around, eventually helped cause the shaft failure of one of the two fan motors (documented in a previous post, if interested). I doubt the shaft will break on this treadmill motor, as it's much thicker, but it will wear the bearings faster than normal.

Gotta work tomorrow (and probably the next day), so probably no more blog entries for the next few days.

No comments:

Post a Comment

Alternate suggestions or improvements to anything that's been posted is very welcome, and extreme detail is preferred to brevity.

Keep in mind that unless you leave an email address in your comment, I haven't any way to reply to you except to reply to your comment here. That means if you want a reply, you'll have to come back to *this* blog entry and it's comments to see my reply to you, unless you leave some method of contact within your comment.