Transverse belt resonance

UNH-OE / wave-tow-tank

Configuration files and issue tracking for the UNH wave & tow tank.

MIT License

2 stars 3 forks source link

Transverse belt resonance #65

Closed petebachant closed 9 years ago

petebachant commented 9 years ago

This has been happening ever since the balloon towing.

Suspected causes

Something fell in the timing belt housing
Belt preload somehow changed and now has a resonant frequency near the tooth passage frequency, which is tow_speed/0.02. Getting bad resonance at 0.6 m/s, or 30 Hz.
Servo tune got a bit stiffer over time. This seems unlikely as the vibration can't be felt at the motor, just at certain points along the path when the taut/slack sides of the belt are certain lengths.
Building settling making the master linear shaft less level.
Possible solutions
~~Vacuum out timing belt housing~~
Change belt preload. It can only be reduced a small amount, though it could help.
Tune servo drive.
Add mass or damping to belt near the tensioning clamp (where it won't go into the pulley housings). This shouldn't be necessary since everything was fine before. It would be smarter to find the root cause first.

petebachant commented 9 years ago

Increasing belt preload seems to make it worse.

petebachant commented 9 years ago

Tried putting a little WD-40 on the sliding surface--no real change.

petebachant commented 9 years ago

Actually, this may be the twisting mode.

petebachant commented 9 years ago

Reduced GV in the S700 drive from 10 to 6, which reduced some noise from the motor but not the belt.

petebachant commented 9 years ago

Put a block of plastic under the belt on the beach side of the tensioner near that end of the carriage. This stopped the flapping as it changed the unsupported length of the belt, and maybe kept it off some bumps that were perturbing it. I'm thinking the issue here is the belt sliding surface. It may need a reapplication of UHMW tape, or even to be replaced with a slightly thicker sheet.

Another idea is that something is deflecting (pulley housings?) meaning belt shortening will not preload the belt enough. Doubtful...

petebachant commented 9 years ago

Tried putting the block of plastic on top of the belt near the same location. This didn't help. It seems the problem is smaller sections of suspended belt that resonate from bumps on the sliding surface. I think the next course of action is

Remove the timing belt guards (will help identify the mode of vibration)
Inspect sliding surface for damage
Repair if necessary
Attempt to reduce belt preload even further and see if that helps

Still it is important to identify what caused this, since it was fine for the RVAT-interference-frame experiment.

petebachant commented 9 years ago

I'm almost positive this is the upper belt sliding surface. Going to open a new issue for that.

UNH-OE / wave-tow-tank

Transverse belt resonance #65

Suspected causes

Possible solutions