I'm using two adafruit 16-channel PWM/servo shields with 26 tower pro sg90 micro servos, and I've had multiple servos die in what seems like a very short timeframe.
The first round of dead servos was after a week of continual sweeping, running 24/7. Inspecting the servos didn't reveal any stripped gears; when I first unplugged them they were slightly warm to the touch.
I replaced the 9 or so that died, calibrated the pulse width for each of them individually, and set them all at 90 degrees. They have been holding that position for a week, powered 24 hrs/day, only moving for about 30 seconds when I test them briefly every few days by sweeping between 75 and 115 degrees. About 6 or 7 have failed so far.
They're not moving much of a load at all, all of the weight they are moving is supported primarily by traxxis u-joints.
I'm planning on buying a Futaba micro servo to test under the same conditions, assuming that at least part of the problem is that my sg90s are cheap POSs; but it would be really helpful to have some insight from anyone with more servo experience in case there's something I'm doing wrong that are causing these to fail. Or is this just to be expected of servos that are powered and holding position for ~200 hours?
I'm happy to provide more info on my wiring & code if that's helpful. Thanks!
The reviews on Tower servos are pretty hit and miss. Typically with servos you get what you pay for. What is the weight they are moving? I wouldn't push any micro or sub micro servo any more than 1/2 of its torque rating.
Another micro servo you may try is the Hitec HS-85BB or the MG version- they have twice the torque rating of the Tower Pro and I've found them to be very reliable (plus Hitec has an awesome warranty.) If you can go to a sub micro I'd try the HS-53 or HS-55. All are available from Servocity.com