So the team is planning to use motors without clutches. Do motors “break” from stripping internal gears or by overheating motors repetitively.
I think it can be both. Last year we stripped a ton of gears, which thankfully can be replaced. However, we also “killed” a few motors which just stopped turning at all. I think it was the brushes or something, I’m not really sure. We’ve had about an equal number break each way, so we take the good gears from the dead motors to replace the broken gears in the otherwise good motors.
As Owen said, both stripped gears and overheating can damage motors. The gear stripping problem does not happen as often with the new 2 wire motors, since these use metal internal gears, as opposed to the old 3 wire motors which used plastic gears that stripped more easily. Stripped gears are simple to fix (you can just replace them), so if this happens to you you should still be fine. You should of course check your design if this happens to you, as there may be some component that is causing stress on the motors, but oftentimes gears can break simply because they have been overused, so it is not always a design problem.
Overheating motors are harder to fix but much rarer, at least in my experience. Generally, if there is a lot of power demand in a given part of the robot, the Cortex’s internal fuse will give out before the motor does. That having been said, I have seen situations, such as on elevators, where a lot of stress accumulating over a long period of use causes the motors to heat up. Usually, when you “burn out” a motor like this, the effect is only temporary, and after the motor cools down it will be good to go again. But, repeated overheating does damage the motor and this damage will build up until your motors are unusable. So if you touch a motor and it is hot to the touch you should stop what you are doing and find out what is causing it.
That being said, if your motors don’t overheat or break normally with clutches on, you should be perfectly fine ommitting them. From experience I’ve seen that clutches actually don’t make a huge difference when it comes to stress on the motors.
For clarification (although we have discussed this to death before), both the motor and cortex have PTC protection devices to limit damage caused by over current. It depends on how many motors are connected to a protected cortex bank or power expander as to which PTC will trip first, the cortex or the motor. If the motor is used for an extended period it will gradually heat up along with all the internal components, this will mean that it’s PTC will trip sooner due to the elevated ambient temperature. There was a theory that the PTC devices could degrade over time, I ran a test where I tripped an equivalent part to the cortex PTC for 500 cycles and was unable to see any change in performance. Most motors that I have seen that were non functional were found to have destroyed the brushes.
PTC trips seem to happen more than gear breakages. Then then internal gear teeth tend break before total gear failure (see pictures below).
We had a real rash of PTC trips on the first sets of 393 motors but not as much last year’s orders for some reason. Some kids got even ice bags as our matches seem to run every 20-30 minutes during the match day. Once the PTC starts going, it seems to go a lot sooner next time. Practicing/programming for hours on end is another way to load up the motors with wear for PTC trips or gear failure.
I can’t find pictures of teeth failing but we’ve seen that too. The motor might still turn a bit if only one tooth is broken, but you’ll notice something is wrong as it hits the flat spot.
Of course having 393’s programmed to oppose each other will not help matters either. Seen that too. 393’s don’t like that. Shafts might bend first but I bet the motors don’t like that either limiting their lifetime.
Excessive loads on an arm can cause some pain. (Rear gearing to a low gear ratio on a lift arm with 10+ sacks sound like anyone’s plans?) Driving with slamming front to back direction is another source of common pain for the motors and one I have seen most commonly for 393 failure. The pictures below are from a scissor lift under wicked load.
But on the driving side, the sudden stops and starts hurt. The derivative of acceleration is called jerk which a lot of people don’t generally consider. It’s really just computing the impact of the sudden stops and starts. Don’t do that! is the simple answer. Everyone wants a top performing robot so you will tend to abuse these guys in matches. The tiny powder metal (not poured/cast) gears don’t like getting slammed with big loads. Controlling the accelerations/decellerations under load is the biggest spike on the motors.
http://en.wikipedia.org/wiki/Jerk_(physics)
Here’s a good picture of the front of the motor motor gears with failure (393 strength config):
P1010221
Be careful with the screwdriver to open these up as the screws can strip rather easily making the next repair more difficult.
Here’s a picture of the back of motor gears (where you swap out the one for the new I2C encoders). I have not seen these fail but I bet they could.
P1010220
Implementing jpearman’s slew rate management on the motors seems to have helped with the +127 to -127 sudden swings the kids can whack on the motors. (search the forums for this) But too much slew rate makes you slow down much slower too.
More power management for 10 393’s seems to be the programming task for this year.
The motors do have a set lifetime too. While we have had good luck with a bunch of 3-wire motors, they do have a set lifetime of abuse. I think someone mentioned a fairly low (like 50 hour) continuous running life. So 2 Vex seasons is what I think an average motor should be able to last. if you get more, yeah for the long tail on the bell curve!
This is the post Team80_Giraffes is referring to.
https://vexforum.com/showpost.php?p=225727&postcount=25
I’m also currently re-evaluating some of the motor control code. The slew rate control is part of a motor control library that team 8888 uses that also implements a non linear mapping from requested control value to output control values. It has occurred to me that this LUT should be used after slew rate rather then before as is currently implemented. The LUT is a power function that was created before the work this summer looking at the non-linear characteristic of the motor controllers, the current LUT we use is as follows.
[ATTACH]6829[/ATTACH]
The blue line shows how the (unloaded) motor speed varies with control value. The black line is the non-linear mapping applied to the control value and the red line shows the resultant motor speeds with the revised input. It’s still not linear but a better approximation. One day I will also measure a loaded motor but it’s low on my list of to do’s.
Vexmen - can you point out where the failure is in your first photo? I can’t see anything; I have lots of dead motors I look at and can’t see anything wrong with the gears, so I’m not sure when replacing the gears might be useful? Is there something less obvious than broken teeth?
QUOTE=Team80_Giraffes;317907]PTC trips seem to happen more than gear breakages. Then then internal gear teeth tend break before total gear failure (see pictures below).
We had a real rash of PTC trips on the first sets of 393 motors but not as much last year’s orders for some reason. Some kids got even ice bags as our matches seem to run every 20-30 minutes during the match day. Once the PTC starts going, it seems to go a lot sooner next time. Practicing/programming for hours on end is another way to load up the motors with wear for PTC trips or gear failure.
I can’t find pictures of teeth failing but we’ve seen that too. The motor might still turn a bit if only one tooth is broken, but you’ll notice something is wrong as it hits the flat spot.
Of course having 393’s programmed to oppose each other will not help matters either. Seen that too. 393’s don’t like that. Shafts might bend first but I bet the motors don’t like that either limiting their lifetime.
Excessive loads on an arm can cause some pain. (Rear gearing to a low gear ratio on a lift arm with 10+ sacks sound like anyone’s plans?) Driving with slamming front to back direction is another source of common pain for the motors and one I have seen most commonly for 393 failure. The pictures below are from a scissor lift under wicked load.
But on the driving side, the sudden stops and starts hurt. The derivative of acceleration is called jerk which a lot of people don’t generally consider. It’s really just computing the impact of the sudden stops and starts. Don’t do that! is the simple answer. Everyone wants a top performing robot so you will tend to abuse these guys in matches. The tiny powder metal (not poured/cast) gears don’t like getting slammed with big loads. Controlling the accelerations/decellerations under load is the biggest spike on the motors.
http://en.wikipedia.org/wiki/Jerk_(physics)
Here’s a good picture of the front of the motor motor gears with failure (393 strength config):
P1010221
Be careful with the screwdriver to open these up as the screws can strip rather easily making the next repair more difficult.
Here’s a picture of the back of motor gears (where you swap out the one for the new I2C encoders). I have not seen these fail but I bet they could.
P1010220
Implementing jpearman’s slew rate management on the motors seems to have helped with the +127 to -127 sudden swings the kids can whack on the motors. (search the forums for this) But too much slew rate makes you slow down much slower too.
More power management for 10 393’s seems to be the programming task for this year.
The motors do have a set lifetime too. While we have had good luck with a bunch of 3-wire motors, they do have a set lifetime of abuse. I think someone mentioned a fairly low (like 50 hour) continuous running life. So 2 Vex seasons is what I think an average motor should be able to last. if you get more, yeah for the long tail on the bell curve!
We just broke one of those gears today
Was it the original, or an I2C gear?
High speed, or torque?
It was a High Speed motor with an I2C sensor on it. The gear that brok was the one next to the I2C gear. The small part of the gear sheared from the larger part of it.
Since nobody has brought this up, surprisingly, the clutches aren’t meant to be used on the 269 or 393 motors; they’re meant to be used on servos and on the really old 3-wire motors (VEX Continuous Rotation Motor), both of which have plastic internal gears which break very easily. The new motors use metal internal gears.
This is why the new motors include a shaft coupler. As the description on that page explains:
Our motors usually break from broken internal gears but this happens so rarely that it’s not worth putting clutches on because you sacrifice much of the potential torque of the motors because the clutches pop way earlier than the motor’s breaking point. Broken gears usually occur when you make mistakes like gearing 4 motors together but having one of them plugged in backwards, at least in our experience.
We have broke a few gears but since the motors come with speed gears, we have spares of speed, and torque lying around. We replace them, and continue on.
I would like clear motor housings so we can see if it is broken, without opening it.
See page 13 of this thread:https://vexforum.com/t/new-vex-products/24917/1&highlight=clear+motor+housing&page=13
Or this thread:https://vexforum.com/t/imagine-if-the-motors-used-clear-plastic/23756/1&highlight=clear+motor+housing