Quote:
|
Originally Posted by scottm87
Couldn't you just measure the temperature of the motor? (some kind of thermistor or temperature sensor...) Given that it takes only one digital or analog input (depending on how much circuitry and what sensor) and skips 5 steps (and the associated error), that would be my solution. You may have to scale it considering you are measuring the temperature on the outside, and the temperature on the inside is proportionally hotter.
Although it might be useful to have the current for the motors measured anyways...
- Scott
|
The only reason I think a model for the temp rise would be better than a thermistor or some other measurement technique is that the temp you really care about is the temp of the armature. This is not so easy to get a real temperature for because it spins. The delay between what you can measure and what the true temp of the armature is could be significant.
This brings up another possible measurement scheme. If you know current and voltage AND speed, you can infer temperature. Basically, you know voltage and speed so you know, theoretically, what your current SHOULD be if the armature were room temp. To the extent that the current is different (in this case lower) than predicted, it must be that the resistance of the motor has increased due to hotter wire in the armature.
Since the resistance of copper wire is pretty well understood, you should be able to get a pretty good estimate of the temperature of the armature using this method, assuming that the current dynamics are fast with respect to the motor dynamics which is typically a pretty good assumption.
Again, this may sound complicated, but once it is worked out, it runs in the background more or less for free, keeping your motors at their peak without allowing them to overheat. I suppose with a bit of tuning, a generic code could be used by any team for any motor that they measure current and speed.
Once I am king and have minions at my disposal, I will be sure to put a few of them on this task ;-)
Joe J.