help detecting dc motor death

Go To Last Post
8 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi guys
I'm building a watercooling sytem for my computer. I'm attempting to make it as silent as possible.To that end I'm going to try and control the speed of the coolant pump with a small avr (probably a tiny25) what I'm trying to figure out is how do I measure DC motor health. I tried googling but I got nowhere.way to many whitepapers on motors that are bigger than I am.I figure its fairly easy to find an open or short in the winding by measuring the current thru the motor. Are there any other failure modes that I'm missing? I figure dead pump has a pretty good chance of equaling dead computer so I'm trying to cover all my bases on this one. Also it occurs to me that it might be possible to measure motor speed by counting current spikes when the motor commutates. Is this even remotely feasible? Any useful links or whatever would be greatly appreciated

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

AC motors would be more effective, much longer life, and there is easy rpm counting.., however they need more complex phase switching controller.
DC motor will open or allmost short in either case I can recall. If seeking for theoretical example when the current is allmost ordinary, but motor is not turning, I would bet that the heat must burn the point of high resistance and thus turning this situation into open or short circuit soon - but it depends also on current of course.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Motor current is the easiest to measure and for the usual DC motor, the current is related to load. You could add a sensor to measure the motor speed as is done with most of the motherboard fans and add a temperature sensor to the cpu. Thus you have three ways to measure if the system is working as opposed to the motor just turning.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

you could use a motor with a simple shaft encoder, even an IR LED hooked and some black and white stripped paper would do and use that as input to an Ext IRQ on your MPU.

there are plenty such motors out there.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

take a piece of metal, stick it to the shaft of the motor, take a CNY70 and count the number of impulses you get per sec. if it goes under a critical number sound the alert and release a big waterbag directly on the processor to cool it ;-)

There are pointy haired bald people.
Time flies when you have a bad prescaler selected.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Measuring DC motor speed based on commutation spikes should be possible. An example:
http://www.romanblack.com/encoder.htm

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Maby you could check the flow of coolant, would also stop the system if it was clogged.

Some bios versions have a thermal shutdown feature if the processor gets to hot, cant get easyer than that (if it is supported by your bios)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

My 2 cents:

1) You might be able to monitor motor current to see if it is rising from nominal. This MIGHT be a predictor to early motor failure.
2) You can monitor motor temperature.
3) I would suggest monitoring CPU temp as your main fault however. That is the parameter you want to fail on regardless of the reason for the failure. Knowing it was the motor that died just helps in servicing.
4) With all this monitoring how does your cooling system shutdown the PC? It does no good to monitor the motor for death if you can't do anything about it!

Go electric!
Happy electric car owner / builder