Motor Speed/Rotation counter discussion

[3v0]

I have been reading up on efforts to sense brushed DC PM motor speed and/or count rotations without optical sensors. It is possible to buy motors and controllers with sensor ability. I am looking to a DIY slant for use with sensorless motors.

A web hunt has resulted in 3 methods.

A) Use a hall effect sensor to get some number of counts per motor rotation.

B) Sample on which touches on it.

C) Read the zero crossings of the inductive spikes caused by the brushes switching poles. Maybe use a comparator.

Neither A or C will work with PWM. Maybe one could allowed the motor to run full time during a periodic sample time. Sampling would change the speed but maybe not enough to be a problem.

C can be effected by the state of the brushes. A does not have that problem and since it does not connect to the motor power supply there is no voltage level problems.

One problem with B is that as PWM goes to 100% you do not have enough time to make the reading. One could either uses a voltage supply high enough so that the motors would develop enough output at less then 100%. About 50% more voltage would work. Or one could toss in a measurement period knowing that it would slow the motor some.

A and C look like they would be fun to play with but B looks like it is the most mainstream and easiest to get going.

I would be interested in hearing from anyone who has played with or is interested in any of these.

EDIT:
Sparkfun Discussion

PayDirt: **broken link removed** (PIC16F88 and CCS C).

 

[Nigel Goodwin]

Stick a low value resistor in one end of a DC motor, connect a scope across the resistor, then apply a DC supply (a battery is good) across the series resistor an motor. Watch the waveform on the scope, consider how you could use it!.

This technique was used extensively for speed control in motors for cassette recorders.

Bear in mind PWM will completely mess this up though!.