I’m implementing control algorithms for a centrifugal pump motor. I’m a contractor, so this is just the latest stop on the itinerant engineer rodeo circuit.

The motor being used is a SARNS 164267, 6 pole pair bldc. Old. Part of the companies legacy systems. There are no specs. Well let me correct that: I have a spec sheet that bears very little resemblance to the motor, signed by A. Crapo.


1. Primarily, I’m looking for a method of measuring the actual motor output in terms of rotations per time period. The pump head is magnetically coupled, therefore no shaft to use for an rpm measurement.

2. I’m interested in how to derive BLDC motor transfer functions, and motor specs from a black box motor.

3. I’m always interested in information on FOC, sinusoidal control, flux estimation, Park and Clarke transform implementation, torque control techniques etc.

Useful suggestions appreciated.

 

Do you have access to any of the motor leads? If so, then measure the electrical frequency. That will have a fixed relationship to speed if the motor is synchronous.

 

Verne Lund... use a Vibrating-Reed Tachometer!

Regards, Phil Corso

 

I can measure phase frequency and Hall sensor feedback. Which is another issue: the PI algorithm is not holding the speed constant. But I'll fix that.

I need an instrument to measure RPM for system verification of algorithm from a black box point of view.

 

Put a scope on the signal and see what it looks like. If it is fairly clean you may be able to use an isolator and low pass filter and then a digital tachometer (e.g. ones that may normally use pick-ups off gear teeth).

If it's really noisy, then you could do some advanced digital signal processing on it, but at that point you should be re-evaluating the project to see if the motor is really worth that much time and money. I have written motor analysis software, and a system to do that properly costs a lot more money than any small motor/pump combination that I've seen.