I have a design based on ucc1625. The rpm needs to vary from 400 to 4000 rpm (closed loop speed control), At low RPMs, the speed control is unstable. I have noticed that the hall effect sensors (120 degree) are not exactly lined up - one is out by 16 electrical degrees. The electronic tack circuit converts this error into a ripply on the tach signal that I believe is caused the motor to run unstable at low rpms--any ideas anyone? I have run loop stability plots with a Venable Analyser and it shows good phase margin and nothing unexpected. I guess there are two questions in here: (1) could the instablilty be caused by tach ripple, and (2) what should I impose a max deviation for hall effect sentor position for the next design?

 

Harold

A hall effect sensor is much to coarse to use as a position feedback for the BLDC motor. A digital encoder or a resolver is used in most commercial brushless DC drives for the position feedback to the servo amplifier for the purposes of phase current commutation.

George W. Younkin, P.E., IEEE Fellow
Staff Engineer
Bull's Eye Marketing, Inc.
Industrial Controls Consulting Div.
104 S. Main St., Suite 320
Fond du Lac, WI 54935
Tel: 920: 929-6544
Fax: 920: 929-9344
E-mail: [email protected]

 

I have used the 3625 version of this chip.
I have not used the tach feature though. I opted to use current feedback loop with PI. Our app was low rpm brushless motor < 300 rpm. We used the hall signals to generate tach and used linear hall sensor voltages for finer resolution.
Part of your problem might be because the tach is really only a 48 count/rev encoder [very coarse]. I don't remember how the chip uses the tach signal to control speed--not sure if it filters the signal into dc voltage and compares it to its triangular reference to set the internal pwm rate. If it uses an RC filter then this has to be sized for low speed.

The hall sensors should all be lined up. It is possible that one of the hall sensors in your bldc motor has shifted?


good luck,
oj

 

Consider sinusoidal motor and 6-step (block) commutation that is almost ideal at low speed.

Imagine 3 sinusoidal constant speed back phase-to-phase EMF curves that represent in a proper scale DC brushless motor torque curves when feeding 2 different phases with a constant DC current.

For correct Hall sensors placement, DC brushless torque generation for ideal block commutation is pretty much similar to ideal voltage rectification - "most positive" of 3 curves is always selected, continuous transitions from curve to curve are made at crossing points every 60 el.deg.

For inaccurate Hall sensors placement, you can easily imagine (or draw) corresponding DC brushless torque curve - less average torque, more torque ripple.

I assume that increased torque ripple that is discontionuous for ideal block commutation is a reason for low speed instability in case of inaccurate Hall sensors placement.

For some other aspects of DC Brushless commutation read "DC Brushless Commutation Theory" article overview at
http://www.drbrushless.com