e2e.ti.com

[MKing.2364]

Dear community

I have developed a bldc motor driver using a DRV10983 chip from texas instruments target speed 3000.00 RPM

It works of BEMF and dose not use hall effect sensors

I have achieved good stability but the instrument the motor is designed for needs very very stable speed at 3000 RPM

My question before I waste many stressful hours is can the achieve better stability with hall effects than +/- 1.5rpm in 3000RPM

The only way to understand whats happening is to read my post here in the DRV10983 forum

http://e2e.ti.com/support/motor-drivers/f/38/p/860404/3182399#3182399

basically the raw frequency when i look at hall effects drifts 200.1 Hz 199.9Hz 200.0Hz +/- 0.1Hz

This equates to plus minus 1.5 rpm as the hall effects does 4 pulses per ev rpm = 200hz x 60 / 4pulse = 3000Rpm

Can the STEVAL-SPIN3201 device by ST

by st be finely tuned in sensored mode with hall effect sensors to beat +/- 1.5 Rpm drift in 3000RPM I know this is a tiny percantage of +/- 0.05% of 3000RPM but the instrument has a spinning optical measurment system and any speed variance causes a drift in the readings. The DRV1093 speed hunts from 199.9 to 200,1 Hz around the target 200.0Hz causing a drift in the measurement system

To understand this better please read this many many thanks for looking and advice needed from BLDC experts

http://e2e.ti.com/support/motor-drivers/f/38/p/860404/3182399#3182399

[Andrew Lohmann]

The best you can achieve would be to drive the motor as a stepper motor but there will be a phase shift as it warms up (therefore frequency change). I have found a bought in small Escap or MiniMotor brushless DC motors work very well they probably commutate by sensing the winding current?

You will probably need to work-around the issue some other way and except some frequency fluctuation.

[MKing.2364]

Dear Andrew,

Thanks so much for taking the time to read my post.

Before using a BLDC motor we actually used a high rpm stepper motor in the early stages.

The stepper motor gave excellent frequency or RPM stability,

But the design is for a spinning optical measurement system we found even using half step and 16th step micro stepping the motor cogging between stepper phases made the measurment system bumpy.

The system we have achieved is now very good. We have a tiny RPM speed hunt of +/-1.5RPM in 3000 RPM wich is only 0.05% of target.

We can filter the signal to mask the speed variance and we can use the system as is.

My question to anyone that has done speed control loops with BLDC motors is could +/- 1.5 Rpm drift in 3000 RPM be beaten i.e +/- 0.15 RPM tens times better more stable using a microprocessor and the 3 HALL effect signals that are available on the motor in a tuned PI speed loop. The chip Im using DRV10983 dose not use halls and uses back emf but its a good chip +/-1.5 rpm in 3000 is very good result

I have purchased a STEVAL STM32SPIN3201 motor kit to try and find this out for myself but I am having great difficulty getting the motor to run even in open loop mode I have posted a question on it in the Motor Driver IC's And Boards section https://community.st.com/s/profile/0050X000009yJNaQAM?t=1575445163310

thanks for the help Andrew I appriciate it

[Andrew Lohmann]

I found that putting a shaft encoder in the control loop reduced the drift but not the fluctuation. To improve on that take the shaft encoder out of the control loop and run it feed-forward so that you can correct for fluctuation.

[MKing.2364]

Hi Andrew thanks for the tip.

Unfortunately the system I have has an I2C set point register.

Changing the speed set point register by plus and - 1 bit of the 1024 bit set point makes the speed move by plus or minus 20 RPM so it cant be set better than one bit and drifts 1.5 rpm in 3000rpm

.

If you add an encoder do you measure the frequency of the encoder pulses ? or do you count how many pulses pass in a certain time frame to more accurately measure the motor speed.?

I presume you mean the type that has quadrature encoder so might give for example 1000 pulses per revolution etc.

For now they can accept 1.5 rpm drift and can filter the results to Get a more stable reading.

But for my own interest I am playing with the STEVAL spin3201 kit to see if i can get better than 1.5rpm drift in closed speed loop with hall effects.

I will do the steval STM32SPIn01 in my own time just out of interest over christmas as I am intruiged interested to see if 1.5rpm drift can be beaten with a better control system using the hall sensors instead of backemf