3 shunt current measurement

ivan239955 · ‎2015-11-16

Posted on November 16, 2015 at 16:16

Hi everyone,

I'm using the following evaluation boards:

- STM32303C-EVAL

- STEVAL-IHM023V3

I'm trying to make FOC algorithm for ACIM. The whole FW is written by myself excluding the standard peripheral lib from ST. At the moment the motor is controlled with the U/f algorithm.

PWM timer is configured to 200 us. ADC is configured for measuring the currents, when the bottom IGBTs are ON. The trigger of the ADC interrupt comes once per PWM cycle in the middle of the current pulse, when the current flows through the shunt.

I've soldered up the 150 pF capacitors on the STM32303C-EVAL board, so that the RC time constant is 150 ns.

The measured current is sent via UART and built in Excel.

The result of measurement you can see in the attached picture.

As you see the current indeed has a sine form, but it also has some low frequency noise, that I can't filter. In my opinion this kind of current signal isn't good enough for FOC.

Can you advice me something, how can I improve it?

Thank you in advance,

Ivan

Gigi · ‎2016-01-07

Posted on January 07, 2016 at 18:21

Ciao Ivan

First of all congratulation, you did a great works all by yourself.

My personal suggestion is to try to understand from where is coming this noise. It can be due electrical noise due switching inside the PCB? The sampling time is to low or to high? Did you consider that the Ton of the low side can become too small for current reading? The motor current have the same noise?

We didn't use any filtering network and adopt the technique to skip the noise. Maybe you can think something like that.

Another tip is to turn off any SMPS that can be present in the board to see if the noise is coming from that side.

I have no other particular brilliant ideas, for the moment.

Ciao

Gigi

re.wolff9 · ‎2016-01-10

Posted on January 10, 2016 at 16:36

I have a feeling the ''spikes'' I see are always the same height. As if it is a quantisation error. On the other hand, that is not the case, as you can see the vertical resolution is enough. So it seems that your ADC is reading XX mV too much every now and then. Could it be that a significant ground-current is occasionally present in your circuit?