pwm switching noise

Michael Mina1

Other Parts Discussed in Thread: TMDSHVMTRINSPIN

I'm running a National Instruments dynamic data acquisition system using LabVIEW (that I have written) on the TMDSHVMTRINSPIN and other kits. I calculate the power spectrums on multiple accelerometers mounted on the HVACIM motor, and the servomotor from TI. When I run the kits (seperately of-course), and perform power spectrum analysis on the motor, I see multiple spikes on the spectrum, at 120 Hz, after trubleshooting, I found out this was due to the PWM frequency on the kits. This is unacceptable, can you suggest something to minimize the noise? I'm not using the ADC on the kits to perform the spectrum analysis. This is all seperate data acquisition PCI cards.

over 11 years ago

0 ChrisClearman over 11 years ago

TI__Guru** 119590 points

Have you tried different PWM frequencies?

I know some people implement a dithering of the PWM, varying the frequency slightly ever X cycles. This is pretty complex to do though IMO.

0 ChrisClearman over 11 years ago

TI__Guru** 119590 points

also, how do you know the 120 Hz isn' t from something else in your lab?

do the spikes change at all as you change the speed of the motor (frequency of the inverter)?

0 ChrisClearman over 11 years ago in reply to ChrisClearman

TI__Guru** 119590 points

From some others: The noise is most likely coming from the input rectifier on the HVkit. (it could be coming from anything in your lab that is rectifing line voltage). It creates really good current spikes at 120Hz. You would need a PFC input stage to address this most likely.

To prove, try using a DC supply for the kit and you should see it go away.

C2000™︎ microcontrollers

C2000 microcontrollers forum

pwm switching noise