PWM Motor Control Problem - Spikes on Supply Rail

Discussion in 'General Electronics Chat' started by surfline, Apr 23, 2010.

  1. surfline

    Thread Starter Active Member

    Aug 12, 2009
    30
    0
    Hi,

    I am controlling a 30 amp motor through PWM from a microntroller to a driver to four parallelized mosfets (to split up the current and power dissipation). The core of my schematic is shown in the attachment. The input is coming from a micro controller.

    This part of my circuit works, however I have been having some trouble dealing with the noise and voltage spikes on the supply line from what I believe is the switching motor collapsing field effect. That is the reason for diode D1. Without those Cap's and Chokes (inductors) the 14 V line is all screwed up, however those filters help significantly. However, I am still having a voltage spike up to about 20 volts every 20 kHz (my PWM switching frequency) that these filters don't seem to be getting rid of. That spike is not allowable for the rest of my circuit (rectangle box) and I was wondering if anybody has any idea's to help get rid of the voltage spike.

    I have tried a 15V zener diode to ground with no success. I've also tried putting power diode in series on the 14V supply rail in order to try to block the spike from going back onto the line, however this causes my mosfet driver IC to break. I'm guessing that the current had to flow somewhere and since it can't go back out on the supply rail from my diode blocking it, the current flows into the driver, thus breaking it.

    Any help would be greatly appreciated.

    Thanks.
     
  2. rjenkins

    AAC Fanatic!

    Nov 6, 2005
    1,015
    69
    I'd have both high value reservoir capacitors and smaller ceramic or plastic film to handle the high frequency spikes.

    Is D1 a fast recovery type? If not that could cause current or voltage spikes.

    Also make sure you are using a star point ground common and star point power common.

    In other words, the power supply cables, capacitor cables, motor cable, power FET cables & power out to the control circuit should all be connected at a single point each for ground & power.
     
  3. R!f@@

    AAC Fanatic!

    Apr 2, 2009
    8,740
    759
    All diodes should be ultra fast recovery types across each MOSFET, or you are asking for trouble
     
  4. surfline

    Thread Starter Active Member

    Aug 12, 2009
    30
    0
    Thanks for your suggestions.

    Attached is a picture of the scope of the spiky 14 volt line. This is for an automotive application and the GND i'm using for the scope is the battery GND or chassis GND. If I change the GND to use the GND on my PCB, there is not a spike. Similarly, if I measure PCB GND with respect to battery / chassis GND that same spike also shows up. So I have a feeling its actually the GND that the noise is being created on, not the 14V supply line.

    Any suggestions for this problem?

    Thanks.
     
  5. retched

    AAC Fanatic!

    Dec 5, 2009
    5,201
    312
    Dont use it in a car. ;)

    Diode on the gnd line to suppress back leading spikes.
     
  6. surfline

    Thread Starter Active Member

    Aug 12, 2009
    30
    0
    diode from GND to where?
     
  7. retched

    AAC Fanatic!

    Dec 5, 2009
    5,201
    312
    from gnd to Gnd.

    In line from the PCB GND to the (-) battery terminal
     
  8. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    First, the spikes are not 20V but about 5V.

    Diode D1 should be an ultra-fast or schottky diode.

    The spikes might occur due to the parasitic inductance of the wires or the PCB tracks. It is important to keep the wires/tracks as short as possible for the motor circuit, especially the path around the battery-motor-mosfets-battery.
     
Loading...