mosfet Half bridge help, Blowing mosfets

Discussion in 'The Projects Forum' started by woerr, May 18, 2016.

  1. woerr

    Thread Starter New Member

    Apr 5, 2015
    21
    0
    In the following circuit I am blowing the P channel mosfet. some help would be much appreciated.

    [​IMG]
     
  2. woerr

    Thread Starter New Member

    Apr 5, 2015
    21
    0
    Oh, I should probably add that the motor is a 250w 12v brushed dc motor.
     
  3. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,804
    1,105
    It would help to know what the supply voltages are, and what switching frequency/duty cycle the 4429 is using.
    Is the circuit on a breadboard or pcb?
     
  4. Sensacell

    Well-Known Member

    Jun 19, 2012
    1,132
    267
    What is the purpose of the P-FET? Do you need to brake the motor?
     
  5. BobTPH

    Active Member

    Jun 5, 2013
    784
    114
    What is the low / high input voltage? Unless it is near V+, the high side MOSFET might be on at the same time as the low side ones.

    Bob
     
  6. Lestraveled

    Well-Known Member

    May 19, 2014
    1,957
    1,215
  7. woerr

    Thread Starter New Member

    Apr 5, 2015
    21
    0
    Supply is 12v
    frequency is the core speed of the uno32 if im not mistaken, so 96khz (please correct me if im wrong)
    duty is about 50%
    mosfets are mounted on a heatsink, everything else is on pcb

    The voltage spikes coming through the supply voltage is blowing the tc4429 when I dont run seperate supply to the power source. Thus i was told that in a half bridge setup the voltage spikes are recycled thus protecting the rest of the circuitry.
     
  8. woerr

    Thread Starter New Member

    Apr 5, 2015
    21
    0
    I figured that is what the problem was, Is the h bridge setup far superior to the normal nchannel mosfet switching as far as voltage spikes go or should I rather be looking at incorporating a dc-dc power supply to protect the rest of the circuitry?
     
  9. Lestraveled

    Well-Known Member

    May 19, 2014
    1,957
    1,215
    The fix is to slow down the turn-on time and speed up the turn-off time. Put a 10 ohm resistor in parallel with a 1N4148 diode, in series with the gate drive. On the bottom Fets the anode is toward the gate. On the top Fet the cathode is toward the gate.
     
  10. crutschow

    Expert

    Mar 14, 2008
    13,041
    3,243
    Diode D2 will recycle the spike but it should be a larger diode that has at least a 10A rating (preferably a Schottky type).
    The P-MOSFET serves no useful purpose if you have the diode.
     
  11. woerr

    Thread Starter New Member

    Apr 5, 2015
    21
    0
    That is how I was switching, and I have gotten the spikes down to 2v using a rurg5060 and a 100v 220uf cap, I would like to get get it even better though. when the motor starts up the spikes are still too high.
     
  12. woerr

    Thread Starter New Member

    Apr 5, 2015
    21
    0
    [​IMG]

    Top left scope: Diode and 100v 220uf diode over power source
    Top Right scope: only diode
    Bot left : 0.1uf and 1uf over motor terminals
    bot right : close up shot of 100v 220uf cap over power source.

    Yellow line shows power source
    Blue line shows motor terminal signal
     
  13. crutschow

    Expert

    Mar 14, 2008
    13,041
    3,243
    How high is too high?
    I would add some extra 220μF capacitors on the power source.
    250W from 12V means you are switching over 20A and that's a lot of current to rapidly switch.
    A 2V spike at that current requires only 100mΩ of impedance.
    Use as short lead lengths as practical on all the connections, since wire inductance can generate significant spikes at those current levels.
     
  14. woerr

    Thread Starter New Member

    Apr 5, 2015
    21
    0
    I was wrong, Spikes when starting the motor or stopping the motor are almost nothing, When you say add more 220uf caps, do you mean add multiple in parallel as close to the isolated circuit as possible?
     
  15. crutschow

    Expert

    Mar 14, 2008
    13,041
    3,243
    Yes, you want all leads, especially the cap leads. to be as short as possible directly between the supply line and ground. (And you want large gauge supply and ground leads or a ground plane).
    Lead inductance is critical in a circuit with such high switch currents.
    The inductance of a wire can be 20 or more nH per inch.
     
  16. woerr

    Thread Starter New Member

    Apr 5, 2015
    21
    0
    This is a scope shot with 2x 220uf caps, how do I work out what is the best capacitor value to use for a specific size spike.

    My setup is as follows.

    12v battery --(5m long 10mm wire)--> rail -10cm long 1.5mm wire)-> mosfet -(1m long 5mm wire)-> motor -(1m long 5mm wire)-> rail
    from rail -(3m long 1mm wire)-> control circuit

    scope shot is from same points as above using two 220uf caps
    scope2caps.png
     
  17. mejlby

    New Member

    Apr 26, 2013
    28
    1
    Hello Woerr

    I have read your thread

    If you want to make a motor controller that can handle up to 50 amp so I recommend you this circuit

    I have made several of these. and they work perfectly without mosfet heat up.
    The latest I made was to a 800W motor. To this I put heatzink on all 4 mosfet.

    You will with this circuit get regen-braking on your motor

    I have a PCB layout if you are is interested


    Best regard
    Mejlby ( Denmark )
     
    Last edited: May 19, 2016
  18. crutschow

    Expert

    Mar 14, 2008
    13,041
    3,243
    There is no "best" value.
    For high frequency spikes, low inductance ceramic capacitors are usually used, such as the 0.1μF ones typically connected across digital IC pins.
    For lower frequency spikes, such as from the motor switching, the more capacitance, the lower the spike value until you are limited by the capacitor ESR values, and the line resistance and inductance.

    If you use twisted wire pairs for the power/common and to the motor, that will reduce the wire inductance and thus the spikes also.
    If you don't have twisted-pair wires available, you can make them by placing one end of of two identical length wires in a vise and the two other ends in a drill chuck. Then pull the wires taut and operate the drill until the desired number of turns (typically a few twists per foot for large wire).
     
    Last edited: May 19, 2016
  19. ronv

    AAC Fanatic!

    Nov 12, 2008
    3,292
    1,255
    The other thing that might help is to use a twisted pair for the wires from the battery to the circuit. Or if they are to big to twist glue them together in parallel to reduce inductance.
     
Loading...