MOSFET h-bridge question

Discussion in 'The Projects Forum' started by benchaz, Feb 4, 2009.

  1. benchaz

    Thread Starter Member

    Feb 2, 2009
    14
    0
    Hey guys,
    I have a few questions. In this schematics of a mosfet h-bridge, http://www.cadvision.com/blanchas/hexfet/np-s.htm
    he has a 47k resistor connected to the base of the npn bipolar transistor, and 33k to the pnp bipolar. Why is it designed like that? Can't both values be the same? What differences does it make?

    Also, I've read about using opto-couplers to connect a h bridge to a microcontroller to eliminate noise from the motors. I have a nand logic circuit that takes two inputs, enable and direction, and outputs it into the A, B, C, D terminals in that h bridge. Should i put the opto-couplers between the h bridge and the nand circuit, or the microcontroller and the nand circuit?

    Thirdly, I want to do PWM for the h bridge. would it be better to keep one mosfet on the whole time it's going in one direction, while switching the other, or swtich both mosfets at the same time, on and off. Does it make a difference? If so, should i send the pwm signal to the N channel or P channel?


    Thank you guys so much!
     
  2. thingmaker3

    Retired Moderator

    May 16, 2005
    5,072
    6
    That's a good question. I've spent the last 20 minutes pouring over the datasheets for the transistors. I can see no reason for the different resistor values. Use a bare minimum value of 120 Ohms for a 12V system if using older BJTs with 100mA max Ib.

    I'd put them as close as possible to the noise source.

    I see negligable difference with these 4 particular transistors. The minimum pulse width from your uproc won't be as small as 625nSec, will it? If so, you can switch the 3904's a smidge faster than the 3906s. (The MOSFETs will be waiting patiently for the BJTs to make up their minds.) Make darn sure you allow time for one set of transistors to turn off before turning the other set on. Shoot-through causes pretty blue flashes and a curl of smoke.
     
  3. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    The majority carriers is a npn transistor are electrons and in a pnp transistor are holes. Holes have less mobility than electrons and thus a pnp transistor can conduct the same amount of power in a bit more time than a npn transistor.
     
  4. benchaz

    Thread Starter Member

    Feb 2, 2009
    14
    0
    But if i wanted to save a few chips, would it be fine for the nand chips to get the noise from the motor? and only separate the microcontroller from the nand circuit?
     
  5. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    Post your schematic.
     
  6. thingmaker3

    Retired Moderator

    May 16, 2005
    5,072
    6
    I think it would be good enough - if you keep your power and ground traces for the NAND chip in wiht the traces for the H-bridge instead of with those for the uproc.
     
  7. benchaz

    Thread Starter Member

    Feb 2, 2009
    14
    0
    Thanks, ill do just that
     
Loading...