N Channel MOSFET Bootstrapping

Discussion in 'General Electronics Chat' started by AetherNZ, Mar 23, 2008.

  1. AetherNZ

    Thread Starter Member

    Mar 23, 2008
    21
    0
    I am trying to design a motor controller using a H-Bridge with all N channel MOSFETs. I will be controlling the H-Bridge with a PIC and only have 12V on the positive rail. I dont really want to use an IC as the high side MOSFET driver and would prefer using basic components with capacitors ect.

    Im after any useful information on how to design the bootstrapping circuit. I have been searching around but have come up empty handed. Has this method been made redundant from the MOSFET driver ICs?

    Thanks for the help
     
  2. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    You might find relevant information by looking up "charge pump". Your guess about it having been made redundant by the ready availability of high side drivers is probably correct.

    Briefly, a high side FET can only stay in conduction when its gate is 10 volts above the source. The gate voltage has to be at least 10 volts above, but never exceed 20 volts over the source voltage. A high side driver takes the ON signal as an enable to run a controlled charge pump that will raise the gate voltage and maintain it dynamically under changing load conditions.

    You can probably look into the description of how a high side driver works and duplicate the function with some transistors, diodes, capacitors and a comparitor or two.
     
  3. Søren

    Senior Member

    Sep 2, 2006
    472
    28
    Hi,

    Do yourself a huge favour and get P-ch devices for the top transistors and then drive them with eg. small signal BjT's.
     
  4. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Here's a boost circuit you could modify to output Vdd+10v:
    http://www.dos4ever.com/flyback/flyback.html#boost
    That one's intended for 12vdc =>180vdc, but you could certainly change values around to get much lower voltage.

    Using P-Ch MOSFETS would certainly simplify things, however it's hard to find P-Ch MOSFETS with Rds(on) as low as the N-Ch MOSFETS, which translates to heat generation and wasted power.
     
  5. AetherNZ

    Thread Starter Member

    Mar 23, 2008
    21
    0
    Ah yes this is bringing up alot more relevant information :)

    That would be alot easier but the aim of the project is to learn how to implement the N-channel ones as they are all you have in power applications.

    Thanks. This is definitely an option. Prefer if I could avoid using inductors as I dont have many around.


    Have attached a voltage doubler which works by closing S1 & S3, charging the cap. Then S1 & S3 are opened and S2 is closed which provides twice Vdd on the output for the high side N-channel MOSFETs.

    I can use BJTs for S1 & S3 but how do you implement S2? Im stuck with the same problem of needing a higher voltage than my supply.


    Some more info: Its driving a small 12V BDC motor by PWM and using FQP32N12V2 MOSFETS.
     
  6. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    All you need to make a perfectly serviceable inductor is some wire (preferably, magnet wire) something to wind it on, and something to keep it from unravelling once it's wound. And of course, a formula is a big help:

    L(mH) = (AverageDiameterSquared x TurnsSquared) / (1000 x 18 x AverageDiameter + 40 x LengthOfCoil)

    Get a Bic pen that's about used up. Pull the guts out.
    The pen's barrel measures 0.330" diameter.
    Now get some stiff cardboard or plastic stock, and cut some 1/2" diameter disks, then punch a couple 0.330" diameter holes in the center of the disks. Might be easier if you make the 0.330" holes first. Then epoxy the disks to the barrel of the pen exactly 1/4" apart. Wind on 112 turns of AWG 32 magnet wire (Radio Shack has it, if you don't have any) - this will take close to 11 feet of wire. Try to make the layers neat and level. When you've completed 112 turns, put a bit of superglue or epoxy on the finished coil, and wrap a layer or two of masking tape on to hold it in place.

    You will then have a coil that measures approximately 100uH.

    You use a P-Ch MOSFET for S2.

    But if you're going to be an N-Ch purist - better look at the inductor option ;)
     
  7. AetherNZ

    Thread Starter Member

    Mar 23, 2008
    21
    0
    Ingenious haha. If I cant get the capacitive way to work I'll fall back on this.


    Of course, I totally overlooked this. A PNP BJT would work alright here aswell, right?

    I'll draw up a schematic...
     
  8. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Yes, but you would naturally have more of a voltage drop across a BJT than a MOSFET, so you'd lose efficiency.

    Don't forget that everything's backwards from NPN to PNP.
     
  9. AetherNZ

    Thread Starter Member

    Mar 23, 2008
    21
    0
    Here is a quick design. Am I on the right track?

    Hoping to control each MOSFET with the TTL logic of the PIC.
     
  10. AetherNZ

    Thread Starter Member

    Mar 23, 2008
    21
    0
    Any obvious improvements that I have overlooked?

    Thanks
     
  11. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Wow, I completely missed this.

    I'm afraid that's not going to work very well. You don't have a boost circuit for the upper side; the cap alone won't do it for you. Also, once you DO get the gate turned on for the upper side, it'll turn off before it gets to Vdd. That's the point of having the charge pump circuit (which you have yet to include) - it creates a level around 10v over Vdd. But also, you have no circuitry to turn OFF the high-side MOSFETs if you DID get the gate boosted to Vdd+10v - they would just stay on.
     
  12. AetherNZ

    Thread Starter Member

    Mar 23, 2008
    21
    0
    Thanks for the reply. I ended up using a more elegant method which worked a treat.
     
Loading...