Is there a H bridge circuit that uses only N Mosfet?

Discussion in 'General Electronics Chat' started by KaiL, Oct 19, 2015.

  1. KaiL

    Thread Starter Member

    Aug 30, 2014
    57
    0
    I try google around but I can only find H bridge that uses both N and P Mosfet.

    I tried to modify it but it seems to be wrong.
    The attached files are my circuit

    Can anyone helps me with this?
     
  2. GS3

    Senior Member

    Sep 21, 2007
    408
    35
    I have in my hands an H bridge which uses IRF830 but but, obviously, you need the correct switching circuitry. I have not looked at your design in detail but it looks like it is missing some necessary control circuitry.
     
  3. ian field

    Distinguished Member

    Oct 27, 2012
    4,413
    782
    When your H-bridge has only N-channel MOSFETs, you need a high-side driver that can take the gates more than VGSthr above Vdd.

    There are high side driver chips off the shelf that incorporate a charge pump bootstrap circuit to develop the required voltage rail.
     
    cmartinez likes this.
  4. crutschow

    Expert

    Mar 14, 2008
    12,991
    3,227
    As Ian said, you need a proper high-side (normally bootstrap) driver to generate the required Vgs voltage to properly turn on the upper N-MOSFETs.
    A standard MOSFET typically requires a Vgs of 10V to fully turn on (not Vgsthr), and logic-level types require 3 to 5V depending upon their design (as shown in the data sheet where the Rds value is specified).
    This Vgs voltage must be above the N-MOSFET drain supply voltage for the bridge.

    Note that the bootstrap driver, depending upon its design, may not work for very low frequency or static switch bridge application. Those types require a continuous AC bridge drive to charge the charge-pump.
     
    cmartinez likes this.
  5. MaxHeadRoom

    Expert

    Jul 18, 2013
    10,510
    2,369
  6. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066
    That is about the worst possible circuit. It has terrible power dissipation and shoot through.
    Read this forum thread.

    You use fewer parts if you buck up and buy some PFETs. Using PFETs in the top half of the Bridge avoids having to dick around with complicated gate drivers for the top FETs.
     
    Last edited: Oct 19, 2015
    cmartinez likes this.
  7. ian field

    Distinguished Member

    Oct 27, 2012
    4,413
    782
    In any equipment that runs from a mains transformer, you can run a voltage doubling charge pump from the secondary to get the required high side Vgs. Its possible with a SMPSU, but you have to understand the shape of the switching pulses and the rectifier configuration to design the circuit.

    With a DC supply there's little choice but to bootstrap the bridge node with a charge pump to generate the required voltage. If all else fails - its possible to include an inverter such as a blocking oscillator to produce a rail higher than Vdd.
     
  8. Russmax

    Member

    Sep 3, 2015
    81
    12
    A bootstrap circuit is needed to power the gate drive for the high-side NMOS. The supply for the high-side gate drive uses the output node as its "gnd".

    For the high-side gate drive positive supply node (bootstrap node), hook a diode between VDD and the bootstrap node. Also hook a capacitor between the bootstrap node and the output node.

    When the output is low, the cap charges bootstrap to VDD-0.7 through the forward-biased diode. When the output is high, the diode reverse biases, and bootstrap rises to OUT+VDD-0.7. The circuit driving the high-side gate see a constant supply of VDD-0.7, but the whole circuit rises and falls VDD-0.7V above the output node.

    Regards
     
  9. GS3

    Senior Member

    Sep 21, 2007
    408
    35
    Yup. I just posted this schematic in another thread. Here it goes again:

    bootstrap01.png
     
    Russmax likes this.
  10. Russmax

    Member

    Sep 3, 2015
    81
    12
    Yep, the 1N4007 and the 4 uF are the bootstrap circuit. I'm a little wary of that optocouple gate drive, though. MOSFETs tend to need an active discharge path that's at least as fast as the charge path, and this circuit only provides active charging of the gate. It all depends on the size of the FETs and the switching frequency, I suppose.

    Regards
     
  11. GS3

    Senior Member

    Sep 21, 2007
    408
    35
    Yes, that is switching at 50 Hz so no problem. It is from a 50 Hz inverter.
     
    Russmax likes this.
  12. ronv

    AAC Fanatic!

    Nov 12, 2008
    3,282
    1,234
    What is the voltage at the top of the H? The supply should be 8 - 10 volts higher that this.
     
  13. GS3

    Senior Member

    Sep 21, 2007
    408
    35
    What supply? The Fet's Gate is 12 V higher than the Source (C1).
     
  14. ronv

    AAC Fanatic!

    Nov 12, 2008
    3,282
    1,234
    Your schematic in post #9 does not show the voltage on the drain of the top FET, but it shows 12 to 15 volts going to the opto. So if the voltage at the top of the H is 12 volts, it won't work.
     
  15. GS3

    Senior Member

    Sep 21, 2007
    408
    35
    Edit: that diagram is just half of the H bridge to illustrate the question being asked.
     
  16. ronv

    AAC Fanatic!

    Nov 12, 2008
    3,282
    1,234
    I understand, but can't answer the question without knowing what the voltage is at the top of the bridge (and now) where it says 12 to 15 volts. :D
     
  17. GS3

    Senior Member

    Sep 21, 2007
    408
    35
    What question can you not answer? Because the question being asked here is Is there an H bridge circuit that uses only N Mosfet?". I supply that drawing as an example of an H bridge using N Mosfet. The details of the specific circuit and application are irrelevant here. That H bridge can switch whatever you want within its limits. 80 Volts? Yup. 150 Volts? Yup. 230 Volts? Yup. 300 volts? yup. And If you need to switch more volts than those FET can withstand then just substitute higher voltage FETs.

    Why the hangup? Whatever the specific voltage of the specific application is irrelevant here in this thread. But since you seem so obsessed I will tell you that this specific case is an inverter switching 230 volts AC rectified which off the top of my head I believe comes to a bit over 300 Volts. And it switches at 50 or 60 Hz.
     
  18. ronv

    AAC Fanatic!

    Nov 12, 2008
    3,282
    1,234
    There, not so hard. :D

    Yes, is the answer.
    Look at something like this:
    http://www.irf.com/product-info/datasheets/data/ir2110.pdf
     
    cmartinez likes this.
  19. ian field

    Distinguished Member

    Oct 27, 2012
    4,413
    782
    In an earlier reply, I made the comment that the Vgs bootstrap rail needs to be higher than the bridge Vdd. That could be a little ambiguous, the gate voltage needs to reach about 8 - 10V higher than the source to turn it fully on, since then the source voltage should be practically the same as the drain, it amounts to the same thing - but what I said didn't make that perfectly clear.

    If Vdd is no higher than about 15 - 18V, the 1N4007 can supply the bootstrap rail from it, that gets things started although the high side MOSFET can't pull the output node fully up to the Vdd rail until the bootstrap rail has been charged up to its steady state voltage.

    Most power MOSFETs have an ABS-MAX Vgs rating around 15 - 18V or so, its not a bad idea to clamp the bootstrap rail to protect the gate from excess voltage.
     
  20. MaxHeadRoom

    Expert

    Jul 18, 2013
    10,510
    2,369
    The one's in Tahmid's blog I posted all work fine.
    Max.
     
Loading...