Increasing an 12V H-bride to 24V

Discussion in 'The Projects Forum' started by Kminant, Jun 1, 2010.

  1. Kminant

    Thread Starter New Member

    Mar 7, 2010
    4
    0
    [​IMG]

    I have been lurking around the forum for some time now. Most threads have been of great help as I'm a noob in electronics. Currently I'm working in a project that consists of moving two 24V DC Motors (250W) with H-bridges. The motors will be driving a manual wheelchair by friction applied to the wheels. At the moment the problem arises at the h-brige.

    The input signals of the h-bridge are PWM's. The duty cycle of the PWM is being varied by a pic18F1230. So far the programming is done, PWM's vary as commanded from 60% to 100%, and the motor can be run at 12V without heavy load with the current h-bridge circuit. Once constant heavy load is applied to the shaft the h-bridge breaks. My guess is that the currents passing through the MOSFETs are too high.

    I want to raise the voltage to 24 volt in order to diminish the current. But I fear the Vgs of the MOSFET will be to high since R6 and R5 are the one creating the voltage drop to open and close the gates. I'm I right? Is raising the voltage a good idea to run the motors at full capacity and lower the current?

    Please critique my work so far and help me in any way.

    Thanks,

    Kminant

    P.S. The parts on the Schematic are the one being used.
     
  2. Bychon

    Member

    Mar 12, 2010
    469
    41
    I'm no expert on H-brides, but I've seen zener diodes used to protect the Vgs of Mosfets. Think about adding a resistor to limit current on Gates 3 and 4 and install 4 zeners.
     
  3. Ghar

    Active Member

    Mar 8, 2010
    655
    72
    Raising the voltage won't lower your current. For a motor the current is proportional to torque and voltage to speed.

    I suspect you'll be running into some cross-conduction issues with this as well... the low and high FETs are briefly on at the same time causing large current spikes. I had a lot of issues with this personally with a similar gate drive scheme.
     
  4. Bychon

    Member

    Mar 12, 2010
    469
    41
    Could you fix the crossover problem by putting (2) 13 volt zeners between the gates and driving the signal in between the zeners? Or would that cause other problems?
     
  5. Kminant

    Thread Starter New Member

    Mar 7, 2010
    4
    0
    Zener Diodes limit Vgs to 12V hence protect the MOSFET. Am I correct? Zeners are always used between gate and source?

    Making this assumption i came up with this. Please feel free to scream if something is about to POPcorn!
     
  6. Kminant

    Thread Starter New Member

    Mar 7, 2010
    4
    0
    Sorry for the double post! I forgot to say thank you!

    :D
     
  7. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    You MUST provide "dead time" between turning off the high side of the H-bridge and turning on the low side of the bridge (or vice versa) or you WILL fry your MOSFETs to a crisp.

    Both MOSFETs that you are planning on using have large gate charges. They will take (in electronic terms) forever to charge/discharge via a 1k resistor. They will spend that time getting extremely hot, and burning up.

    The 1N540x series diodes are very slow 3A power rectifiers. They will not work well in your application. You need fast recovery diodes.

    Take a look at the attached. You'll likely need beefier MOSFETs than I've shown, but the gate charges for the MOSFETs are a small fraction of the MOSFETs that you are using.

    Note that each MOSFET has individual controls. They also have the gate voltage limited by a Zener diode controlling the min/max gate charge/discharge voltage.

    Rs1 and Rs2 are current sense resistors. If you do not wish to monitor the motor current, you can replace them with wire.

    If you wish, you can replace R3 & R10 with 1k 1/2W resistors and R4, R11 with 2K 1W resistors. There will be a small performance hit, but you'll use less power in the driver circuit.

    At the same time, replace R6 and R13 with 2k 1W resistors.
     
    Last edited: Jun 1, 2010
  8. Kminant

    Thread Starter New Member

    Mar 7, 2010
    4
    0
    SgtWookie,


    Can a 10k resistor do the work of charging/discharging? Hence increase "dead time". Or am i missing something here?

    Are 1N4742A (zener) with 1N4148 (swtch diode) in series fast enough? Reverse recovery time is 4ns, is this what you mean by recovery time?

    Thanks,

    Kminant
     
  9. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    If you don't mind your MOSFETs glowing orange right before they explode, I suppose you could.
    Yes, you are missing something.
    You will need to program the dead time delay in the microcontroller, so that it is impossible for both the high and low side MOSFETs to be on simultaneously.

    The way to do this is to look at the total charge on your MOSFET gates, look at how long it will take to discharge the MOSFET gate via the gate drive circuit, and look at the datasheet specification for the turn-off time. Once you get that all figured out, add in a bit more for safety.

    If your MOSFETs emit great quantities of smoke as they explode, you probably did not have enough dead time.

    1N4148 diodes are plenty fast enough. However, they are only rated for 100mA, so they too will disappear in a bright flash of molten glass.

    Did you look at my schematic?

    I think you believe that you can fix your schematic. When it looks like the one I posted, then it will be fixed.
     
Loading...