Shop OBEX P1 Docs P2 Docs Learn Events
My H-Bridge is overheating. Can somebody help me? — Parallax Forums

My H-Bridge is overheating. Can somebody help me?

potatofacepotatoface Posts: 23
edited 2014-02-13 05:54 in General Discussion
I made this H bridge motor controller using four MOSFETS from sparkfun. .

https://www.sparkfun.com/products/10213
https://www.sparkfun.com/products/10349

and this bipolar transistor.

https://www.sparkfun.com/products/8928

circuit.jpg




I have made two versions of this H bridge, one with 2n2222 bipolar transistors from radio shack, and the other with the sparkfun BC547.
Im trying to toggle the FWD connection on for about a second, just to test if it works.
but both boards do the same thing: Overheat really quick.
The first one actually melted the tinning on the mosfet heat sinks.

So, as soon as I connect the DC power supply (12v, 9v, 6v ...etc), the mosfets get really really hot within seconds. The 12volt computer power supply im using wont even start. It detects a short and wont even power up. (ive got the forward and reverse lines connected to ground with a 100k resistors and also connected to the P1 and P2 pins on the prop demo board.

The 9 volt will spin the motor for a few seconds before the h bridge overheats, but it seems like the H bridge isnt completely shutting off and the motor doesnt exactly stop right away when I toggle P1 on my propeller demo board. And it seems like the voltage from the battery has been halved when it gets to the motor. The motor barely spins. Its not a big motor. Just a little cheap motor that you would find in an old battery powered cassette player, so I doubt I'm pulling over 10 amps.


So what the heck am I doing wrong here? Using the wrong Bipolar transistors? Not putting diodes on each transistors ground? Are the connections floating so each mosfets is fighting the other?


thank you for the help!

-Joe.
circuit.jpg
610 x 700 - 43K

Comments

  • jmgjmg Posts: 15,173
    edited 2014-02-12 19:53
    potatoface wrote: »
    I made this H bridge motor controller using four MOSFETS from sparkfun. .

    Such cross-coupled at the power stage circuits are rather primitive, and prone to crow-bar current effects.

    With Logic level MOSFETS you are better to drive all the gates from drivers, not from the 'other-side' of the power stage.
    Better designs include a lock out, so it is impossible to try to drive all 4 FETS at once.
  • potatofacepotatoface Posts: 23
    edited 2014-02-12 20:03
    Do you have a link to a decent beginner/DIY H bridge schematic? I am trying to drive two vehicle wiper motors for the end result.
  • jmgjmg Posts: 15,173
    edited 2014-02-12 20:19
    Google finds this example, which looks better (ok for modest voltages, as there is no Gate Voltage protection )

    http://letsmakerobots.com/node/24086

    This one has lock-out so both Top FETS cannot be on at the same time.

    It Drives normally each way on A=H, B=L OR A=L , B=H , and if A=H, B=H then both PFETS are off and Enable allows both N fets to be on, so you can control a short across the motor for braking.
    A=L,B=L (or floating) is all 4 FETs off.
  • ercoerco Posts: 20,256
    edited 2014-02-12 20:20
    Wiper motors take a lot of current, much more than your 2222 transistors can handle. For just $2.86, give this dual H-bridge L298 a shot:

    http://www.ebay.com/itm/Dual-H-Bridge-DC-Stepper-Motor-Drive-Controller-Board-Module-Arduino-L298N-2014-/181298308755?pt=LH_DefaultDomain_0&hash=item2a3638a693

    See discussion at http://forums.parallax.com/showthread.php/144881-2.86-Motor-Driver-Board-(Must-Try-to-not-Buy...)
  • Bob Lawrence (VE1RLL)Bob Lawrence (VE1RLL) Posts: 1,720
    edited 2014-02-12 20:34
    http://www.bobblick.com/techref/projects/hbridge/hbridge.html

    http://www.edn.com/design/power-management/4314217/Microcontroller-drives-H-bridge-to-power-a-permanent-magnet-dc-motor


  • Beau SchwabeBeau Schwabe Posts: 6,566
    edited 2014-02-12 20:57
    As mentioned, when you cross couple the NMOS transistors on the bottom like that, you have a crowbar effect.

    One thing is that you want your Supply to the H-Bridge to be more than 9V, AND it needs to be solid, meaning that a heavy load shouldn't pull the voltage down. If it does you can run into runaway issues.

    It's also a good idea to have a separate power supply to the gates that isolate any voltage sagging mentioned above that can happen due to a heavy load. This can be a simple diode capacitor arrangement, but the important thing is that the power supply to the gates is equal to or greater than the supply to the Source of the PMOS transistors.

    Also note: there is a 10k resistor on the upper left PMOS gate and the lower left NMOS gate that should probably be a 1k to be in character with the other pull-up or pull-down resistors in the circuit.

    Quick question... Are the PMOS transistors getting hot and the NMOS transistors remaining cool? This is typical of load voltage sag with inadequate gate voltage drive.
  • potatofacepotatoface Posts: 23
    edited 2014-02-12 21:42
    jmg wrote: »
    Such cross-coupled at the power stage circuits are rather primitive, and prone to crow-bar current effects.

    With Logic level MOSFETS you are better to drive all the gates from drivers, not from the 'other-side' of the power stage.
    Better designs include a lock out, so it is impossible to try to drive all 4 FETS at once.


    could you explain that a bit? What exactly is the crowbar effect?

    The mosfets are being driven by bipolar transistors. Could you also explain what you meant by "not from the other side of the power stage"?
  • jmgjmg Posts: 15,173
    edited 2014-02-12 22:17
    potatoface wrote: »
    The mosfets are being driven by bipolar transistors. Could you also explain what you meant by "not from the other side of the power stage"?

    If you look at your image, the N-FETS are not driven by bipolar transistors, they are driven by the P-FETS.
    In contrast, the link I gave has all mosfets driven by bipolar transistors
    potatoface wrote: »
    could you explain that a bit? What exactly is the crowbar effect?

    Imagine a slow ramp on a P-FET Drain, and no drop yet across the Motor.

    BOTH N-FETS are conducting, so the P-FET has to fight the N-FET on its own side, to raise Gate Voltages.
    It is that direct current pathway, that is called a crowbar effect - best avoided.

    Eventually, the drop across the Motor starts to help, and the crow-bar-ing N-FET starts to turn on less than the opposite side, and will finally turn off as the other N-FET saturates.
  • Mark_TMark_T Posts: 1,981
    edited 2014-02-13 04:16
    That circuit should not have any issues with shoot-through (not "crow-bar effect"), if its driven correctly - you must not
    activate both lines simultaneously. Note that both p-channel gate pull-up resistors should be 1k, not 10k. The circuit
    has some weaknesses (in particular the devices aren't being driven from low-impedance drivers)

    You always need to commission a high power bridge carefully - for instance always first test with a current-limited supply and
    a low-current load, checking all the voltages on each input, gate and output before proceeding to a higher current setting and
    more demanding load. Just connecting at full power and hoping is foolish, you are likely to blow things up that way if something
    is up.

    The behaviour you describe strongly suggests something is mis-wired or not working.

    [ and get rid of the 1N4001 rectifiers, they aren't doing anything useful ]
  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited 2014-02-13 05:54
    Hi potatoface;

    That circuit diagram of the H-Bridge is not correct.
    It shows the High Side P-Channel MOSFETs are wired in backwards, or upside down.
    The Sources should be connected to +12V.

    With the P-Channel MOSFETs connected correctly your circuit should work fine.

    As others have said, there may be other problems.
    But your primary problem in the P-Channel MOSFETs connected incorrectly.

    Duane J
Sign In or Register to comment.