1
\$\begingroup\$

I am just getting into electronics and my first project is a 555 timer PWM (0-100) and an H-bridge to drive a motor for my brewing equipment (500 mA max. rating).

Here is the setup:

enter image description here

I have a 12 V circuit, a mechanical switch to set the direction (powers up either "left side" or "right side" MOSFETs of the H-bridge). As I was afraid of shoot-through I added an extra P-channel MOSFET that is driven by the PWM and supplies the load.

The problem: Occasionally one of the MOSFETs (low-side) just burns after a while (gets hot/ smells and drain and source are shorted now), the motor stops and only has short flickers of movement.

Observations: I have the feeling that the pot position is connected to the problem. I think that it happens when the motor is insufficiently supplied with voltage and stops moving. If I stay in the 50% range or full load range, I have the feeling it works well, but still only a feeling.

I experienced that this happens much faster if the pot is 100 kΩ and not 10 kΩ, but I have no clue why.

Now that one of the MOSFETs has burned, I tried checking if the PWM is flawed and surprisingly it only does not work if the switch is supplying the side of the broken MOSFET (I disconnected the out pin from the circuit for checking that). The other way round the PWM is fine.

Can you tell me where I messed up?

\$\endgroup\$
16
  • 1
    \$\begingroup\$ Sounds like it is related to low duty cycles and induction pulses of the motor, but I need to think a bit about your free wheeling diodes. What make/type of motor is that actually? \$\endgroup\$
    – datenheim
    Feb 5, 2023 at 9:41
  • \$\begingroup\$ It is a Grothen G928 motor: see here link \$\endgroup\$
    – hAndyman
    Feb 5, 2023 at 9:47
  • 2
    \$\begingroup\$ And please note NPN and PNP has nothing to do with MOSFets, please, those have N- or P-channels ;-) \$\endgroup\$
    – datenheim
    Feb 5, 2023 at 9:53
  • 1
    \$\begingroup\$ Can you try an experiment for me. Reduce R3 and R4 to 1k and see if that makes any difference. \$\endgroup\$
    – RoyC
    Feb 5, 2023 at 10:21
  • 1
    \$\begingroup\$ (Currently, a double-pole switch would functionally replace the H-bridge.) Do you actuate Switch1 while the motor is running? \$\endgroup\$
    – greybeard
    Feb 5, 2023 at 11:12

1 Answer 1

0
\$\begingroup\$

I think I fixed it. It seems that the MOSFET gates on the low side were powered via the negative pole and therefore the N-channel MOSFET on the "off" side was slightly ON which lead to fast burn out of the MOSFET. Found out via an LED that I placed instead of the MOSFET between Gate --> Source of the N-channel MOSFETs. Although the breadboard worked and the through hole soldered version did not, I found out that the LED also responded on the breadboard. But much much less bright. So it seems that both circuits were faulty but the breadboard circuit had too much resistance for the gate to turn the MOSFET even slightly ON. I checked all the resistances and found that the connections on the breadboard had much more resistance (I had to wire 2 breadboards together because one was too small). Sometimes the difference was even 10 to 100 Ω! Here is the new circuit. (The pulldown resistors did not change much/anything. The MOSFETS burned at 50k and at 1.5k) And the circuit has a different, namely a "ON-OFF-ON" switch. Therefore there is a small pause between forward and reverse. I don't know if shootthrough would occur with the depicted switch but I did not have a proper model for the switch. All the best Andi enter image description here

\$\endgroup\$
3
  • 1
    \$\begingroup\$ The correct terms for the state of a MOSFET (or BJT) is ON and OFF. "Open" implies an OFF state, but from context it seems you mean ON. In your new schematic, the diodes D43 and D44 will limit the low gate drive to about 0.7V and may be above threshold. It would be better to put a diode from the common sources of X15 and X16 to GND. You probably have other problems as well, due to use of proto-boards. And you really need to get a scope to troubleshoot properly - even a cheap $30 unit would be OK. \$\endgroup\$
    – PStechPaul
    Feb 13, 2023 at 20:26
  • \$\begingroup\$ Thank you for the feedback! I corrected it. What kind of troubleshoot unit do you have in mind? I don't understand your approach of replacing D44/D43 with a diode from the source of X15/X16 to GND. The voltage that powered the gate of X15/X16 originated from the ground side as far as I could tell. I disconnected the switch from the MOSFET in question and checked it. So the problem would persist if I allow current from the GND side to the MOSFET gates?! \$\endgroup\$
    – hAndyman
    Feb 14, 2023 at 8:06
  • \$\begingroup\$ You really need to check the gate voltages, which need to be solid. And you need to consider that you are only driving the load high during the PWM ON. If you will be driving a motor, this will be problematic, I suggest that you use a conventional bridge circuit. By deciding on this weird design, you "own" it and need to troubleshoot accordingly. IMHO this design is fatally flawed, and I can't help. \$\endgroup\$
    – PStechPaul
    Feb 14, 2023 at 18:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.