MOSFETs causing microcontroller restart

[stoyanoff]

Hi, everyone! I have a problem with mosfets. I`m controlling the circuit with microcontroller. Here is the schematicsView attachment for the forum.pdf A small remote pcb is controlling an IGBT. I`m using 6N137 optocoupler to optoisolate the controller from the power circuit. The frequency if the pulses passed to the 6N137 is 25kHz. When I isolate the mosfet inverter and connect the IGBT directly to 4049 everything works fine, BUT with this second mosfet inveter connected the microcontroller starts reset. I`m using shielded cables to connect the microcontroller with the remote pcb. Can you tell me how to suppres this disturbances???
Thanks!

PP: the 2 inverters are not pointless, because this way I can use smaller gate resistor and drop down the turn on and off times of the IGBT.

 

[emontllo]

Dear stoyanoff,

It seems that the "Mosfet Inverter" (working as a IGBT gate driver - I agree, a good practice to increase switching speed and reduce loses ) may be having a shoot-trough because both P-channel MOSFET and N-channel MOSFET are connected at the same time and this causes the Vcc 15V to drop temporarily and provably cause a supply reset to your microcontroller.

If this hypothesis is correct, I would suggest two options:

1) Maintain the current architecture but solve the shoot-through with different MOSFETs, NPN or a specific circuitry... this website talks about this and provides good tips http://www.talkingelectronics.com/projects/MOSFET/MOSFET.html

2) (preferred) Use a gate driver IC (specially designed for this purpose). There are inverting ones such as ZXGD3005 and non-inverting ones such as MCP1406.

 

[stoyanoff]

The power supplies of microcontroller and MOSFET inveter are different. As I said I`m using 6N137 for optoisolation. The microcontroller is working on 5V(galvanic separrated) and optocoupler(left side), 4049 and MOSFET inverter are on 15V. I have a disturbance on the 15V part but I fixed it by adding 1000uF cap. I think my problem is somekind of EMP, but I`m not sure.....

 

[stoyanoff]

I`ll try this different chart. May be the small short circuits are generating EMP bursts. I`m out of ideas. I hope this will work!

 

[emontllo]

The 1000 uF capacitor solution is a bit shoddy... having shoot-through means losing power, overheating MOSFETs and may cause premature failure.
Have you checked your microcontroller supply with a scope?
In addition the 6n137 cannot be driven at 15V and the Vo pin needs a pullup resistor to the supply rather than a pulldown resistor. Aplying 15V to 6n137 may cause internal device failure and drain too much current from the microcontroller. Decoupling capacitors in parallel with U1 and U2 supplies would be also needed. Which value of R3 and microcontroller voltage are you using to set the 6n137 LED current?
https://www.fairchildsemi.com/ds/6N/6N137.pdf (detector absolute maximum voltage is 7 V).

I can't understand how can this circuit even work...

 

[picgak]

Pin 6 must be pulled up as shown in the attached fig
6n137 recomended Vcc is 4.5V to 5.5V

 

[stoyanoff]

Don`t be so linear! The high voltage can`t destroy anything! The high current other way.... When I`m drainning very small current because the 4049 has big input resistence everything is OK. But if I try to reach the 6N137 limits it will be impossible at 15V....

 

[emontllo]

The high voltage can`t destroy anything!

The high voltage can destroy an integrated circuit if above its absolute maximum supply voltage.

 

[mtwieg]

Yeah I have to echo what other people have already said. You're not using the 6n137 properly and I'm surprised the circuit works at all given your schematic.

As for the reset issue, shoot through current in the FET inverter could be the issue. How exactly are you generating the supply voltages for the MCU and gate driver? Isolated bench supply?

 

[stoyanoff]

Hi, again! I made a few changes as emontllo adviced me. As you can see in the attached file I`m using only 1 complimentary couple. It works fine without voltage on the power line, BUT agian when I turn on the voltage there, my microcontroller starts reseting. If I remove the charge pump(the complimentary couple) it works fine even with the high power supply?! Do you have any ideas what`s happenning here??
Is there any chance some kind high frequency disturbance generated by the complimentary couple to pass through the optocoupler??
Thanks!

 

[stoyanoff]

The problem is the so called "ringing". High current passes through the MOSFET group causing EMP burst. Any ideas how to deal with this???

 

[pplus]

Poor routed ground on PCB.

 

[stoyanoff]

Impossible! I tried on 2 boards and the second one has 2mm lagre traces!

 

[mtwieg]

The problem is the so called "ringing". High current passes through the MOSFET group causing EMP burst. Any ideas how to deal with this???

Reduce shoot through in the inverter, or use a proper gate drive IC. Also add better bypass capacitance near the inverter.

 

[pplus]

Impossible! I tried on 2 boards and the second one has 2mm lagre traces!

It's not just only width of the traces. Current of the power unit should not be run on the traces of digital ground. Power traces should be seized minimum length. Must be installed blocking capacitors. And so on. Miracles do not happen.