motor driver with FETs

[joco000]

Hello,

I made DC motor driver with 4 FETs ( IRFZ 44, IRF 4905).
When FETs are on board, driver works OK.

Then I put FETs on wires ( 20cm long) and mount them on big heatsink.

Fets get imidately very hot and fuse soon has burned.

I dont see any shortcut...

Does this wires changed circuit properties so much that controller doesnt work anymore??

Here is schematics that I used
https://www.electro-tech-online.com/custompdfs/2011/10/k166.pdf

thank you for your help
Joe

 

[carbonzit]

Did you insulate them (electrically, not thermally) from the heat sink? Check whether one of the device's pins is connected to the metal tab of the MOSFET if there is one.

 

[joco000]

yes, they are electricaly isolated from heatsink and each other.Drains of FETs are connected alredy on PCB. Drain is connected to the housing of FETs, so D and housing have electrical connection of Q3 /Q4 and Q5/Q6. Please see schematic.

Anyway, I also tryed without heat sink, only FETs on wire and it is the same.

 

[alec_t]

Are you absolutely sure that you have correctly paired the FETs so that each heatsink carries an IRF44 paired with an IRF4905?

 

[joco000]

I will check again.

From your answeres i see that you think ,that additional wires are not cause for problems ?

Ate these FETs very sensitive for electrostatic? maybe i destroyed them when soldering wires ??

 

[Reloadron]

I assume you did not use the kit and pcb board? Should that be true as was mentioned make sure the FETs are electrically isolated (remember the tabs are pins) and make sure they are wired correctly.

Ron

 

[audioguru]

Mosfets have a high gate capacitance that resonates with the series inductance of connecting wires causing oscillation at a VHF frequency which makes them very hot. Therefore a low value resistor (10 ohms to 47 ohms) is used in series with the gate at the gate pin. Try it.

 

[ronsimpson]

My bridge driver and FETs are running at 250khz and the distance from driver to FET is 1cm. 20cm=ouch
Try tightly twisting the gate and source wires together.
I think I can get 25 volts across 20cm of wire for 50nS. The source voltage at the FETs is not the same as at the PCB. You might be killing the FETs with over voltage due to long wires.

 

[joco000]

PWM frequency is 270 HZ, so quite low.

Audioguru, I understand you to put resistor on Gate pin at FET, not at PCB ??

I also have snubber circuit mounted on PCB.

thanks
joe

 

[ronsimpson]

You may be switching at 270hz but the turn on/off edges are fast. 50nS? The FETs are capable of near RF speeds and if they oscillate it will be at full speed. If the circuit worked with 0cm and does not work with 20cm, and the heat sink is properly isolated, then the answer is clear.

 

[joco000]

would be any help if I put some small value capacitors also between Gate and GND at FETs pins to prevent RF oscillations ?

 

[audioguru]

I repeat: Mosfets already have a high gate capacitance that resonates with the series inductance of long wires. A low value resistor in series with each gate at the gate pin swamps the inductance and stops oscillation.
Adding more capacitance will probably make it worse.

Nearly all Mosfet circuits use low value resistors in series with each gate pin.

 

[joco000]

Ok, Thx.
I will come back with results

regards,
Joe

 

[joco000]

Result is a little bit better, but still far away from acceptable.
I read that ferrite bead should help in series with resistor....

 

[misterT]

The schematics show 10k gate resistors (R11, R12).. that is way too much.

Here is something you can try:
If you can, move transistor Q1 and Q2, resistors R10 and R13 and all zeners close to the fets.. you can solder them directly to the fet legs. Add good bypass capacitors to V+ and GND close to the transistor and pullup-resistor. This could help at least for the lower fets.

Why you need the fets to be 20cm away from the board? Are they close to the motor? Try running the circuit withouth the motor.. does the fets get hot?

 

[joco000]

This is a part of bigger project and all hosings are already made. For this driver I need bigger heatsink, but there isn`t enough room for it so I need this long wires.
I think that moving parts directly to FETs is quite good idea.

Anyway, here is a link where author describe problems with FET oscillations
https://www.electro-tech-online.com/custompdfs/2011/10/APT0402.pdf

regards,
Joze