3
\$\begingroup\$

I am working on a high-current MOSFET switch project. I would like to implement short-circuit/overcurrent protection, that would be fast enough to prevent MOSFETs from blowing up.

In the actual application, I am using a HAL sensor to measure current, but it won't allow me to take any action quickly because of its delay.

How should I realize this protection? Should I do it by using a shunt and a high-quality op-amp, or is there a better solution?

\$\endgroup\$
11
  • 2
    \$\begingroup\$ Details of the MOSFET switch schematic are needed. \$\endgroup\$
    – Andy aka
    Commented Dec 30, 2020 at 11:56
  • 1
    \$\begingroup\$ In terms of speed, shunt resistor will be hard to beat. \$\endgroup\$
    – winny
    Commented Dec 30, 2020 at 12:09
  • 4
    \$\begingroup\$ How much is "high current"? How fast is "very fast"? \$\endgroup\$
    – Justme
    Commented Dec 30, 2020 at 12:26
  • 1
    \$\begingroup\$ MOSFET switch is composed of 5 MOSFETs in parallel, with 22 Ohm gate resistance each and 1.5A driver. It has to handle 150A continuous current, has been tested on the bench. I don't know how long does it take to blow up MOSFETs while short circuit of battery (passing like 1-2kA), maybe I should ask question about what solution is the fastest one. \$\endgroup\$
    – Bratw
    Commented Dec 30, 2020 at 12:36
  • 3
    \$\begingroup\$ A possibility would be placing a inductor in series that limits the speed at which current can ramp up, buying time for a overcurrent circuit to act. \$\endgroup\$ Commented Dec 30, 2020 at 13:22

2 Answers 2

1
\$\begingroup\$

Use Vds detection with minimal dead time; drivers with short circuit protection use that methods.
EG for high voltage
https://gate-driver.power.com/products/scale-idriver-ic-family/sid11x2k/

If you have inductive loads or long cabling, detecting the current with a shunt is too slow and when you trigger the shutdown the current could take to much time to reverse. Specially with 5 Mosfet in parallel you can have problem to handle the energy in the circuit.

\$\endgroup\$
1
  • \$\begingroup\$ can you explain to me why detecting over-current with a shunt would be slow, for an inductive load? \$\endgroup\$ Commented Aug 20, 2021 at 13:36
1
\$\begingroup\$

One cheap option is an in-PCB current transformer: run the output trace in an inner layer of the PCB, and use vias and perpendicular traces on top and bottom layers to form a solenoid around it. Load the secondary winding with a resistor, and use a fast comparator to detect overcurrent.

Given the currents involved, the output trace can end up as two parallel fairly wide traces on layers 2 and 3, with the current transformer wrapped around them on layers 1, 4 and the vias.

Of course this will only detect transient short circuits, not a slow increase in current - but should be enough for such a purpose.

VDS measurement with deadtime would be a more typical option.

\$\endgroup\$

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.