Protected 24V, 200A Switch

[luky315]

I need to switch 200A to an approximately resistive load (slightly inductive) and i would like to protect the switch from a short curcuit on the output (R < approximately 0.1Ohms). So the solution would be a high power e-fuse design, but the question is which solution would be the most robust and cost effective? Is a hot swap controller worth the additional cost or would a simple solution with a comparator controlling a few paralleled MOSFETS be ok?

[Faringdon]

200A is your 'normal' current, so the short cct current must be immense.
One good way for you is to do multiple current source paths to the load....and have cheap B2B IGBT switchs in each of them...so then each has less current to break.
Also have a fuse in each of the current source paths.

[luky315]

The point is that the short circuit current has to be limited by opening the FETs in time. And IGBTs are not the right choice with 24V

[T3sl4co1l]

You would need quite a large MOSFET for a regular hot-swap controller to do the job, and you can't parallel them when linear operation is used; or not without a considerable amount of source degeneration, which defeats the purpose (big voltage drop at nominal current).  Even then, at some point, too much gate charge will make the control unacceptably slow, or unstable.

A switching design, probably multi-phase, would be effective, but you'd have to create your own as I don't think such a thing exists.

Or single phase if the "slightly inductive" load is sufficient, and is okay with being chopped (and this doesn't emit destructive EMI throughout your system, etc..), but honestly at these currents, I'm okay with the added complexity of multi-phase control to save on filter capacitance, switch node limits, and scaling in general (SMD transistors and off the shelf inductors will suffice).

Tim

[coromonadalix]

some ssr relays  ??
https://www.ato.com/industrial-solid-state-relay-200a
https://www.aliexpress.com/item/4000295498163.html?


you do have IGBT based switches too

[Faringdon]

The point is that the short circuit current has to be limited by opening the FETs in time.

So you want to stop the current before it gets to your short circuit level?
Though your 'normal' current is 200A...so you cant have it trip on that.

...So these MOSFETs are going to have to break in excess of 220A......its very unlikely that MOSFETs exists that can break eg  220A of current without blowing up...unless they have a  large amount of auxiliary circuitry around them.  IGBTs on the other hand, can break currents  of that size.

As Tim allures, the alternative is to have multiple parallel current paths, from which the current of each is limited to well under 200A....but the total current into load is 200A.

[uer166]

200A isn't that crazy to do with paralleled FETs driven hard (no linear operation).
 
It can be dome with a comparator circuit with suitable trip and delay limit to avoid nuisance trips, but take care of FET SOA. You can even have multiple thresholds with multiple RC delays to simulate a fuse (slow and fast trips).

@farington it is likely better to over-size the FETs and let a short circuit current through, I'd guess in the 600-1000A range in this application.

Since load is inductive, you almost certainly need a freewheeling diode that can support the short circuit current for a time.

[Psi]

A system I used was a Isense amp IC measuring current across 6 layer PCB track resistor and providing ground ref output as input to MCUs built in comparator.
The mosfet bank was 160A 12V (bank of 4 fets in parallel controlled by opto gate drivers from MCU)
The comparator could have its output connected in hardware to a mcu output pin which I connected to a FET that immediately clamped the opto gate drive system turning the high current mosfet array off practically instantly. I had an RC filter on the Isense Amp signal to set the filtering/delay
The MCU input pin used for the isense/comparator could also be selected as input for the MCU ADC
The MCU comparator had a built in resistor ladder you could connect to the other comparator input to set a trip threshold.

All that meant I could
- Read and monitor the current using the ADC.
- Set the comparator threshold programmatically.
- Get an immediate shutdown all done in hardware if the comparator ever tripped and also get a comparator interrupt do take any software actions desired.
- Implement a second software-based over-current trip system at a lower current using the ADC as the first line of protection before the hardware/comparator system tripped.

[Faringdon]

Thanks, also the following shows a load switch...maybe higher current versions are available for your requirement

[luky315]

Yes, of course I would need a e-fuse or a "Protected Switch", the question is which controller shall be used to controll the (paralleled) MOSFETs. At the moment I'm evaluating LM5066 / LM5069 / TPS2490 based solution

[mag_therm]

If intending to limit a fault current by commutating paralled MOSFETS to OFF,
be sure to check the north east corner of the SOA limits for a hot device.

Also when Mosfets fail there is a tendency for the Drain to melt to the Gate,
which connects the Drain voltage to the driver components.