Is a PWMed load a bad idea when using a switched-mode power supply?

[Pinkus]

I need some help:
A friend has a Meanwell switch-moded power supply (type is actually unknown to me). It is a Meanwell 24V 250W.
Connected is a H-Bridge driving a 24V brushed motor with 13A current at full load (though starting current will be probably much higher).

Problem: The H-Bridge is ramping the motor up to speed by PWM (probably some khz, though I do not know yet). This is working fine with no load, with load two of the Mosfets (40V, 18A / 150A pulse) are burning through (=shortage) immediately. Interestingly these were two Mosfests not driven simultaniously (=of the two Mosfets switchend on to drive the motor, only one Mosfet did burn through).
Btw: the motor output of the bridge / mosfets are protected by a bidirectional TVS diode against inductive peaks from the motor. But there is no such protection (yet) at the supply input from the power supply.

The friend is abroad, thus I currently cannot go there and measure what is happening.

My guess is: the switch mode power supply does not like the PWM load and then outputs peaks of a too high voltage which is killing the Mosfets. However I never had experience with a pulsed load at a switch mode power supply and are seeking some advise if the above would be the most logical reason. Maybe somebody has some experience with these supplies.
Any other ideas?
[Ignore the H-bridge, this is proven to work fine and is well designed.]

[T3sl4co1l]

Well, easily verified with a scope.

I don't know that any random supply should behave badly under this condition.  It's certainly possible.  I would not buy one that does...

It might not be the supply.  If there is substantial length of wiring between supply and bridge, the stray inductance can kick when load is switched.  A properly designed bridge will account for this (typically with a lot of electrolytic capacitors for bypass).

That the H-bridge is capable of burning itself up, is proof to me that it is not "work[ing] fine and is well designed"!

Tim

[TheMG]

Provided the output filter capacitors in the power supply are in good working order and the PWM frequency isn't super low, the power supply doesn't even "see" the PWM at all, it just sees a relatively steady load as the PWM current draw is filtered/smoothed by the output capacitance.

That being said, do you have power supply decoupling capacitors of sufficient size located as close to the H-bridge itself? If not, as the previous poster mentioned you might be getting effects of the wiring inductance between power supply and H-bridge.

[David Hess]

I second T3sl4co1l's suggestion about stray wiring inductance.  No power supply should have a problem with a PWMed load, but PWM loads can be very susceptible to stray wiring inductance.

[Pinkus]

Thanks to all.
As I said: I have not seen the cabling situation nor can I go there and measure with my scope. I already told him to use a much larger capacitor at the h-bridge (is currently only 220µF).
I once had a switched-mode bench power supply on my bench (replaced by linear ones long ago) and I remembered that this supply did not like PWMed loads at all and went crazy then. This is why I thought this could be a common problem with these kind of supplies.

[2N3055]

What does "switch-moded power supply" mean. A switch mode supply or that it was moded, aka, something was changed with it.?

Like Tim said, a scope is a first thing to do...

[Pinkus]

I know that not being able to take an oscilloscope measurement is not satisfactory. Unfortunately, he does not have one, I am too far away. The supply is a regular Meanwell RSP-500-24 https://www.mouser.de/datasheet/2/260/RSP-500-spec-1109676.pdf
As said: I will propose him the connection of a much larger capacitor at the bridge.