48V PSU can put out 85V....Need output Schem for PSU?

[cupoftea]

Hi,
Why do OffThe Shelf PSU's not give the schem of their output?
Its clear from the attached LTspice simulation, that if an OffTheShelf PSU of 48Vout (say) gets a load suddenly connected to its output, then the PSU can put out 85V transient peaks....which may well blow up the load.

Clearly this is due to the LC in its output (always present in offtheshelf PSU's). So why dont manufacturers tell us what they've put in the output so we can investigate and anticipate such overvoltages?

 

[mtwieg]

I'm having trouble believing a designer would be dumb enough to choose those sort of component values (22uH and 1uF) for the filter, and with no damping. 22uF and 1uH would make much more sense.

And IMO it's not up to the manufacturer to assure you that everything will be fine when you hot-plug your load to their PSU.

 

[cupoftea]

Thanks Mtwieg....the attached shows much more rational components and still it goes up to 70V.

We're currently doing a prototype , where we have 12V wall warts providing power to 18V Abs Max gate drivers...........so we are needing to put voltage clamps in there.

AYK they might well use non optimal components...they may just re-use from somewhere else in the BOM......to save their cost.

I have actually known kit that got messed up by this.

One was a PSU for a LED driver.....in the product, the PSU soft started into the led driver and all was well....but during production assembly and test, they often turned the PSU on before connecting it to the led driver.....and when they then "hot" connected it........blew up the led driver...in many cases, didnt blow it up but weakened it and it blew up in the field.

--- Updated Jun 3, 2022 ---

...also, in our prototype, we cant keep unplugging and plugging the wallwarts........a switch is far more convenient...but it does then mean "Hot-connection"....

 

[scopeprobe]

Hi,
Why do OffThe Shelf PSU's not give the schem of their output?
Its clear from the attached LTspice simulation, that if an OffTheShelf PSU of 48Vout (say) gets a load suddenly connected to its output, then the PSU can put out 85V transient peaks....which may well blow up the load.

Clearly this is due to the LC in its output (always present in offtheshelf PSU's). So why dont manufacturers tell us what they've put in the output so we can investigate and anticipate such overvoltages?

The question is why would they? You purchase a PSU dependant on your requirements. Wall warts vary in quality so i wouldn't advise using them for a development such as yours anyway.
Your making a lot of assumptions regarding the accuracy of your Model. Your hot switching a pure capacitive load which is 3 times larger than the psu output filter capacitor which in itself would change the characteristic of the output filter in a simulation. You have no LR or ESR included which means you have a rather pure LC circuit with a high Q and no actual resistive load of note which would cause most SMPS to go into pulse skipping mode or something similar. You also seem to have damping for leakage that doesn't appear to exist in your model. You would probably be better served replacing the entire 'Off The Shelf PSU' block with a simple spice voltage source and start from there. As i mentioned in your previous thread, your probably better off designing power rails as part of the overall power supply, using wall warts is just adding more unknowns into an already difficult problem. If you need a clean source of 18V use a couple of 9V Batteries until you have the other rails stable

 

[cupoftea]

You have no LR or ESR included

Thanks, my apologies for not making it display in LTspice...but it is there.

--- Updated Jun 5, 2022 ---

our hot switching a pure capacitive load which is 3 times larger than the psu output filter capacitor which in itself would change the characteristic of the output filter in a simulation

..Yes i agree, thanks, and that is a great point.....i guess my values are possibly a bit unlikely in the RW....but its possible.
There is also the situation, AYK, that a downstream cap in an LC filter in an smps post filter is likely to be on the small side.....since the res freq of the post filter should be >3 x crossover frequency..(in academic theory at least)...however, this isnt a hard and fast thing, as often the phase margin detriment from disobeying that isnt so bad.

Thanks for suggesting alternatives,,......but i think dreaded wall warts are the cheapest way, and to be honest, i cant change and redo now as i must deliver something pronto.