Making an offline PSU more immune to mains transients

Hello,

In line with pages 7 to 8 of the following, can we increase mains transient immunity of an offline PSU by reducing the Y capacitance?

View attachment TN-203-Ensuring-Surge-Compliance-Rev-0.2.pdf

Having understood this, the obvious temptation is to entirely dispense with all Y-caps on the PoE board.

Click to expand...

I don't see how you could possibly apply that sentence to PSUs in general. The app note is focused purely on protecting PSEs (which apparently don't have any Y caps in themselves) from surge currents. Certainly reducing the Y caps in the PSU will reduce the surge currents through the PSE, but this doesn't address the effects on removing the Y caps on the PSU itself, or any other part of the system.

can we increase mains transient immunity of an offline PSU by reducing the Y capacitance?

Click to expand...

Basically - NO.

Filter and MOV's and/or GDT's, ideally MOV's with fuses built in

Thanks, I wonder then, by increasing the voltage rating of Y capacitors can you increase mains transient immunity?…..…I say this because some years ago I was working for a multinational lighting company, and we managed to reverse engineer a competitor product and got the schematic for it…..it was a 15W , 230VAC , Offline Flyback LED driver. There was the usual Y capacitor across the isolation barrier from primary to secondary…but in series with the Y capacitor was a high voltage diode. This surely must have been put there to increase mains transient immunity?…..but it makes you think, hey, hang on, having no Y capacitor at all would surely make it have even better mains transient immunity.

but in series with the Y capacitor was a high voltage diode

Click to expand...

Makes no sense at all.

The purpose of Y capacitor across transformer is to short interference currents injected by the switcher between primary and secondary winding. Any series component must be expected to thwart the wanted effect.

The original thread point was possible conversion of common mode into differential mode surge by inappropriate (e.g. excessive) Y capacitors. I agree with the previous answers that you shouldn't overate the effect and that state-of-the-art surge protection is still required. But whether the discussed effect exists or not, it's not specifically related to Y capacitor rating. Y capacitors have high surge capability and overvoltage margin by design.

I feel that you are somehow poking around in the dark about the frequently discussed surge capability and possible surge induced failure of your LED lamps.

The purpose of Y capacitor across transformer is to short interference currents injected by the switcher between primary and secondary winding.

Click to expand...

.
Thankyou, yes i agree the Y capacitor diverts these "noise" currents back to the primary side. If the Y capacitor was not there, then this “noise” would jump across to the secondary, then somehow couple to earth from the secondary and go that way to complete a current loop and create common mode noise.
The thing is, isn’t it strange in situations when the isolated offline SMPS has earth connected to the primary side… In such cases, they often then use an insulated wire to connect primary side earth to the secondary side ground so as to earth that aswell. …-In such cases, you strangely often also see a Y capacitor connected from primary side to secondary side . –Surely there is no point in that Y capacitor to be connected like that in this case?
..I mean, surely t it would be better to let the noise get back to the primary side via the insulated wire from secondary side earth back to primary side earth, and then use a Y capacitor from primary side earth back to primary side live or neutral. –I mean, that, surely, must be a much more effective way to filter out common mode noise?

Actually a large Y cap from HVDC 0v to output 0v can carry enough current under mains surge conditions to upset secondary side control and cause a blip or worse in converter performance ...

Considering Y capacitors connected from Earth to Live and Neutral for the purpose of Common Mode Noise reduction…surely such Y capacitance makes the product less immune to damage from Line to Earth transients?… (eg from lightning strikes)
I mean, a big enough transient will break over these Y capacitors and short them…….thus destroying the power supply....surely you agree with this?

perhaps you have never read a spec for Y2 class caps? they are rated to about 5kV and designed such that they should never break down to to a short ( as far as that is possible) which would be bad for safety issues - hence their large size relative to other caps.

Y1 to about 8kV and ditto ...