Electronics > Power/Renewable Energy/EV's

Wrong use of Y capacitors on a Boost converter PCB?

(1/1)

Faringdon:
Hi,
The attached shows the bottom layer of a Boost converter, 5vin, 24vout, 3A75 out, 350kHz.
Customer has asked us to “make it work”. Customer is at “end of tether” with this PCB, and is even hard to get hold of to ask questions.
This PCB sits on a metal plate. (you can see the 4 screw holes). The  “Y” capacitors C25, C26, C17, C18 are 100nF 2220 1000V X7R capacitors. They are all connected from circuit ground to chassis ground.
This PCB has no common mode chokes on it.
Would you agree that the use of 100nF , 2220 capacitors makes no sense for common mode filtration? A 100nF, 2220 capacitor will have too much internal inductance (ESL) to be of good use in common mode filtration? Alternatively we should use around 1nF , 0603 capacitors for this purpose? (due to their lower ESL).
Also, would you agree, that without the use of common mode chokes, the actual use of Y capacitors is almost a waste of time here?…..they are better than nothing, but really need to be used with common mode chokes?

Also, these capacitors are of X7R dielectric. Would you agree, it is officially known that X7R is of no use for Y capacitor dielectric? X7R cannot “heal” after an overvoltage transient. Only Z5U or Y5U dielectrics can heal in this way?
Also, these “Y”  capacitors are rated 1000V. This is presumably due to an isolation requirement between chassis (metal base) and circuit nets. However, the clearance distance from circuit ground to chassis copper pour is just 0.25mm in places. (and that’s on the top layer). So do you agree, that 0.25mm would not meet any currently existent isolation safety requirement  for creepage and clearance distance?

tggzzz:

--- Quote from: Faringdon on December 03, 2022, 12:20:46 pm ---Hi,
The attached shows the bottom layer of a Boost converter, 5vin, 24vout, 3A75 out, 350kHz.
Customer has asked us to “make it work”. Customer is at “end of tether” with this PCB, and is even hard to get hold of to ask questions.
This PCB sits on a metal plate. (you can see the 4 screw holes). The  “Y” capacitors C25, C26, C17, C18 are 100nF 2220 1000V X7R capacitors. They are all connected from circuit ground to chassis ground.
This PCB has no common mode chokes on it.
Would you agree that the use of 100nF , 2220 capacitors makes no sense for common mode filtration? A 100nF, 2220 capacitor will have too much internal inductance (ESL) to be of good use in common mode filtration? Alternatively we should use around 1nF , 0603 capacitors for this purpose? (due to their lower ESL).
Also, would you agree, that without the use of common mode chokes, the actual use of Y capacitors is almost a waste of time here?…..they are better than nothing, but really need to be used with common mode chokes?

Also, these capacitors are of X7R dielectric. Would you agree, it is officially known that X7R is of no use for Y capacitor dielectric? X7R cannot “heal” after an overvoltage transient. Only Z5U or Y5U dielectrics can heal in this way?
Also, these “Y”  capacitors are rated 1000V. This is presumably due to an isolation requirement between chassis (metal base) and circuit nets. However, the clearance distance from circuit ground to chassis copper pour is just 0.25mm in places. (and that’s on the top layer). So do you agree, that 0.25mm would not meet any currently existent isolation safety requirement  for creepage and clearance distance?

--- End quote ---

If you can't deal with the customer ----- don't.

If you haven't the skill to sort out the board ----- don't try.

Don't ask random strangers to remotely do your work for you, based on inevitably incomplete information. You never know, some might adopt the old usenet technique used for students that tried to get people to do their homework for them: feed them subtly wrong answers so that their tutors could see they didn't understand

Navigation

[0] Message Index