Breakdown on PCB versus air

[Klaus Kragelund]

Hi

I'm having a problem with a surge test causing a spark on a PCB I
testing. The problem is that the voltage from Line/Neutral to ground
exceeds the dielectric breakdown of the PCB leading to a spark

We cannot increase the distance due to size constraints, but we can
add a cut-out slot in the PCB to increase the creepage distance.

Normally about 2kV/mm is the breakdown voltage for air. And the
breakdown for a FR4 board lies at approx. V/mm

So adding a slot would help us. But ofcourse one needs to be aware of
polution/humidity of the air and polution of the PCB

I'm looking for numbers of the breakdownvoltage for both FR4 and air
versus polutions degrees and not just the standard numbers, but actual
calculations/tests

I've searched a lot on the web and found next to nothing on this
subject. I have ordered the www.ipc.org IPC-D-330, but will take days
to arrive.....

So does anyone have a clue about this?

Thank

Klaus