creeping voltage

Here you can take an Idea.

Dielectric strength of normal dry air is about 3200Volts/mm, so 3000Volts should not jump on 4mm gap anyway.

Not really sure what you are trying to do, but dielectric breakdown voltage of air is about 3kV/mm. That value is however affected by multiple different factors that (might) reduce it.

http://en.wikipedia.org/wiki/Dielectric_breakdown

I believe both of you are a decimal point off.

The 60 Hz dielectric strength of dry air at sea level and a temperature of 25°C is 10 kV rms per inch or 393.7 volts rms per mm (556.8 peak or crest).

The originator did not specify whether the voltage is AC or DC but, if it is DC, the corresponding breakdown voltage gradient would correspond to the AC peak or crest voltage (rms x sq rt 2).

When you are bringing the breakdown of air into the discussion, I believe you are talking about the strike voltage. The creep voltage is usually higher, since the spark has to "creep" along surfaces. If the surface of the structure has material on it, or is otherwise not clean, this strike voltage can be much lower.

-Bluezone

I'm guessing your running a "Hi-Pot" test of your fixture?

Have you tested the fixture in the same conditions, but without the Aluminium (I guess bulk-head) in the circuit, or completely removed?

from my old BS159 standards minimum creepage distance for 3.3kv is 50.8mm

A rule of thumb that has worked in industrial application for many years is 1KV per inch of creapage distance. This takes into account some dirth and humidity but not rain... This is what we use in our electrical cabinets.

We use some Phoenix connectors. If they are the same type that we use, you should not use them at more than 250VAC. Not enough distance between metal parts.

It is very difficult to find a connector that will hold 3KV under industrial conditions. Most will need rubber boots to keep them clean and dry and even silicon grease to provide you years of reliability if you intend to cram everything within mm of spaces.

There is a big difference between an air gap and a surface creapage. Dearth doesn't (usually) accumulate in a gap but it does on a surface. Also, coronas can travel on a surface from ion accumulation and static charges. This is why the V/m is much higher in a gap than on a surface.

Often adding a layer of conformal coating will break the surface creapage by providing a thin insulation layer for the conductors. Is is recognized by UL and CSA that the surface creapage distance is cut in half when a proper coating is added to printed circuit boards and similar.

In my last company, we used to use "Selly's Estapol Clear Varnish" on our HV circuit boards for Electrostatic air cleaners to limit the amount of corona and flash over, worked well after 3-5 coats.

I wonder if you can use standard SHV BNC connectors, those are rated at 3500VAC/5000VDC typical. Diameter of the mounting hole is about 11mm or so for the panel types.

Yes, air is 3kV/mm if dry, real air though is more like 1.5kV/mm. I do a fair amount of prototyping with HV and find a good rule of thumb is absolute minimum 1kV/mm (DC) gap for bare conductors on surface of insulating substrate (e.g. PCB). In practice humidity and dirt will influence that, so 0.6kV/mm is a more conservative value (e.g. 5mm gaps for 3kV). If you design the gap according to PCB standards the figure is considerably lower - something like 0.1kV/mm.

Can anybody explain

What is creeping voltage?