What is a correct method to measure safe maximum voltage of a PCB?

I have a small (24 x 15 mm) PCB for a custom connector. It is intended for low voltage (< 48 VDC) applications, but I am interested in measuring it in order to provide customers a reasonable maximum voltage rating (de-rated of course).

It is a 1.6 mm thick two-sided FR4 board with 1 oz copper and soldermask. The smallest gap between traces is 1 mm. Using this calculator from Sierra Circuits, the maximum voltage should be about 528 V. They define this calculation as:

...the minimum spacing between the two copper features that can handle that voltage difference without any dielectric breakdown.

Using an insulation tester that produces a 1000 V test voltage, the insulation resistance is measured as ≥2 GΩ (at 25°C); which is the tester's maximum.

I suspect the calculator is erring on the side of safety. I also assume it is using the IPC-2221B standard like Saturn PCB Toolkit, which effectively only goes to 501 V or 0.8 mm trace gap.

Lots of questions are based on faulty assumptions, and I am sure this is no exception. I'd like to know:

• What is a better or more proper technique to measure various maximum voltages of a given PCB? (For example, I'd like to know its insulation resistance and at what voltage it starts to arc. I know that temperature affects the result.)

• Is it common/accepted to assign a conservative (e.g. 300 V) rating even though I don't actually know its limit?

• I would like to measure its insulation resistance value. With a basic (Fluke 1587 FC) insulation tester at its maximum, I'd be OK accepting a value of ">2 GΩ," however, what would the next steps be if I wanted to obtain a more precise value?

It really depends on what standard you need to adhere to. This involves whatever NRTL or ETL you'll take to get the product tested (if you do so). If you are going the NRTL route, you might be able to get some info from them on what req's and standards you need to apply for certification. At the end of the day, you want to be able to make a design safe and not have to rev it.

I can give an example of some specs (but do your own homework because standards get updates): These are the the IPC-2221 standards, and most designs can use these clearances.

This doesn't apply if you have a product that will be operated in an commercial or industrial environment, then pollution degree can apply: Your standard will have a table like this:
Source for table and quote

How to define working voltage class net to calculte the creepage?

What is a better or more proper technique to measure various maximum voltages of a given PCB?

Use the calculator and specs at this site for temperature rise

Is it common/accepted to assign a conservative (e.g. 300 V) rating even though I don't actually know its limit?

If that's the nominal voltage then probably yes. ESD or other spikes need to be accounted for, man designs with AC mains uses filters to suppress short term spikes so they don't affect the design.

I would like to measure its insulation resistance value. With a basic (Fluke 1587 FC) insulation tester at its maximum, I'd be OK accepting a value of ">2 GΩ," however, what would the next steps be if I wanted to obtain a more precise value?

Use a highpot tester.

IPC standards for minimum electrical clearance provide for voltages above 500VDC (or peak AC) at 0.00305mm/V past the 0.8mm/500V range of table 6.1 of IPC-A-610. Your board's 1mm spacing with (edit:) polymer coating/soldermask would therefore be in compliance up to ~565V. The calculation is probably fairly conservative but aside from sending it off to get hipot tested, I think "rolling your own" test setup to get a number higher than this could expose you to unacceptable liability without much potential (no pun intended) gain, especially since the intent is for ELV and since soldermask thickness tolerances, etc could affect the result.

• I definitely agree on the variability being a primary reason to select a conservative rating. Just a note that this board is soldermask only (no conformal). In short, I am reading that if we needed/wanted to get a certification, the solution would be a professional hipot testing provider. Commented Oct 20, 2022 at 17:54
• Those numbers are for "permanent polymer coating", e.g. soldermask. There is a separate set of values for conformal coating. I would definitely go with third party testing on this to have an independent data point in case it ever becomes a liability issue. Also take into account the possibility that the soldermask might be damaged by the customer if the board isn't well protected.
– vir
Commented Oct 20, 2022 at 17:59