# How to measure leakage current?

I have an AC powered device, and a customer asks for exact leakage current. Current that can trigger a leakage current circuit breaker, so it's obviously something flowing through Y capacitors to the ground wire.

The question is, what is usually measured and what actually matters? Peak to peak? RMS? Pulse width?

• Obviously you can just measure whats running through PE but what matters depends on the kind of breaker, there are quite different ones that measure in quite different ways – PlasmaHH Sep 6 '18 at 9:20
• This is why i asked about "usually". No problem to measure. Problem is that the customer asks for "below 3.5mA", but they don't know is it RMS, peak, or whatever. – Gregory Kornblum Sep 6 '18 at 9:22
• And of course personally I am not aware of any specific breaker, which is the problem in the first place. – Gregory Kornblum Sep 6 '18 at 9:23

The required insulation test uses high voltage DC where one clamp is put on both the AC inputs and the other probe is put where people can touch (earth wire, metal chassis, low voltage DC components, ...). Because it is DC class-Y capacitors will not be leaking current.

The simplest test to test for AC leakage is to put it on a non-GFCI protected circuit (to avoid trips) and put a clamp meter around live and neutral (but not ground) and set to measure AC. You can frankenstein an extension cord for that purpose. This is how a GFCI works, when that current exceeds the rated value it will trip.

There are more advanced devices that will used high voltage AC instead which will include the effect of the class-Y capacitor.

If that leakage current if because of high frequency noise on the line you can add a common mode choke on the AC input as filter.

• It needs to be a high-sensitivity clamp meter, capable of measuring down to milliamps. Not all can do that. – Simon B Sep 6 '18 at 14:11

How to measure leakage current? I have an AC powered device, and a customer asks for exact leakage current.

Of your device? I believe you are referring to current to ground that should be on the neutral. If you want to find how much current is leaking to ground in a device, electrically isolate it and measure current to ground through a known path.

Current that can trigger a leakage current circuit breaker, so it's obviously something flowing through Y capacitors to the ground wire.

Hmm in retrospect there may have been something wrong with that first sentence and maybe it was supposed to be part of this one. Anyway I think you're talking about an AC GFCI circuit breaker. Most GFCI devices work by passing both live and neutral wire through an induction loop. If all of the current flowing out on the line wire flows back on the neutral, the magnetic fields will cancel out and no current flows in the induction loop. If any of the current flows to ground instead, it won't return through the loop on the neutral and a portion of the magnetic field won't cancel out, causing a current to flow in the sense loop and breaking the circuit. This article goes into detail about what's legally defined as a class A, C, D, E device and why trip level requirements are what they are.

The question is, what is usually measured and what actually matters? Peak to peak? RMS? Pulse width?

The circuit I'm aware of mentioned above is basically measuring the absolute value of the total lost current, meaning both that it is measuring quantity and that it measures current regardless of direction of flow. The legal requirements for NEC specifications are clearly based on both current and exposure time, but because faster is always better in this case it matters more that it measures up to the minimums set by the government and performance may be significantly better in reality.

At any rate, Class A GFCIs must trip at 5mA and class B GFCIs at 20 mA, class B GFCIs are for old underwater fixtures in swimming pools where leakage current exceeds 5 mA and can cause nuisance tripping.

• Slightly off topic: I love the way you structure your answers. – Harry Svensson Sep 6 '18 at 13:24
• @HarrySvensson Very kind of you to say. A very nice thing for me to hear from an actual engineer. I really hope my helping with some of the easier and electrician/safety questions on here helps free up you fancy folk for the tough ones. – K H Sep 6 '18 at 23:23
• So, what matters is peak current to either direction? – Gregory Kornblum Sep 7 '18 at 11:16
• @GregoryKornblum I think first can we confirm that this is the type of device you are asking about? Also are you referring to AC or DC equipment? – K H Sep 7 '18 at 22:58
• It's AC servo drive – Gregory Kornblum Sep 7 '18 at 22:59

If someone is asking for an AC current, and they don't specify how it is measured, then it will be RMS.

RCDs (GFCIs) are set to trip at a specified current imbalance between phase and neutral.

For an AC powered device, finding leakage value in mA can be performed using a flash tester of some kind. You can use a 500V or 1400V (depending on TVSD) AC HIPOT test to highlight if there are any issues with flash over, leakage and clearance between earth and live/neutral. In the event of having a high leakage I usually look for damaged/nicked wiring, poor clearance and take a look at any filters that may present at mains input.

It's good practice to perform an IR test after a flash test, in case any harm has been done.

• By 'flash tester', do you mean a megaohm insulation tester? – MrGerber Sep 6 '18 at 13:07
• I mean like a multipurpose safety tester, like the H103 model as an example. Not an insulation tester. – TestDeviant Sep 6 '18 at 14:06
• The H103 model of what? If I google "multipurpose safety tester" I don't get any results. Are you talking about a dielectric strength tester? Answers are more usable for others if they use correct and universal language, in favor of "local lingo" or at least explain niche abbreviations and terminology. – MrGerber Sep 6 '18 at 14:21
• Google multipurpose safety tester h103 – TestDeviant Sep 6 '18 at 14:29
• Is OP supposed to acquire a "multipurpose safety tester h103" for his small (one-time?) leakage measurement? – Harry Svensson Sep 6 '18 at 14:59