# Insulation resistance tester: orders of magnitude difference in measurement depending on connection polarity

https://youtu.be/eGjd7HoYyz8?t=115

The man on the video is measuring a transformer's insulation resistance, the resistance measurements are very different depending on how the meter is connected:

1. When the positive lead of the meter is connected to the chassis of the transformer and the negative lead is connected to a low side winding the resistance measures as 30MOhm

2. When the positive lead of the meter is connected to the low side winding and the negative lead is connected to the chassis the resistance measures as >1GOhm

The question is what is causing such a behavior of the tester.

• That's a roundabout way of saying measuring one way gives 30 MΩ, switching the leads around give you >1 GΩ, and the question is why the resistance is different when they should be same. – Harry Svensson Mar 11 '18 at 22:00
• Yes, sorry if it's not clear and could be formulated in a shorter way. – axk Mar 11 '18 at 22:08

It could be geometry (pointy things spray electrons when made negative), or the copper wire could have enough of a coating of copper oxide (a semiconductor) to be rectifying the leakage current through its insulation.

A typical gas-filled rectifying tube used a filament, but some were simply points opposite plates (similar to the field-emission cathodes that are still being experimented with), and still rectified. This would be consistent with the given results if the steel of the transformer had a rough surface compared to the copper wire, and was field-emitting at the high points.

Copper/copper oxide is a Schottky diode structure, and copper oxide was used for rectifiers in years past. The leakage of practical CuO rectifiers is relatively high in the reverse direction, so it took stacks of junctions in series, but the asymmetry in conduction was maybe 100:1 between forward and reverse bias.

A good strong screw or solder connection bypasses any oxide, but the varnished wire in a transformer hasn't any oxide-destroying properties.

The tester measures a current of around micro-Amps (1 GOhm at 50 v to 2.5 kV). It is difficult to have accurate measurements with this kind of clamps. There could be a lot of things.

The terminals are likely made of different metals, steel for enclosure, copper/whatever for terminals, different metal plating, etc. There should be some termocouple effect, micro-amps one way versus another, temperature of clamps will make a difference, dirty hands will make a difference, nearby AC can be rectified by dissimilar metal junctions, etc, etc.

More, the transformer is likely filled with transformer oil, and it might have substantial capacitance with unknown charge level. And the charge level will change depending on how long the guy takes readings.

The whole tester looks like a child's game controller, not like any professional instrument. The whole thing is goofy.