And do you ever test the tester? Nobody hears the dog that doesn't bark.
While your advice was correct, you appear to be answering some other post.
It turns out my prototype outlet tester is tripping the GFCI.
I'm not sure exactly what is going on but I also smell the distinctive odor of burning resistor and the proto board is getting hot... This particular proto has a bunch of extra parts patched in to test the 100M input Z, and may have too small of an effective resistor in the ground leg. The board is too small and parts are too close together to tell exactly which part is causing the excess heat dissipation.
Without the outlet tester plugged in I still measure 2 mA leaking into the ground (with kitchen appliances plugged in.). Oddly it remains unchanged at 2 mA with the mixer plugged in or not, so the mixer is NOT the leakage path.
Found it...
the 2 mA leakage was being caused by a cheap outlet strip that was "protected", so probably a leaky clamp device across the line to ground (?). Seems like a clamp should be line to neutral but I'll take it apart later.
=====
I have other work to do but now I can drop the GFCI outlet back into the junction box and ground it.
My next generation outlet tester PCBs are due in a few days from now... currently at the DHL sorting center in Shenzhen.
So logic has returned to this exercise and the laws of physics and house wiring are sorted. I've tested the outlet testers with GFCI before so this appears to be a fault specific to this one prototype tester. I sent away all my other testers, but never smelled the burning resistor smell before. And the field testers have not reported issues with GFCI. The design is specifically running ground current at less than 5 mA.
In fact the UL spec for outlet testers that I am designing to defines the max ground current at 2 mA, so unlikely to trip a 5 mA GFCI.
JR