Has a regression snuck in to the PPK stack?

We've had good success measuring power consumption with the PPK2 until now. But recently, it started reporting some hard-to-explain numbers. Specifically, t told me a nRF9160 was drawing 40 uA, but I had an UART RX interrupt active, which should mean the HFCLK was online. Furthermore, when I toggled the measurement (without turning off the supply or doing ANYTHING to the DUT itself), the numbers immediately jumped back up to what I would expect for an active HFCLK (500-600 uA or whatever). So something was definitely up.

Suspicions raised, I plugged in a Joulescope in series with the PPK and the PPK was consistently off (reporting way lower values).

Ordinarily, I would suspect some hardware defect in my PPK but a customer has seen recent numbers when doing their own measurements (same thing there - HFCLK active, PPK reporting numbers way too low for what I would expect to be possible with HFCLK active).

Does anyone have comments here? We love working with the PPK because the form factor, UX and price are all great so ideally we can keep using it without worrying about reliability...

Related