Just guessing, but we're probably staying below the limits of the resolution of the ADC.
For sure
why not go for the "voltage" method of https://lygte-info.dk/project/Charger%20UK.html? Charge at CC, and periodically measure the voltage under 0 charge current.
That charger doesn't have relay, and have a protection diode.
I don't think that it is a good idea just to stop PWM without disconnect the battery...
no, Then it would show 0dv/dt and not -dv/dt.
Also when the Voltage threshold is set to 0.90V it should not have triggered at all. To me it looks more like a Bug.
What RD60xx model?
Need to see the graphic, to understand what happens.
PSU detects the peak of output voltage, and "-dV/dT" settings sets the Voltage threshold from the peak.
I am also tested but only on 6018.
When I set -dV/dT to 0.03V, it triggered too early, after about 54 minutes, it looks like the output is too noisy.
Then I set -dV/dT to 0.05V, it triggered after 02h 33m with 0dV/dT state, analyzing graph can see the same problem, output noise.
I will try to add some filtering.