Anomalous modulating power pickup

[DOCR]

During some long-term datalogging for unrelated tests, I've observed some occasional anomalous electrical activity at one of the circuit nodes being monitored and I wondered if anyone else has seen anything similar

The logged trace is usually fairly smooth, with just a very slight amount of noise, but approximately once every day or two the voltage starts changing quite significantly (eg. up to 0.5V pk to pk on a 3V background level)

The trace 'looks' like a fast transient noise glitch, but in fact the anomaly often lasts up to around 15 minutes, with each single peak or trough lasting about a minute

Each anomaly shape is unique but some have a few common features, eg. a marked central spike (or trough), flanked on each side by slightly symmetrical troughs (or spikes)

I've tried swapping the 2 logging probes between the 2 nodes being monitored and the anomaly stays with a particular node (which suggests it's happening within the circuit, not the logger)

The node of interest is an output terminal of a ferrite transformer, so I'm assuming that the anomaly is magnetic rather than electrostatic - however, the transformer core is small (only 1-2cc in volume) with just a few turns per winding, so only capable of very small capture of stray fields (eg. no pickup from 3kW charging equipment being switched on/off, on other side of wall using same utility circuit)

The affected circuit branch is small (see below) and very low impedance, so whatever is able to induce up to 15% variation of the DC level must be a significant source of energy, lasting many minutes

If you've seen something similar occur in one/some of your circuits please post a few details

thanks

[DOCR]

It seems that this power pickup isn't due to magnetic or inductive linkage with some source - I placed a duplicate circuit next to the original one and only one of them gets affected at a time

So, maybe the effect is happening within the silicon of the pulse switching transistor?

Here's another (differently-shaped) example of the anomalous influence on the output line - a symmetrical 'sweep' of positive and and negative DC peaks/trough, again lasting about 15 mins in duration

[Jeg]

Hi
I wonder if it is some kind of ferroresonance. At specific range of frequencies with specific duty cycle lower than 50%, ferrite cores are getting excited due to that effect, producing erratic behaviour which is picked up by the secondary. Usually you can hear it like hissing coming out of the core.

[DOCR]

Hi
I wonder if it is some kind of ferroresonance

Hi Jeg, good call !

The drive current is probably too low to cause 'squealing', but if it is ferroresonance that could explain the generally symmetric nature of the effect

The oscillation frequency is voltage-dependent on the supply, so as the battery voltage gradually decreases with time, the transformer can definitely get tuned through possibly-resonant frequencies

This would also account for the duration of the effect, lasting many minutes

Yes, it all seems to fit, I think you've nailed it!

Many thanks

[Jeg]

Hi again.
A member from here posted those videos before few years. It seems that he took the advantage of this effect to make a self sustain circuit. In any case it looks interesting.

https://www.youtube.com/watch?v=Z2vrfyqmYvU
https://www.youtube.com/watch?v=vgYLVyswgeQ