Issue with ENS160+AHT20 board - AHT noisy data

I’ve been having power issues on my ENS160+AHT20 boards, fixed with an extra cap as per kbenton’s post, but this issue highlighted another problem I wanted to ask about without hijacking that thread:

In the screenshot you can see the point at which the ENS160 stopped sending readings due to the power issue (logs reported: ens160 readings unavailable - normal operation but readings not ready).

Notice how the AHT20 data is very noisy, but only when the ENS160 is working?

Why would this be? The noise remains now I’ve added the extra capacitor, but why did I get more stable data when the ENS160 failed?

There is also a ~2°C positive offset associated with this behaviour that I had originally attributed to local heating from the other side of the board, but now I’m not sure.

I will post more as I look into this further, but wondered if anyone more experienced with these boards had any wisdom they could share.

I don’t know but my guess would be that it’s still power supply related.

How are you powering it? 3.3V or 5V?