LDO linear voltage regulator for ESP8266

[n0f8r]

Hi All,

I have a board with an ESP8266 module on it - this one: https://www.aliexpress.com/item/32311028307.html?spm=a2g0s.9042311.0.0.27424c4dtzgihN.
The board takes 24v and uses a switching regulator to drop that down to 4.5V which is then fed into a 3.3V linear regulator - this one: https://www.mouser.co.uk/ProductDetail/621-AZ1117EH-3.3TRG1?r=621-AZ1117EH-3.3TRG1 which then feeds into the ESP module with a pair of 22uF mlcc caps across the LDO regulator output and the power in of the ESP module.

My question is on LDO Linear Reg output and whether it's normal to see 3.3v with +- 0.25V variation, ie spikes down to 3V and almost up to 3.6V?

I have a high board failure rate - where sometimes the ESPs just seem to go haywire (stop responding, slow to respond, random reboots) something like 10% of them in fact. Initially I thought maybe I'm getting bad quality ESP modules form China - and I'll just have to live with it. I tried de-soldering the ESP modules on a few bad boards and re-soldering fresh ESP modules on, they seemed to work ok for a bit, but then two of them started acting up again.

So now I'm wondering if it could be the linear regulator I'm using, or how I'm using it. I'd guess the 22uF power filter caps I'm using are probably too big for purpose, maybe way too big, but that's only a BOM expense issue right - it wouldn't cause voltage spikes?  Google tells me 3.6v and higher can kill the ESP8266 - I put one of the re-soldered boards on the scope and I'm seeing 3.3V plus/minus 0.25v on the ground and VCC of the ESP module - to be clear, at 100 micro seconds per division the voltage graph line looks very stable to the human eye, but the scope stats tell me there are peaks over 3.5V and lows almost down to 3.0V. Is that normal to expect coming out of a 3.3v LDO regulator?

any tips/pointers/questions appreciated.

thanks,
 Kevin.