SF32LB525UC6 boot looping on custom board

Troubleshooting
1

Hello everyone! Sorry to bother you again about this custom board of mine, but I have yet another hardware problem I can’t figure out the origin of…

My problem is the following : the chip keeps boot looping every few milliseconds. I can tell because whenever I power it on the serial port is flooded with “SFBL” messages. If I flood the uart input with ENTER DEBUG commands, I can get the chip to output a “ENTER DEBUG MODE SUCCESSFUL” packet correctly sometimes, but it resets anyway shortly after.

The chip is currently connected to a NOR flash (powered from the VDD33_LDO1 output of the chip), but the behaviour doesn’t seem to change if I remove the flash chip. I have observed this behaviour across two boards I’ve manufactured, so it’s unlikely to be a manufacturing defect, though I managed to access and modify the FLASH memory content on the previous board. It doesn’t work anymore now, after a while the boot looping issue reappeared, and afterwards I compromised the device after too much board modification to test the individual components.

I will make further hardware probing to try and narrow the problem down, but I’d like not to fry my new prototype due to too much testing, so I would like to know beforehand if anyone has a guess about what could cause the chip to reset in a loop. I’d be happy to answer any technical questions you might have, and will provide further information as I perform more tests in the coming days. For now what I can say is, I have done the self-checklist available on the wiki, and I don’t think I have missed anything, except for the BUCK coil which is a bit off spec (it has a DCR of 1.6Ω instead of the 0.4Ω recommended, but unfortunately I couldn’t source the correct coil and footprint variant from my local supplier). Maybe it could be enough to cause instability in the internal power supply ?

2

VSYS vs VOUT33
VSYS vs VOUT33800×480 27.1 KB

Well I think I see one potential problem… On this graph VSYS is in blue, VDD33_VOUT1 is in yellow. It seems that as soon as VDD33 turns on, the system voltage drops considerably. That at least seems to be a good explanation as to why the chip keeps rebooting, such a voltage drop on the main power input would surely induce brown-out reset.

So the problem might very well be the out of spec coil after all… I don’t have a battery plugged in to the board and almost no peripherals (only a small NOR flash chip, and a switch to turn on the memory) so I don’t expect to hit any kind of current limit. I will search further for a solution, either I will try to obtain a good enough coil if I find any, or the problem comes from somewhere else. I still don’t get how that could have worked in the past but not anymore, very strange…

3

Well, after pondering on this graph for a bit, I reached the conclusion that the BUCK converter was almost certainly not the culprit, and I figured that perhaps the battery charger was not designed to run the chip without the added capacitance of the battery attached. So I tried connecting a battery to the device, and it worked perfectly. So I guess that concludes this problem, sorry to have bothered you about it, maybe the lesson here would be that it would be nice to indicate clearly that the internal battery charger cannot support the chip on its own (although it’s very possible I simply missed the information…)