[SOLVED] The circuit works with a common LDO, but does not work with separate LDOs.

[GitRaccon]

Hi.

I have simple PCB, with STM32 and RFM69. STM32 is suppiled by 2,5V LDO (TPS7A0525, permamently enabled), and RFM69 is suppiled by another TPS7A0525, which is enabled by STM32 GPIO pin (TPS has EN pin, output voltage its OK). When i enable LDO for RFM69 and im trying to transmit frame by RFM69 radio, it doesn't transmit, SPI communication work well (readed registers values, waveform on the oscilloscope). But when i disable this LDO and short RFM69 supply pin with STM32 supply pin (both ICs suppiled by common LDO) RFM69 properly transmitt radio frames. All ICs are placed relatively close, in 1,5x2" area. Its not my first time when i use multiple LDOs to supply multiple IC, but this situation suprises me. Did You have meabe some ideas?

Best regards

 

[KlausST]

Schematic?
Or at least a sketch?
Links to datasheets?

 

[GitRaccon]

There is nothing special in schematic (below). I have tried various combinations of connecting the RFM69 to the STM32 powering LDO: before, after the ferrite bead, etc., and it works in each combination, but when I use a separate LDO for the RFM69, the RFM69 does not transmit radio data.
LDO datasheet: https://www.ti.com/lit/ds/symlink/tps7a05.pdf
RFM69 datasheet: https://www.tme.eu/Document/8ad31f80918318ab98bba02e2011e54c/RFM69HW.pdf

 

[betwixt]

The schematic looks fine. Are you perhaps enabling the RFM69 but not allowing it time to initialize before 'talking' to it?
An oscilloscope should be able to check the supply rise and data timings.

Brian.

 

[GitRaccon]

betwixt - bingo - that was software issue. I was working on a longer developed software and I didn't think there could be a problem of this type there, but in fact, I didn't have RFM69 initialization module after turning on its LDO. Thank You

 

[dick_freebird]

The two have different loads and different ref grounds physically so likely electrically. So different responses to load actors and any drool or overshoot might make one chip's supply push in, or ground, to make substrate current (leasing who knows where?) or simple logic glitch as seen by pin w.r.t. either.

First step? Bug hunt for such spurii.
Next step? More / better close-in decoupling.
But not -too- good because ESR min for
stability.

 

[KlausST]

Hi,

when powering down the RFM module .. you also MUST NOT drive the digital inputs HIGH.
But usually SPI signals are HIGH when idle.

When powering down the RFM module .. no communication via the according SPI signals is possible / allowed.


We can not find out if you do this correctly ... so you have to do it on your own.

When I design such circuits with partly powered down sections, I ususally use hardware to be sure digital signals are within allowed levels.
This ensure that during powering up the microcontroller and/or during programming the microcontroller .. or other non_normal situations. ..nothing bad can happen.
I miss such hardware in your schematic.

Klaus