Gents,
I do have a mysterious behaviour in some PCB's and we don't find an idea how to tackle:
we do have two different designs (form factors), let's call them A (larger) and B (smaller); working in 434MHz transparent mode.
Testing is done via a proven sender, fed with 2kHz data signal, to a DUT working as transparent receiver, showing output data at a scope.
Or (for TX testing) vice versa: proven receiver and DUT as TX.
Both designs use exactly the same components (RF: TI's reference with Epson TCXO in a 4ML), exception is B has a bit smaller MCU but with similar code.
Both are TX/RX based on CC1125. We already built some prototypes and now we got the pre-series PCB's, parts have been placed by our professional supplier, placement looks fine.
"A" works fine, both TX and RX. No temperatur issues (air heating and ice spray cooling).
"B" works fine in TX, but all 12 samples show DATA-Out like receiving only noise. Measuring signal path doesn't show any obstacles, even at PIN 18/19/20 I can see the input signal as it should be (no shortage, no open).
Pressing on the PCB doesn't change anything.
But: some of the PCB's start working if I cool them down with ice spray. It's not possible to sweet spot an area, but it's always the arena of antenna network/CC/TCXO.
I single stepped and stopped the MCU, while CC works still in TX mode - same behaviour, so I can exclude the MCU.
I tested again some of the 'older' samples, which have been hand made and which worked well at room temperatur - they still work well, but if I heat them they also start giving 'noise' at the data output.
"A" again doesn't show any of this, so it can't be a parts problem (both are build with the same parts).
SW should also be out of scope, as this happens even if I stop the MCU.
It looks like a short/bridge, but even with microscopes we don't find it.
I know it's hard to remote help on these issues, but perhaps you guys have an idea how to ...
Thanks in advance