Bug #3248
closed
osmo-e1-xcvr fails to talk SPI
Added by laforge almost 6 years ago.
Updated almost 6 years ago.
Description
At least Dieters osmo-e1-xcvr seems to have trouble to work at all. He's not able to solicit any SPI response at all from the chip. Voltage is stable, clock looks great, his SPI waveform looks vaild, but the SDO line stays low no matter what he does on SPI.
It's currently unclear what might be causing the issue, but it's likely that the other osmo-e1-xcvr boards are also affected. Please keep this in mind when playing with your boards :/
I've re-checked the pinout of the data sheet against the schematics/symbol and the pin-out is right, and I couldn't find any other obvious error in the schematics/board design so far.
- Related to Bug #3247: osmo-e1-xcvr should put pin 44 to GND added
- % Done changed from 0 to 20
it seems at least on Dieters' board the GND end of the 3kOhms R1 was somehow not connected to GND. If that connection is manually created, the board consumes >= 50mA more current and the red "LOS" LED shows up at start.
Dieter states more information will follow tomorrow.
roh: please have a look at the unsoldered PCBs if the R1 pad has conncetion to GND, or - if it hasn't - why. At least according to the eagle BRD file in git it should have.
- Related to Bug #3251: osmo-e1-xcvr: R1 doesn't connect to GND? added
- Status changed from New to Closed
- % Done changed from 20 to 100
After fixing the GND connection of R1, Dieter has made some success:
- red LOS LED is on at startup (expected)
- SPI registers can be read/written using (STM32) SPI_CPOL_Low und SPI_CPHA_2Edge
- for E1 operation, it's sufficient to set THZ (Transmitter High-Z) in register 0x06 (TCF1) to 0
- if Tx is looped back to Rx, then the LOS led is gone
I've created #3251 to investigate the R1-missing-GND connection which appears to be the actual bug here.
Also available in: Atom
PDF