I have developed an energy metering module based on ADE7763 with an ATMega8 MCU. We have never had any Problems while prototyping the boards. Now, we started production of 100 boards. But: We have 30% boards with a same bug: The ADE Chips returns nothing on SPI. The SPI bus is isolated by optcouplers.
I have attached two Images to demonstrate the bug. The first one is a signal plot of a working board.
0: CS measured at MCU
1: CS measured on ADE
2: MOSI measured on MCU
3: MOSI measured on ADE
4: MISO measured on MCU
5: MISO measured on ADE (but after driver Transistor)
6: SCKL measured on MCU
7: SCKL measured on ADE
B2: decoded SPI Bytes
Everything is working perfectly.
The second image is a Signal plot of a non-working board.
The ADE chip Returns 0x000000.
But i think the ADE chip understands the first reading command from MCU, because the MISO line is set to low after that. Any ideas?
We have done following without any effect:
1) Replaced ADE chip to a new one
2) Replaced all optocouplers
3) Replaced crystal
4) Checked +5V and VREF (is 2,4V)
5) Checked analog inputs: same on each board
6) Replaced MCU to a new one
Thanks for Input!