We need help with a problem with the L9963T (SPI to isolated SPI transceiver).
Please help us find the cause and solution.
PROBLEM DESCRIPTION:
- We are using L9963Ts on each side of a communication link between two boards. Board #1 transmits to board #2 which responds. Board #1 starts to transmit shortly after power up and then continues to transmit.
- If we power up both boards at the same time, everything works.
- But when we power up board #2 after board #1, the L9963T on board #2 does not indicate that it is receiving (i.e. BNE is not set), even though we can see that board #1 is transmitting. Since board #2 is not receiving, it will not respond. We suspect that the L9963T on board #2 is somehow thrown off by the switching of signals on the transmission line while its power supply is ramping up, and is therefore not working properly. Has this effect been reported before?
HARDWARE CONFIGURATION:
Both boards:
- Slave mode (NSLAVE tied to GND)
- 1MHz SPI clock (SPICLKFREQ tied to GND via 9.09k resistor)
- BNE/CPOL and TXEN/CPHA used as GPIOs
- Transformer-Based insulation as stated in datasheet chapter 4.1 (Application information)
Board #1:
- ISOFREQ, TXAMP tied to GND
- DIS tied to GND via 10k resistor
Board #2:
- ISOFREQ, TXAMP, DIS tied to GND via one 1k resistor
