TCA9546A Communication

Hi. 
Thanks for prompt reply. Sorry i got confused in the command i wrote to you. I did manage to communicate with SD0/SC0 lines using the command:

i2cset -y -r 2 0x70 0x1

When i do that, i can see new device on address 0x50 in the chip. When i read the address i can see the contents of the EEPROM of the PHY transceiver.
According to the transceiver datasheet, the eeprom is on i2c address A0 (which in 7-bit addressing becomes 0x50 if i understand correctly). The datasheet says the PHY registers address is 0xAC (which in 7-bit addressing becomes 0x56), but i can see nothing on this address. Actually when i scan the bus i see only addresses 0x70 and 0x50.

Is there something i am missing. Why aren't all the addresses of the device showing up??

Thanks for you help :)

Hi Jonathan.

I do understand the difference between device address and register address. The PHY address is as mentioned 0xAC, and the EEPROM address is 0xA0 (which shifted left becomes 0x56 and 0x50 respectivly). and I do actually see the EEPROM at address 0x50 on the bus (when i use i2cdump on that address on the bus i see the EEPROM contents). my question is why can't i see the addres 0x56 on the bus. This address should contain some phy registers (which i am interested in reading/setting). My problem is that i can't see anything (no device) on the PHY address but can see the EEPROM address. 
The device PHY transciever and for that reason should be automatically detected, both PHY and EEPROM. but only EEPROM is showing up.

Is my problem more clear now?

for example if i would like to read AN_RX (auto-negotiation receive) register of the PHY, to see auto-negotiation values, i must write to bus number 2, phy address 0xAC (according to data sheet), register address 0x5, value bla bla. But i can see only EEPROM address 0x50 (0xA0 >> 1), but no phy address 0x56 (0xAC >> 1).

Hi.
The fake write to the address has solved it. 
Thank you very much

Hi Sri.

I think you are confusing the two ICs, The fake write was to the TCA9546 at address 0x70. It seems that you configured your TCA9546 correctly if you can see EEPROM contents at address 0x50. I've had a similar issue with one of the SFP connectors, but it was a specific one which did not show up at address 0x56 (just like yours). The rest were showing up at 0x56 and 0x50.

So it seems to me that your SFP connector have not initlalizaed it's phy component correctly.

Good luck.

Hi Sri.

Your connector is an IC made out of few components. one of them is the EEPROM which has i2c access on chip address 0x50 (which you see), one of them is the PHY which also has i2c access but on address 0x56 (usually so, if you follow standards). It seems that you can see only the EEPROM IC component, but not the PHY IC component. For debugging i would suggest to first try another converter (another company), I would also try and see the manufacturer's EEPROM contents, maybe there is kind of error status register. 
Also there are additional (some optional) lines going to the SFP connectors - tx_fault, tx_disable, dev_detect, rate (optional). 
TX_DISABLE is output from the processor, make sure this line is tied low when you activate the PHY. After a successful PHY initialization, the PHY will de-assert tx_fault line which should indicate that it is ready to transmit. Reset your PHY and check that this line has been tied low shortly after reset.
Last thing you can try is to change the sequence of initializations of the components (in my case when configuring the tx_fault, tx_disable, dev_detect and rate gpios BEFORE asserting ETH_RST would cause the phy not to start, changing the sequence of those made it work for me, i.e first assert ETH_RST, then initialize tx_fault, tx_disable, dev_detect and rate as GPIO and give values).

Hope that helps :)