Hehe, already been through it all with the 6610 a couple years ago - it needs to see an i2c bus with two devices (per psu) on it for the psu's - the psu itself, and an EEPROM
For the psu device address it needs to see 9 or 10 different sensors/registers for things like rail voltages, temps, fan speed etc. There's also some SMbus handshaking going on when it first powers on - which I did manage to capture years back with a logic analyzer, not sure if I still have it. For the EEPROM device address it needs to see a 16kbit EEPROM populated with valid data (psu model id, serial, vendor and revision)
I have dumps of most of these I can paste when I'm not on mobile, but as I've said before it's *really* not worth it. I've been down this road several times before with this switch and it is a hilarious amount of work and ghetto-rigging all to save 2 to 3 watts and maybe 5db at best off of an already quite efficient switch
For the psu device address it needs to see 9 or 10 different sensors/registers for things like rail voltages, temps, fan speed etc. There's also some SMbus handshaking going on when it first powers on - which I did manage to capture years back with a logic analyzer, not sure if I still have it. For the EEPROM device address it needs to see a 16kbit EEPROM populated with valid data (psu model id, serial, vendor and revision)
I have dumps of most of these I can paste when I'm not on mobile, but as I've said before it's *really* not worth it. I've been down this road several times before with this switch and it is a hilarious amount of work and ghetto-rigging all to save 2 to 3 watts and maybe 5db at best off of an already quite efficient switch
Last edited: