Hello
I have a custom board that is using the TLK105 ethernet PHY.
The TLK105 stops responding after a power cycle when the supply voltage only drops to approximately 292mV before rising to 3.3V again. The 292mV is supplied through pull-up resisters needed for external circuitry equivalent to the following:
Multiple different attempts have been regain control of the PHY. This includes timing and the usage of both the power down pin and the reset pin.
The following sequence is by far the fastest and most reliable of the attempted sequences.
-
Disconnect all power
-
Connect RS485 and 3.3V supply
-
Pull the reset pin low for approximately 280µs
(280µs seems to work better than both shorter and longer reset times (1µs and 2ms)) -
System runs as expected
-
Remove 3.3V supply
-
Leave the device disconnected for a few seconds to ensure we have reached the final voltage
-
Reconnect the 3.3V supply
-
Pull the reset pin low for approximately 280µs
-
The PHY now won't respond to communication on the Serial Management Interface.
Ex. reading the BMCR register just returns 0xFFFF. -
Often the PHY returns to normal operation after approximately 25 seconds. Either by adding a plain delay after performing the reset or by reading the BMCR register until it returns a valid value (reset bit is cleared)
However 25 seconds is a long time to wait for the PHY to get ready as the device is practically useless without it.
Can anyone provide me with a fast and reliable way to return the PHY to a working ?
