I am working with a board that has 2 DP83826E PHY devices. The board has a problem in that on powerup, the strap resistors will have random states because the pins have no pull ups or pull downs and the logic for applying power to the pins is faulty. During normal operations there is no voltage at the strap resistors, and during power up there are short pulses during the time the PHYS are being started, with suspect timing.
Given that we know that the PHY is configuring randomly at startup, the MDIO bus is stable and the PHY registers can be read and written.
My question is can we trust anything we are seeing in the register read outs.
The issues observed are as follows:
1) One of the PHYS has bit 10 (Isolation mode) set on BMCR on power up, the other does not. Register SOR2 has value 0, so it is not confirming isolation mode is set.
2) If a reset of the PHY is done, either soft reset on BMCR or hard reset on PHYRCR, BMCR does not return to default state. Bit 10 is still set, and SOR2 still reads 0.
3) On both PHYs, if LEDCFG is written a value in the 2 upper nibbles, those values will be reflected in BMCR. If bit 10 of BMCR is set, it is not cleared with a PHY soft or hard reset.
4) Values written to LEDCFG are not controlling the usage of LED1 or LED2. The board does not use LED3 .
Are these behaviors a side effect of the random start up configuration or significant of some other issue?
