Nice progress at KL7NA! I was able to connect to RX1 for the first time last night, and then RX2 and RX3 today. Just to make sure I understand the setup:
RX0,RX1 are on a v1.1 connected to your yagi. (Listed incorrectly as v1.2 and dipole.)
RX2,RX3 are on a v1.0 connected you your dipole. (Listed incorrectly as connected to yagi.)
Is that correct?
Sometimes different ethernet connections can exhibit the problems you saw with your first v1.1. Did you switch BeMicro boards, ethernet cables, or whatever device the other end of the ethernet cable is connected to?
I have a 9:1 balun on mine and found that it is an improvement.
With the new firmware, there are two LEDs (farthest from the heartbeat) that indicate clipping on the ADC, one led that is a heartbeat, and five LEDs (nearest to the heartbeat) that indicate the DAC transmit level. Something is wrong if you see all of the LEDs light up when not transmitting.
The variation in reset requirements is interesting. I did not forget about the importance of reset in the prototype. I was following the AD9866 evaluation board, which has a separate chip for reset, and knew I would leave that off. I didn't know how I would eventually do reset, and since there was an exposed pad on the pull-up resistor, I left it unconnected with the idea to connect a wire to the exposed resistor pin later for experiments. But when I didn't have a reset issue, I was distracted by other issues. I need to revisit my board that doesn't work and see if it is a reset issue. On page 43 of the datasheet, it recommends reset be held low for at least 10ms after power up. A proper reset sequence in the RTL is coming after I finish some front-end boards for v1.2.
So far, 6 Hermes-Lite variants have been built: 2 v0.9, 2 v1.0 and 2 v1.1. It would be nice to know of any variations within this group.