LMH1297: Closed Eye Opening at HD/3g rates

Hi,

Can you please share a picture of your eye diagram measurement setup? Where is the eye being measured? 

The eye itself looks to be under-equalized. So if you are measuring the eye at the far end of a cable, you can use the LMH1297 OUT pre-emphasis to compensate the loss of the cable and open up the eye.

Thanks

David

Eye is being measured from output of a LMH1297 Working as a CD. Being inputted to a Phabrix QX and displayed from there.

Cables are 4855R 1Meter in length.

Source gen is internal to Phabrix as well as an BrightEye 57.

Circuit in question is a LMH1297 Running as an input EQ with a diff pair with 4.7uf AC coupling feeding into another LMH1297 running as a CD.

Hi,

If you change the LMH1297 output pre-emphasis, do you see an improvement in the eye diagram?

Thanks

David

Adjusted the pre-emphasis with no discernable change to eye pattern.

Did pull the signal going to the driver over via SPF to a different Driver and the eye was better.

Hi,

I have the below setup

PHABRIX TX ---> Belden 1m BNC cable ---> #1 LMH1297 in EQ mode (SDI_IO) ---> #1 LMH1297 in EQ mode (OUT0) ---> #2 LMH1297 in CD mode (IN0) ---> #2 LMH1297 in CD mode (OUT0) ---> Belden 5m cable ---> PHABRIX RX. Below is the eye I am seeing which looks to be pretty equalized.

Have you tried another 4855R 1M cable or are you seeing this issue on multiple boards? I don't expect the LMH1297 output eye would be under-equalized with the 1m Belden cable.

Thanks

David

David,

our setup seems to flow slightly differently:

PHABRIX TX ---> Belden 1m BNC cable ---> #1 LMH1297 in EQ mode (SDI_IO) ---> #1 LMH1297 in EQ mode (OUT0) ---> #2 LMH1297 in CD mode (IN0) ---> #2 LMH1297 in CD mode (SDI_IO) ---> Belden 1m cable ---> PHABRIX RX 

The above eye being driven from the CD, multiple cables and even going back to 1694A 

the Tail end of the LMH1297 chain ends with OUT0 terminating on a SFP the signal from that SFP has a clean EYE

Chris

It was a typo in my setup description, so our setup is the same.

For #1 LMH1297, EQ/CD_SEL = L, SDI_OUT_SEL = H,

For #2 LMH1297, EQ/CD_SEL = H, OUT0_SEL = H, and SDI_OUT_SEL = H. 

Are you using the same BNC connector for both the Belden cable and SFP? If you are, and also seeing a clean eye with the SFP, then the LMH1297 SDI_IO to the BNC connector is ok, and the issue still looks to be a cable issue. 

Below is also my PHABRIX generator setup. I played with different resolution, pathological insertion, none of them made a difference to the LMH1297 output which is expected.

Thanks

David

David,

Going over the design looks like a bad impedance stack is causing the front of the eye to be closed. In contact with board house now to see how that happened but looks like we ended up with a 63 ohm impendence on the output lines.

I will get this addressed and report back if that was the fix for this issue.