cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
mcieslak
Visitor
Visitor
2,615 Views
Registered: ‎10-16-2017

Meaning of UT sign in EyeScan measurement

Jump to solution

Hello all,

 

I'm working on my own IBERT project for Kintex Ultrascale GTH IBERT.

Currently I got the problem, that I do not fully Understand the following statement in the data sheet:

 

UG576(v1.5) p. 220:

When in DFE mode (RXLPMEN=0), due to the unrolled first DFE tap, two separate EyeScans
measurements are needed, one at +UT and one at –UT, to measure the TOTAL BER at a given
vertical and horizontal offset.

 

How does this have to be interpreted in DFE mode of course? My clues are:

At first, I have to do 2 separated EyeScans, one with +UT and one with -UT. That's obvious I think. But how to continue?

  1. Add the resulting open areas of each EyeScan to the 'total' open area.
  2. Add the resulting sample and error counters on each position. And THEN calculate the Open Area.
  3. Only Calculate the Open Area of the overlapping part of both Scans. The Scans look very different and have an overlapping part in the centre of the diagram.

 

Or is there any other way the OpenArea results from these Scans?

 

I realy need some help as the reults influence the the rest of my algorithm.

 

Best Regards

Michael

Tags (3)
0 Kudos
1 Solution

Accepted Solutions
mcieslak
Visitor
Visitor
3,234 Views
Registered: ‎10-16-2017

Hello again,

 

I have finally found an explanation in UG578(v1.3) p. 230f.

There it explains perfectly good how to measure the BER in DFE mode. I still do not know how the UT sign influences the Scan in detail, but this doesn't matter anymore. Now I know how to deal with it in EyeScan.

 

Dear Xilinx,

it would be a fantastic if you can copy this information into the UG576 for GTH Transceivers.

I assume some other users may have searched for this but nobody searches in the manual for GTY transceivers if you work with GTH.

 

Best Regards

Michael

View solution in original post

0 Kudos
3 Replies
roym
Moderator
Moderator
2,584 Views
Registered: ‎07-30-2007

Use the UT = 0 scan for the right side of the eye and the UT = 1 scan for the left side of the eye. Construct a single eye from the 2 scans and then calculate the open area.  See also Manual Eye Scan . See DFE lecture pages 8-13 for an explanation of the separation.

 




----------------------------------------------------------------------------
Don't forget to reply, kudo, and accept as solution
Be sure to visit the Resources post periodically to keep up with the latest
https://forums.xilinx.com/t5/Serial-Transceivers/Serial-Transceiver-Forum-Guidelines-and-Useful-Resources/td-p/1173590
----------------------------------------------------------------------------


0 Kudos
mcieslak
Visitor
Visitor
2,552 Views
Registered: ‎10-16-2017

Hello roym,

 

thank you very much for your quick response.

 

I already read the documents you mentioned. The DFE lecture contains a lot information how the DFE in general works and how to configure the coefficients. The UT sign seems to be part of the Slicer. The lecture does not mention its influence in the eyescan.

And your other reference (AR# 67295) only uses LPM mode. DFE is not mentions. By the way. I used this manual to build my eyescan algorithm.

 

However, I did some measurement and attached it to this post.

As you can see there the UT sign does not influence the right and left part of the eye. It has more impact on top and lower part of the scan. In the centre they perfectly overlap to a nice-looking eye. It is not in the file but I think it can be assumed by the Error Data arrays.

So, why not just overlap the Data (for example by addition) and then calculate the resulting BER of the sum?

 

0 Kudos
mcieslak
Visitor
Visitor
3,235 Views
Registered: ‎10-16-2017

Hello again,

 

I have finally found an explanation in UG578(v1.3) p. 230f.

There it explains perfectly good how to measure the BER in DFE mode. I still do not know how the UT sign influences the Scan in detail, but this doesn't matter anymore. Now I know how to deal with it in EyeScan.

 

Dear Xilinx,

it would be a fantastic if you can copy this information into the UG576 for GTH Transceivers.

I assume some other users may have searched for this but nobody searches in the manual for GTY transceivers if you work with GTH.

 

Best Regards

Michael

View solution in original post

0 Kudos