This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

DS90UB960-Q1: CML output eye opening with the input jitter requirement

Part Number: DS90UB960-Q1

Hi TI Support Team

I want to know which parameter used verify 960 FPDLINK  forward channel performance. 

From datasheet description, the CML out can used to monitor the total jitters so the min eye width is 0.35UI,that means it allow the max jitter is 1-0.35=0.65 UI.

but the input jitter requirement from datasheet: max 0.4 UI for CDR and PLL of the receiver to track and recover the incoming serial data stream. 

Do you means the  1 - 0.4UI (total jitter) =0.6 UI for clean eye diagram.  But the mini eye width is 0.35 UI. So we have 0.25UI margin.  Is it correct?

CML output requirement: 
(1). CMLOUT supplies a buffered loop-through output driver to observe the jitter after equalization.

(2). CMLOUT monitors the post EQ stage, thus providing the recovered input of the deserializer signal.

(3). The measured serial data width on the CMLOUT loop-through is the total jitter including the internal driver, AEQ, back channel echo, and so forth. 

(4).  CMLOUT eye opening of at least 0.35 UI suggests that the forward channel signal integrity is likely acceptable. 

Input jitter requirement:

(1) Chapter 6.6 AC EC : tiJIT = 0.4UI max

Thanks

Yutai

  • Add figure in below. 

  • Hello,

    Let me follow up internally on the 960 jitter requirements. But instead of looking at the CMLOUT signal, a faster, more reliable way to verify link is to run the margin analysis tool. This test directly checks the forward channel link and avoids any user measurement errors.

    Check out the app note here

  • Hi Jiashow

    Thanks for your comments. Please help clarify  below query. 

    1. 960 input jitter requirement. 

    2. CML OUT test condition : how many UI  need used for eye diagram to measurement eye width from oscilloscope persistence mode. 

    3. Does the CML out eye width with the BER of E-12 (bit error rate @ 10^-12) from datasheet? 

    4. Do you have any application notes about how to CML out test setup. 

    Thanks

    Yutai

  • Hi Yutai,

    Please give me some time to look into this.

    Jiashow

  • Hi Yutai,

    I've verified with my team. The jitter requirement at the input of the 960 device is 0.4UI. On the CMLOUT, you should measure at least 0.35UI of eye opening. The CMLOUT has more lenient jitter requirement because the EQ circuitry may introduce more jitter to the signal, which gets extracted by the CDR.

    We have instructions on how to measure CMLOUT jitter from our validation team. If you send me a private message with your email, I can share it with you.

    Once again, I'd like to empathize that MAP tool is an easier way to diagnose and test the health of the link.

    Best,

    Jiashow

  • Hi Jiashow

    I put your comments in below in below and please share the CMLOUT instructions to  yutai.li@flex.com

    Do you have any comments about the item 2 and item 3. Does it include in your CMLOUT instructions ?

    1. 960 input jitter requirement. 

     Jiashow: 960 device is 0.4UI.  CMLOUT need 0.35UI of eye opening

    2. CML OUT test condition : how many UI  need used for eye diagram to measurement eye width from oscilloscope persistence mode. 

     Jiashow: ?

    3. Does the CML out eye width with the BER of E-12 (bit error rate @ 10^-12) from datasheet? 

     Jiashow: ?

    4. Do you have any application notes about how to CML out test setup. 

     Jiashow: We have instructions on how to measure CMLOUT jitter from our validation team.

    Thanks

    Yutai

  • Hi Yutai,

    It's based on BER of 10^-12.

    Duration for the test depends on system target confidence level. For example, 12.5 minutes for 95% confidence.  You can calculate it based on your confidence level.

    Best,

    Jiashow

  • Hi Jiashow

    Please send your instructions on " how to measure CMLOUT jitter from our validation team" to me.

    Mail:  yutai.li@flex.com

    Question: How about your typical characteristics the TJ@BER1 Eye diagram. Does this figure eye width based on BER of 10^-12 ?

                    I don't find any description from datasheet. 

    Thanks

    Yutai

  • Hi Yutai,

    I just sent you an email.

    Yes, it's based on BER of E-12

  • Hi Jiashow

    I will discuss with you in mail. 

    Thanks for your support. 

    Yutai