DP83869HM: TDR function : Cable length

Dino,

Is the customer using our EVM or their own board? If they are using our EVM's we can replicate and then figure out any underlying issues.

In the meantime, here is an application note for TDR:

https://www.ti.com/lit/an/snla253/snla253.pdf?ts=1666637234610&ref_url=https%253A%252F%252Fwww.google.com%252F

Please note that this is for DP83822 and should be used as a reference; Customer can follow similar steps, but have to make sure the registers align with 869 datasheet.

Sincerely,

Jason Lee

Hello Jason, 

Thank you for your comment. 

The customer is using their own board and they are comparing DP83869 and competitor, BCM5482S on their own board. 

As you mentioned before, we followed similar step but the result also has big gap. please refer the below. 

- Convert Hex Value of a peak to Decimal Value(DV).

- Find the Intermediate Location (IL) = (DV x 0.8621) - 8

- Find Final Location(FL) in meters = IL + ((70 - IL) x 0.01)

In case of 30m real cable test, 

DV value is 52(0x34) , IL = 36.8292 , FL = 37.16 meter , gap with real cable length :7.16m 

For 3 meter cable test, 

DV value is 20(0x14) , IL = 9.242 , FL = 9.86 meter , gap with real cable length :6.86m

The gap is bigger then competitor device result. the competitor device gap is only 1 meter.  

Please review it again above equation is correct for DP83869HM and let me know your opinion. 

If above equation is correct, please let me know which factor cause big gap between real cable length and measured value. 

By the way, At the below post, I can see blow comment that the equation you mentioned is not correct for DP83869HM. 

https://e2e.ti.com/support/interface-group/interface/f/interface-forum/1042612/dp83869hm-tdr-details?tisearch=e2e-sitesearch&keymatch=DP83869%2526TDR# 

- Can you confirm the constants used to calculate IL are also valid for DP83869HM?

   + I would first like to clarify that the constant used for IL formula is -17 instead of -8. In addition, this is the entire formula needed to calculate the location. There is no need for FL formula as this is the only formula.

        Peak_Location (m) = (DV x 0.8621)-17

On the first my post, there is the result when we used above equation,(DV x 0.8621)-17 But I mentioned before, there is also big gap. 

Which one is correct? please let me know your opinion regarding correct equation for DP83869HM. 

Thank you. 

Dino,

I am going to check these calculations again and confirm with the Applications team. I will keep you updated with this investigation; Thank you for your patience and feel free to reach out at any time.

Sincerely,

Jason Lee

Hello Jason, 

I am sorry to push you but Is there any update on above request? 

The customer are waiting TI comment. 

Thank you. 

Dino,

Thank you for your patience. I confirmed that the following is the correct equation:

Peak_Location (m) = (DV x 0.8621) - 17

Please let me know if there is anything else I can support you with!

Sincerely,

Jason Lee

Hi Jason, 

Thank you for your reply. 

I mentioned first article, I already used equation, Peak_Location (m) = (DV x 0.8621) - 17 and test cable length 

But The customer got a result 27.824m with real 30m cable, 0.24m with real 3m cable. there is a big gap 

I mentioned first post, he competitor device is BCM5482S. The result using competitor device are 31m under 30m cable and 4m under 3m cable. 

At datasheet of  DP83869HM, We can see comment of +- 1m accuracy but there is a big gap between real cable length and measured value. 

Please review again and let me know your opinion regarding this result.  If above equation is correct, please let me know which factor cause big gap between real cable length and measured value. 

Thank you. 

Hi Jason, 

I am sorry to push you but the customer are waiting TI comment. 

I can't find any comment and document for DP83869 TDR function equation. 

Please review it again and let me know your opinion. 

Thank you. 

Dino,

We are still in the process of creating a DP83869 application note; However, the function that you utilized is still the correct one. In the meantime, in terms of other areas to troubleshoot:

• Have you tried using a DP83869 EVM instead of your own?
• Could we check that they are accessing the registers correctly?

Sincerely,

Jason Lee

Jason, 

James Lee said:

Have you tried using a DP83869 EVM instead of your own?

the customer are already using DP83867 for their other project. I know that DP83867 and DP83869 are similar device and  has same equation for TDR function. Using DP83867 EVM, the customer has similar result. 28.6M as measured value using 30m cable. 

James Lee said:

Could we check that they are accessing the registers correctly?

How can we check that they are accessing the resisters correctly? As mentioned above, the customer has experienced with TI Ethernet PHY.

 the peak location values is from register 0x190 to 0x199. is it right? I don't think they are accessing the registers correctly. 

BTW, you mentioned first , the equation is (DV x 0.8621) - 8, and next, correct equation is Peak_Location (m) = (DV x 0.8621) - 17 

What does -8 and -17 mean? and How is this value derived? Is this changeable? 

If we use -15 value, the result come within the +-1m as mentioned datasheet. what do you think about this? please let me know your opinion. 

Thank you. 

Dino,

We are going to go through testing with our own cables and EVM to confirm these for you. In the meantime, I am reaching out to the Design team in order to ensure we have not missed any factors. Thank you for your patience!

Sincerely,

Jason Lee

Jason, 

Thank you for your strong support. 

When can I get your response? Please let me know the approximate schedule.

Thank you. 

Dino,

We conducted testing with 3m and 15m cables and also found varying results. I have reached out to Design to investigate further into this.

Sincerely,

Jason Lee

Jason, 

Is there any update on this request? Please let me know the approximate schedule. 

Thank you. 

Dino,

Our Apps team is meeting with the Design team this week to discuss the varying results from our testing; I estimate that the Design team in India will give us an answer some time next week.

Sincerely,

Jason Lee

Jason, 

Is there any update on this request? The customer has been waiting for our answer for a month.

Thank you.

Dino,

We are still waiting on Design to respond to us with conclusive information regarding this topic; Thank you for your patience. We really do appreciate it and are working to solve this issue for the long-term.

Sincerely,

Jason Lee