• State
  • Locked Locked
  • Replies 8 replies
  • Subscribers 33 subscribers
  • Views 324 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

DP83TD510E-EVM: Inquiry Regarding Configuration for Long Cable Reach with DP83TD510E

Takashi Usukura
Prodigy 10 points
Part Number: DP83TD510E-EVM
Other Parts Discussed in Thread: DP83TD510E

Tool/software:

Hi,

We are currently evaluating the DP83TD510E SPE-PHY chip using two evaluation boards to assess its performance, particularly the Long Cable Reach (1000m+).

During our research, we encountered the following scenario:
When integrating the DP83TD510E into actual hardware, the output voltage from a SPE terminal on a hardware can drop below the expected 2.4Vpp due to factors such as system internal wiring, EMC filters, and protection circuits. TI Japan advised us that the PHY's VoD swing level can be adjusted via register settings (address 0x0100) to compensate for this. We have confirmed that this adjustment indeed increases the VoD swing level.

However, we observed that increasing the PHY's VoD swing level beyond ~3.0V leads to deterioration in the SQI value (address 0x0A85), which negatively impacts long-distance cable performance. TI Japan suggested that we seek further advice on this issue through the E2E forum.

With this in mind, we have the following questions:

1. How can we achieve Long Cable Reach comparable to default VoD swing level(2.4V) when increasing the PHY's VoD swing level?
We think that proper configuration of the PHY's receive equalizer settings may be necessary. However, specific parameters for these adjustments are not documented in the datasheet. Could you advise on which registers to adjust and what values might be appropriate?

2. Would modifying the termination resistors R5 and R7 value (2kΩ) on the analog front-end allow us to maintain long-distance performance without changing existing equalizer settings and PHY output voltage regulations?

For reference, we have already tried the following approaches to achieve Long Cable Reach:

1. Application of the DSP ANEG script (referenced in Application Note sluaaa9a.pdf).
2. Adjustment of transmission line impedance resistors R1 and R2 (also detailed in sluaaa9a.pdf).
3. Application of the 6428.Force_2v4Vpp_Master_Init.txt script (sourced from a historical E2E forum thread).

We would greatly appreciate any guidance or additional recommendations you can provide.

  • 0
    TI__Expert 6925 points

    Hi Takashi, 

    What is the requirement you are trying to meet with the cable reach? Is this requirement not being met with the additional scripts from the app note sluaaa9a? 

    Takashi Usukura said:
    However, we observed that increasing the PHY's VoD swing level beyond ~3.0V leads to deterioration in the SQI value

    This is expected behaviour. The received signal goes through a PAM3 Slicer, which in turn generates the mean square error based on the sliced output level. Since you are increasing the voltage swing significantly beyond the expected 2.4V, this may be the reason why SQI is higher than usual. 

    Takashi Usukura said:
    How can we achieve Long Cable Reach comparable to default VoD swing level(2.4V) when increasing the PHY's VoD swing level?

    Adjusting the swing level has mostly been used as a tool to ensure that the output voltage is as close to 2.4V as possible, not for increasing cable reach. 

    Takashi Usukura said:
    Would modifying the termination resistors R5 and R7 value (2kΩ) on the analog front-end allow us to maintain long-distance performance without changing existing equalizer settings and PHY output voltage regulations?

    These resistors would not help in increasing cable reach, and are recommended to be left as 2kOhms.

    I wanted to ask some more things about your setup to better understand this application. 

    • What voltage level are you seeing on the recieved signal with the standard 2.4V swing level?
    • Are you able to link up successfully in this configuration?
    • What is the SQI value in this configuration. 

    Best,

    Vivaan

  • 0 in reply to Vivaan Jaiswal
    Prodigy 10 points

    Hi Vivaan,
    Thanks for your comment.


    　What is the requirement you are trying to meet with the cable reach? Is this requirement not being met with the additional scripts from the app note sluaaa9a?

    ->We are indeed targeting a 1500m+ cable reach. Tests were done with the additional scripts from the app note sluaaa9a p.7.


    Takashi Usukura said:
    However, we observed that increasing the PHY's VoD swing level beyond ~3.0V leads to deterioration in the SQI value

    This is expected behaviour. The received signal goes through a PAM3 Slicer, which in turn generates the mean square error based on the sliced output level. Since you are increasing the voltage swing significantly beyond the expected 2.4V, this may be the reason why SQI is higher than usual.

    ->From your explanation, I understand that the comparator thresholds within the PAM3 slicer are currently configured to match the 2.4V VoD swing level.
    　On the other hand, I believe the PAM3 slicer could potentially accommodate different VoD swing levels if it has configurable parameters related to the VoD swing level.
    　So, my additional question is: Could the comparator thresholds or other operational parameters of the PAM slicer be adjusted by modifying PHY settings?


    I wanted to ask some more things about your setup to better understand this application.
    • What voltage level are you seeing on the recieved signal with the standard 2.4V swing level?
    ->I observed 2.2V as signal level.
    • Are you able to link up successfully in this configuration?
    ->No. Regarding waveform observations, it was confirmed that the waveform occasionally shifts from the LSM (Low Speed Mode) pattern to the PAM3 pattern and then reverts back. Then, a complete link up was not achieved.
    • What is the SQI value in this configuration.
    ->The SQI value was found to be random (ranging from 0x001A to 0x7FFF) after reading the 0x0A85 register multiple times.

    BR,

  • 0 in reply to Takashi Usukura
    TI__Expert 6925 points

    Hi Takashi, 

    I am not sure if there is functionality to be able to change the PAM3 Slicer parameters. Let me check internally with our design teams and respond back about this particular question early next week. 

    Given that you are unable to achieve link up in normal operation, I wanted to ask you to run the setup without auto-negotiation. As detailed in the app note referenced earlier, disabling AN should help in increasing cable reach. Since you are using the AN script, I wanted to try to run a test with AN disabled as a whole. More information on using forced mode instead of auto-negotiation can be found in section 3.2. 

    The variance in SQI may be because of the inconsistent link behaviour you described above, but I still wanted to inquire about the cable being used in this setup. Would you have the details such as characteristic impedance, return loss, insertion loss of this cable? Lets also check the SQI values with forced mode to see if the behaviour carries over or not. 

    Best,

    Vivaan

  • 0 in reply to Vivaan Jaiswal
    Prodigy 10 points

    Hi Vivaan,


    I am not sure if there is functionality to be able to change the PAM3 Slicer parameters. Let me check internally with our design teams and respond back about this particular question early next week.

    ->Have you received any advise/comments from your design team?


    Given that you are unable to achieve link up in normal operation, I wanted to ask you to run the setup without auto-negotiation. As detailed in the app note referenced earlier, disabling AN should help in increasing cable reach. Since you are using the AN script, I wanted to try to run a test with AN disabled as a whole. More information on using forced mode instead of auto-negotiation can be found in section 3.2.

    ->I also tested without AN, but the results did not achieve a longer reach compared to when AN was enabled."

    The variance in SQI may be because of the inconsistent link behaviour you described above, but I still wanted to inquire about the cable being used in this setup. Would you have the details such as characteristic impedance, return loss, insertion loss of this cable? Lets also check the SQI values with forced mode to see if the behaviour carries over or not.

    -> I have insertion loss characteristics of the cable. Since the characteristics of a cable I am using is not good compared to that of Siemens 6XV1830-5FH10, it complies with 10Base-T1L insertion loss profile by 1,000 meter.


    BR,

  • 0 in reply to Takashi Usukura
    TI__Expert 6925 points

    Hi Takashi, 

    I have not heard back from the design team yet, but I will be sure to let you know as soon as I do. I also wanted to investigate if the peripherals are meeting the requirements for this device since that is the most common reason for this behaviour that we have seen. 

    Takashi Usukura said:
    it complies with 10Base-T1L insertion loss profile by 1,000 meter.

    Can you please share the details of insertion loss, return loss, characteristic impedance, etc for this cable? If you are saying that this cable only meets the requirements for lengths up to 1000m, then we would not be able to support 1500m+ with this cable. Additionally, I also wanted to check on the CMC as well to ensure that it meets the requirements in the datasheet. It would also be helpful if you are able to share the schematic for this application. 

    Best,

    Vivaan

  • 0 in reply to Vivaan Jaiswal
    TI__Expert 6925 points

    Hi Takashi, 

    The PAM3 Slicer parameters are not able to be edited. Even if we were able to do so, I do not think it would have any effect on the link behavior observed, since it only affects the SQI value. 

    Were you able to verify the cable and CMC characteristics compared to the requirements for this device? I believe the cause may be due to the cable requirements only meeting up to 1000m.

    Best,

    Vivaan

  • 0 in reply to Vivaan Jaiswal
    Prodigy 10 points

    Hi Vivaan,

    Thank you for your confirmation.

    I understand that the PAM3 slicer parameters cannot be adjusted.

    Regarding the cable characteristics, they are similar to those of the Siemens cable mentioned in the application note (SLUAAA9A). The Siemens cable meets the 10Base-T1L insertion loss profile up to 1,200 meters, yet it successfully achieves a 2,000-meter reach. Therefore, I assumed that further optimization of the PHY equalizer adjustments outside of PAM slicer limitations might enable even longer reach.
    (Actually, some long-reach scripts indicate the presence of interesting parameters such as FGCA, AGC Gain, deq offset, MSE threshold, Hybrid Gain, etc. Those parameters could help us, hence I asked you how to use PHY equalizer parameters. )

    Additionally, the CMC device and configuration match those of the EVB, so I believe the CMC characteristics meet the PHY device’s performance requirements.

    BR,

  • 0 in reply to Takashi Usukura
    TI__Expert 6925 points

    Thank you for clarifying the setup, Takashi. You are correct, if the cable parameters are similar to that of the siemens cable used in the app note, it should be possible to achieve link up at these distances. 

    You mentioned earlier that you were not able to see link up using the ANEG script found in the app note. I wanted to try using force modes on this setup with this cable to see if the PHYs are able to link up that way. The app note goes into how to configure force modes instead of Auto-Negotiation. 

    This can be done by using bit 14 of register 0x1834 to set both the DUT PHY and the Link Partner PHY in Host/Client Configurations (One as Host, other as Client) and bit 12 of register 0x7200 to disable Auto-Negotiation. In Force Modes, the output voltage is selected through straps, so make sure that the device is strapped to 2.4Vpp. I am hoping that doing this procedure on both devices enables a link between the two. 

    Another thing I wanted to check on was the characteristic impedance of the cable you are using. You mentioned that the resistors R1/R2 were changed to match the cable impedance, so I wanted to confirm what value they were changed to. Would it be possible for you to provide a schematic so that I can understand the system better?

    Best.

    Vivaan