TJA1103 PHY registers about SYMBOL_ERROR_COUNTER

[forfortuna]

Dear NXP Technical Support Team,

I hope this email finds you well. My name is SeongHyeon Lim, and I am currently developing an automotive Ethernet diagnostic system using the TJA1103 100BASE-T1 PHY.

I am writing to inquire about the specific behavior of the SYMBOL_ERROR_COUNTER, as we are trying to implement a predictive maintenance feature to detect gradual degradation of Ethernet cables and connectors.

System Overview:

• Product: TJA1103 100BASE-T1 PHY on [S32K3X4EVB-T172]
• Objective: To detect communication quality degradation due to cable/connector aging for predictive maintenance.
• Monitored Registers:
  • SIGNAL_QUALITY (Shared MMD, Address 0x8320)
  • SYMBOL_ERROR_COUNTER (Shared MMD, Address 0x8350)

Observed Behavior:

In our tests where we intentionally degrade the signal quality, we can observe the SQI value in the SIGNAL_QUALITY register gradually decreasing from 7 down to 4 or 3. Our expectation was that as the SQI value drops, the SYMBOL_ERROR_COUNTER should begin to increment, at least intermittently.

However, what we observe is that the SYMBOL_ERROR_COUNTER remains at zero, even with a significantly degraded SQI (e.g., 3 or 4). The counter only appears to increment rapidly at the moment the SQI drops to 0, which coincides with a link-down event.

Our Hypothesis: We suspect this is due to the TJA1103's robust internal DSP and adaptive equalizer, which effectively correct nearly all signal errors as long as the link is considered "available" (i.e., LINK_AVAILABLE bit is set). Therefore, symbol errors are only counted when the PHY's correction capabilities are completely overwhelmed, leading to a link failure.

Questions:

1. Could you please confirm if this observed behavior—the SYMBOL_ERROR_COUNTER remaining at zero for SQI values greater than zero—is the intended and expected operation for the TJA1103?

2. If this is the expected behavior, for the purpose of predictive maintenance of gradual cable aging, would NXP recommend focusing primarily on monitoring the downward trend of SQI and the upward trend of MSE, rather than using the SYMBOL_ERROR_COUNTER as an early indicator?

3. Is there any further technical documentation or clarification available regarding the specific conditions under which the SYMBOL_ERROR_COUNTER begins to increment (e.g., a specific internal state, a threshold other than link failure)?

Thank you for your time and assistance. We are available to provide additional information, such as code snippets or logs, if required.

[PavelL]

Hello @forfortuna ,

Thank you for your clear and well-structured question.

Your observations and hypothesis are correct. The SYMBOL_ERROR_COUNTER on the TJA1103 typically remains at zero while SQI is above zero, even if the signal quality is degraded. This is expected behavior.

1. SYMBOL_ERROR_COUNTER Behavior

Yes, this is the intended operation. The counter only increments when LINK_AVAILABLE = 1 and COUNTERS_ENABLE = 1, and usually only when the PHY’s internal correction mechanisms are no longer sufficient - typically when SQI drops to 0 and the link goes down.

2. Recommendation for Monitoring

For predictive maintenance, we recommend focusing on the SQI trend and MSE. These provide earlier and more gradual indicators of link degradation than the SYMBOL_ERROR_COUNTER.

3. Additional Documentation

There is no further public documentation available regarding the internal conditions that trigger the SYMBOL_ERROR_COUNTER.

Best regards,

Pavel