DP83826E: IEEE 10base-T compliance test failed

Hi Chen,

I would like to double check, did you use the following script for compliance?

If possible, may I ask some questions on the compliance setup?

• What is cable length that you used for compliance?
• Did you use compliance fixture when you are performing the test?

--

Regards,

Hillman Lin

Hi Hillman,

Yes, we just set the commands like the picture to make the test, but not set the final line(restart PHY) until measurement ended .

1.The test connector to differential probe is no more than 10cm. Probe is TEK P6248 and its length is 1.2m.

2.Yes, we used the fixture for 10Base-T to do this test

Sincerely,

Ken.Chen

Hi Chen,

If possible, could you try the software reset and see if that help with the performance.

The next step I would like to check is the schematic around the MDI lines or magnetic side.

--

Regards,

Hillman Lin

Hi Hillman,

I think I found the problem, the compliance test we undergoing was 10base-T and the commands setting configuration of DP83826 is 10base-Te.

Therefore, turning off the energy efficient ethernet(EEE) might solve our question.

Could you tell me how to set 10base-T configuration of this PHY?

Sincerely,

Ken.Chen

Hi Hillman,

There is one more question, the harmonic measurement must all-one Manchester-encoded signal,

(Captured in IEEE802.3 page400)

however, when we set 0x16 as 0x7101 and PHY output random packets.

I also tried the address 0x27 but it was only generating a single pulse and not satisfied with TEK's tool.

How could I undergo this test item?

--

Sincerely,

Ken.Chen

Hi Chen,

Are you looking for compliance test on 10base-T?

If so, the following script is used to set 10base-T for compliance testing:

--

Regards,

Hillman Lin

Hi Hillman,

That scripts are setting for 10base-Te and it is write at first row.

It's not 10base-T.

Sincerely,

Ken.Chen

Hi Chen,

Unfortunately, DP83822 does not support 10Base-T.

--

Sincerely,

Hillman Lin

Hi Hillman,

Understood, so DP83826 is also not support 10Base-T, right? 

There is one more question, the harmonic measurement must all-one Manchester-encoded signal,

(Captured in IEEE802.3 page400)

however, when we set 0x16 as 0x7101 and PHY output random packets.

I also tried the address 0x27 but it was only generating a single pulse and not satisfied with TEK's tool.

How could I undergo this test item?

--

Sincerely,

Ken.Chen

Hi Chen,

Yes, you are correct.

For harmonic test, we only have all ones signals without encoding in the DP83826PHY configuration. This is also valid for 10Base-Harmonic setup. Customer could tune this test mode in the scope or the test equipment.

--

Regards,

Hillman Lin

Hi Hillman,

Tek's scope expect the waveform to measure harmonic test. However, all ones signals without encoding will not make this TP_IDLE and tools can't accept this type of signals.

The report I uploaded is setting PRBS configuration. It's can be accepted by scope but sending random packet and result is failed.

Can you tell me how to configurate this PHY to make all ones signals with TP_IDLE?

--

Sincerely,

Ken.Chen

Hi Hillman,

Thanks a lot for your help. However, our tool seems different and it has only 2 options can be changed.

The tool we selected as below. 

Could you tell me what's your full name of software and edition? 

Sincerely,

Ken.Chen

Hi Chen,

Sorry for the delay. We confirm the license for on our scope. The license we are using is TekExpress Framework Version 4.17.0.58.

--

Regards,

Hillman Lin

Hi Hillman,

Can tell me which software you choose to do this harmonic test?

Where you get the UI can change the signal type to "ALL ONES NO IPG"?

--

Sincerely,

Ken.Chen

Hi Chen,

Here is the section I choose for the signal path:

• \

I think it might be software version dependence. Could customer update the software for TekExpress? It should be free to update the software.

--

Regards,

Hillman Lin