DP83867CR: Request for checking my waveform

Hi ttd,

Normally, the signal level you probe is not the signal level we see in the PHY.

There is always some non-ideality due to the probe measurement:

• probe impedance mismatch
• probe position mismatch or variation
• differential impedance disturbance due to the probe measurement.

As long as the PHY is function properly, there shouldn't be any a huge concern on the waveform. If you want to send the waveform, you can send it through the private chat.

--

Regards,

Hillman Lin

Hi Hillman Lin,

Thank you for your reply.

I send the waveform through private chat.

If DP83867 has setting to gain drive strengh for RXD, could you share it.

It seems that RXD drive strengh it not enough.

Or, as you commented, it may be a probe impedance mismatch,

in which case any advice on how to improve the waveform with probing methods would be helpful.

Best Regards,

ttd

Hi ttd,

Thank you for sharing the waveform. It seems like you are concerning on the communication between the PHY and the MAC.

For some of the signal we do see the waveform is not dropping completely down or pulling completely up. Here are some questions we would like to ask further for confirmation:

• Where did you have this probe measurement in the schematic?
• Did you have any series resistors between the MAC and the PHY? It seems like the load might be too high or there is an impedance mismatch between on the RX_Data lines if the measurement is correct.

We do have register that could tune the impedance on the RX-Data lines: register 0x0170 bit[4:0]

--

Regards,

Hillman Lin

Hillman Lin,

Thank you for your reply.

There are series resistors(22Ω) between the MAC and the PHY.

See the schematic I sent through private chat. 

Also if you have any advice to  tune the appropriate impedance from 50Ω by trim on the RX-Data lines: register 0x0170 bit[4:0], please let me know.

In general,  the impedance control should be 50Ω default setting(0x0170) + change 22Ω to 0Ω series resistors.

However changing the resister from 20 ohms to 0 ohms improved the situation somewhat, but there was still no change in the High/Low swing.

The TX side swings fine, so I assume it is due to the ability of the PHY to drive the RXD.
If we lower the impedance control from 50 ohms (e.g. 40 ohms), will the ability to drive RXD increase?

In general, I think lowering the impedance would increase the output current.  

If there is other registers to improve the signal integlity like  0x0170 bit[4:0], please let me know.

 PCB trace impedance is 50Ω on my board.

Best Regards,

ttd

Hi ttd,

Thank you for the information

We have never seems any issue on the RX_Data before due to the driving strength for DP83867. 

Since the DP83867 is operating fine, there might be probe position or probe measurement on RX_data that you are seeing the behavior. If you are really concern on this, please try out tuning 0x0170 and see if you are seeing any improvement on the RX_data.

--

Regards,

Hillman Lin

Hi ttd, 

I shared waveform via internal message.

Could you check it?

It seems stable waveform(overshoot/undershoot and Tr/Tf are improved)  I  adjusted dampling resister and internal impedance of DP83867.

Could you check them just in case and if you have any concern, please let me know.

Best Regards,

ttd

Hi Hillman Lin,

Thank you for checking the waveform and I will evaluate functions after tuning the register setting any functional.

if I see a functional issue, I will ask you about it again.

Best Regards,

ttd

Hi ttd,

Thank you for your confirmation.

--

Regards,

Hillman Lin