DP83869HM: About the SGMII amplitude change settings

Hi,

I don't have much understanding on this. Based on the thread before, it seems like D6[11:10] are able to tune the the swing level. Have you try register 00d6 to 0C00 and see any peak to peak change?

May I also ask what is your SGMII trace length? It seems like your SGMII singal is smaller than we expected. If possible could you also share the schematic on the SGMII lines?

--

Regards,

Hillman Lin

Write 0C00 to register 0d6. The waveform is as follows.
mdio-tool w eth0 0x0D 0x001F
mdio-tool w eth0 0x0E 0x00d6
mdio-tool w eth0 0x0D 0x401F
mdio-tool w eth0 0x0E 0x0c00

Schematics cannot be disclosed. Apologies.
The SGMII trace length is approximately 6.3inch.

Hi,

I think the main reason that we are seeing a decrease in SGMII amplitude are due to the layout constraint. Here are the main thing I would like to check on the layout: impedance matching, trace length, and distance between the trace. We also have the layout checklist for you to review.

Industrial_PHY_Layout Review Checklist.xlsx

--

Regards,

Hillman Lin

Hi　Hillman Lin
I want to reduce the amplitude by register setting from the measured amplitude of 964 mV.

Question 1
B: Does setting 01 to 0x00D6[11:10] result in 0 V. It is strange that there is no output.

Question 2.
I understand that trace length etc. acts to reduce the amplitude.
A: 00 setting to 0x00D6[11:10]: 136 mV drop.
C:10 setting to 0x00D6[11:10]: the measured value is 955 mV.
Is it not reasonable that 920 mV (set value) - 136 mV (dropped value) = 784 mV?
The measured value of 955 mV is larger than the set value of 920 mV and it is doubtful that the setting of register 0x00D6[11:10] is effective.

Please check if it is really possible to change the amplitude with 0x00D6[11:10].

Regards,

Hi,

Again, we have limited knowledge on this two bits because we usually make sure customer is able to pass the eye diagram before we provide further instruction on tuning the amplitude.

My approach would be cleaning up the layout because we usually don't recommend to tune the peak to peak of SGMII.

Based on the waveform you provided, it seems like even you increasing the peak to peak signal won't help with the SGMII link up because your eye diagram does not satisfy the eye diagram requirement of the SGMII signal. Following FAQ would be helpful:

https://e2e.ti.com/support/interface-group/interface/f/interface-forum/1247949/faq-dp83tg720r-q1-why-is-my-sgmii-eye-measurement-distorted?tisearch=e2e-sitesearch&keymatch=eye%20diagram#

--

Regards,

Hillman Lin

Yes; Hillman Lin.

Available at: e2e.ti.com/.../faq-dp83tg720r-q1-why-is-my-sgmii-eye-measurement- distorted?tisearch=e2e-sitesearch&keymatch=eye%20diagram#.

I have confirmed the above.

I checked the waveform again with a wideband oscilloscope.
The waveforms are shown below.
I want to reduce the amplitude because it is over vdiffp-p in the SGMII mask.

Regards,

Hi,

Sorry for the confusion. The SGMII waveform you send us are reasonable. Based on the datasheet SGMII output. the swing level should be somewhere around 550mV.

https://e2e.ti.com/support/interface-group/interface/f/interface-forum/1287391/faq-dp83869hm-sgmii-output-differential-voltage-clarification?tisearch=e2e-sitesearch&keymatch=SGMII#

--

Regards,

Hillman Lin

Hi,Hillman Lin.

Does this mean that the following answers are incorrect?
I understood that 10 - VOD = 920mV could be set.

Regards,

DP83869HM: DP83869HM: SGMII voltage swing_2 - Interface forum - Interface - TI E2E support forums

Hi, 

Again, I am not too familiar with this register tuning.

• Are you facing any issue with the SGMII communication?

--

Regards,

Hillman Lin

Hi,Hillman Lin

>>Again, I am not too familiar with this register tuning.
　→Could someone familiar with the register please confirm this?

>>・Are you facing any issue with the SGMII communication?

>>・Does cannot fit in your application?

　　→he vdiffp-p of the receiver IC used in this project is 1000 mV, so the output voltage of the DP83869HM is over.

Hi,

I double check with the team internally.

They still want to know if you are seeing any functional issue 1100mV amplitude on SGMII communication? 

What is the maximum SGMII p2p that the receiver IC could support. Because based on the standard, 1100mV is valid for SGMII communication.

Thank you for your time

--

Sincerely,

Hillman Lin

Hi,Hillman Lin.

>They still want to know if you are seeing any functional issue 1100mV amplitude on SGMII communication? 
  →he vdiffp-p of the receiver IC used in this project is 1000 mV, so the output voltage of the DP83869HM is over.

>What is the maximum SGMII p2p that the receiver IC could support. Because based on the standard, 1100mV is valid for SGMII communication.
  →he vdiffp-p of the receiver IC used in this project is 1000 mV, so the output voltage of the DP83869HM is over.

★My question is: Why can't I change the amplitude at 0x00D6[11:10]? It is.
Does this answer mean that it is wrong?
e2e.ti.com/.../dp83869hm-dp83869hm-sgmii-voltage-swing_2 e2e-sitesearch&keymatch=DP83869HM%253A%2520DP83869HM%253A%2520SGMII%2520voltage%2520swing_2

Regards,

Hi,

Here are some more configuration that you could try on your side. Try to write register 0x00D6 = 4000 or register 0x00D6 = 6000 and see if you are able to reduce the peak to peak of SGMII swing level?

--

Regards,

Hillman Lin

Hi,Hillman Lin.

I implemented the settings you answered.
I was able to confirm that the amplitude is reduced.
I am attaching the acquired waveforms for your reference.

★Question 1
　 When setting 0x00D6 = 4000: What is the expected value of Vdiffp-p in mV?
★Question 2
　 When setting 0x00D6 = 6000: What is the expected value of Vdiffp-p in mV?

Regards,

# mdio-tool w eth0 0x0D 0x001F
# mdio-tool w eth0 0x0E 0x00D6
# mdio-tool w eth0 0x0D 0x401F
# mdio-tool w eth0 0x0E 0x4000


# mdio-tool w eth0 0x0D 0x001F
# mdio-tool w eth0 0x0E 0x00D6
# mdio-tool w eth0 0x0D 0x401F
# mdio-tool w eth0 0x0E 0x6000

Hi,Hillman Lin.

Thank you very much.
The answers you gave me helped me understand.

Regards