DP83869HM: Request for information/more details for 0x01EC register. Unable to set this register to 0x1FFC

Hello,

Could you please let me know what the value of register 0x3 is?

Sincerely,

Gerome

Hello Gerome,

Thank you for responding to this post, the 0x3 register value read is  0xa0f1

Hello,

What are you using to read/write to these registers? Would like to ensure you are correctly accessing these registers.

Also, before you write these values, what are the initial values for 0x1DF, 0x0, 0x10, and 0x1EC?

In addition, if you only are using the PHY in 100M Media Converter Mode, why not strap the PHY into this mode initially instead of writing to registers upon every power up initialization? We have a list of strap options listed on SNLA318 on section 1.7 for this. See if this fixes your issue.

Sincerely,

Gerome

Hello Gerome,

1. We are reading and writing the registers from driver file fec_main.c, below is the screen shot of the read and write methods,

2. The initial values for 0x1DF, 0x0, 0x10, and 0x1EC are as below and also read STRAP_STS register,

                 Register 0x3 before write is: 0xa0f1
                 Register 0x1EC before write is: 0x0
                 Register 0x0000 before write is: 0x1140
                 Register 0x0010 before write is: 0x5048
                 Register 0x01DF before write is: 0x0
                 Register 0x6e (STRAP_STS register) dump: 0x10b1

3. We have strapped the DP83869 as per below table. I am also attaching the schematics for better understanding,                             

1581.DP_Schematic_forTI.pdf

Please let me know if any further information is needed.

Regards,

G.Shricharan.

Hi,

I will need a couple days to take a look at the schematic. I expect a response by Thursday.

Sincerely,

Gerome

Hello,

Your strap resistor settings look right. It is very curious why the PHY isn't initially going into that mode, instead according to your register dump, it is in RGMII to Copper mode. Can you send a scope capture of your power supplies ramping up?

Sincerely,

Gerome

Hi,

In addition, could you make some voltage measurements while the PHY is in a reset state? I would like to know what the PHY sees on all strap pins when held in a reset state.

Sincerely,

Gerome

Hi,

Please find the attached waveform of all the voltages + Reset sequence.

Please note that we can read all the reg value of the IC also every time.

-

Mahesh

Hi Gerome,

Please find the waveforms Image below as requested.

Let us know if any further information is needed.

Regards,

G.Shricharan.

Hello,

Can you please send a scope capture of pins 22, 35, 36 as the PHY is held in reset mode? These should be strapped to High, High, Low as these are reflected in your schematic. However, your strap register value indicates that its getting strapped in a different mode.

Sincerely,

Gerome

Sure. I will share it.

As we can read / write DP83869, can not we write reg and put it into required mode irrespective of strapping  ? Please suggest.

Hello,

Yes, you can do either. I was just curious after you had originally designed for this in your schematic why the strap register was not accurately reflecting the intended design. This would eliminate the need to do software writes to set the mode in the first place.

Regarding the software issue, I would like to ask if you can run the procedure outlined in section 9.4.9.1.3 of the DP83869HM datasheet to explicitly write 0x1FFC into register 0x1EC. You can then read back using section 9.4.9.1.4.

Another step I would like to ask is if you are seeing activity on the fiber side and copper sides of the PHY, as this may tell you that the PHY is already in Media Converter Mode. 

This other E2E thread faced a similar issue to what you are facing (albeit different setup), but the issue was determined to be a read/write access issue for extended registers so I would like to see explore this route.

https://e2e.ti.com/support/interface/f/interface-forum/901852/dp83869hm-used-as-media-converter-from-wired-phy-to-sfp---works-only-with-evm-but-not-with-other-external-media-converters/3343975?tisearch=e2e-sitesearch&keymatch=1ffc#3343975

Sincerely,

Gerome

Also, another suggestion I would like to make is to ensure that after each register write, you then set the PHY into software restart (keeping the register values intact). This can be accomplished by writing register 0x1F = 0x4000. So you would write 0x1DF, software restart, 0x0, 0x10, restart, 0x1EC.

In addition, this E2E thread has the information regarding register 0x1EC that you had previously requested:

https://e2e.ti.com/support/interface/f/interface-forum/881224/dp83869hm-dp83869hm-in-1000m-media-convertor-mode-not-working/3270109?tisearch=e2e-sitesearch&keymatch=1EC#3270109

The default value of the register should be 0x1FFC, not 0x0, so I would like you to double check the read steps to ensure you are correctly accessing the extended register.