Using I2C0 bus of AM3358 for HDMI DDC Bus...

James Jolliffe

Other Parts Discussed in Thread: AM3358, TPD12S016

I am having difficulty reading HDMI monitor EDID using read-edid v3.02 in linux. Command line prompt response from ARM is No EDID on Bus 0 & No EDID on Bus1... I am using I2C0 Bus of AM3358 which has 3.3K PUs connected to HDMI connector 'DDC bus' via a TPD12S016, TI HDMI companion chip with I2C level shifting. from 3.3V to 5V. The I2C bus going into TPD12S016 without HDMI connection to monitor looks okay but as soon as I hook HDMI cable from monitor to board the ACK line is not able to drive completely high. These are only devices connected in I2C0 bus path from ARM!!!

Thanks,

James

over 10 years ago

0 Biser Gatchev-XID over 10 years ago

TI__Guru**** 393215 points

Hi James,

James Jolliffe said:
as soon as I hook HDMI cable from monitor to board the ACK line is not able to drive completely high.

On which side of the level shifter does this happen?

0 James Jolliffe over 10 years ago in reply to Biser Gatchev-XID

Intellectual 960 points

It appears to be happening on both sides of TPD12S016 I2C level shifting.

James

0 Biser Gatchev-XID over 10 years ago in reply to James Jolliffe

TI__Guru**** 393215 points

Do you have pullups on both sides? What are the values?

0 James Jolliffe over 10 years ago in reply to Biser Gatchev-XID

Intellectual 960 points

Biser,

I have 3.3K PU on AM3358 side. TTPD12S016 has integrated 10K PUs on 3.3V A side (ARM side) and 1.75K PU on 5V HDMI connector side.

James

0 Biser Gatchev-XID over 10 years ago in reply to James Jolliffe

TI__Guru**** 393215 points

I think these 1.75k PU resistors are what causes the issue. As far as I remember these are per HDMI specification, however a few years ago I had a similar issue with a display that wasn't able to pull ACK low enough with these values. In my case these were external resistors so I was able to change them to 4.7k and the DDC started working. I can't give you advice what to do in your case. Maybe you should try with several different displays (in my case it was an older display that was causing the problem).

0 James Jolliffe over 10 years ago in reply to Biser Gatchev-XID

Intellectual 960 points

HI Biser,

Could you possibly move this to the BU forum that supports the TPD12S016 device so they could possibly comment. I assume they have done testing with this device and various HDMI/DVI monitors. The 1.75K PU resistors are integrated into their device.

Thanks,

James

0 Guy Yater over 10 years ago in reply to James Jolliffe

TI__Genius 12190 points

Hi James,

Removing the 3.3K PUs should solve the issue. The TPD12S016 10k PU are in parallel with them and the device is not strong enough to pull the 3.3k||10k down. Pleas try that and let me know if it is solved.

Regards,

0 James Jolliffe over 10 years ago in reply to Guy Yater

Intellectual 960 points

Guy,

With the 3.3K PUs removed, I am now able to successfully read the monitor EDID.

Thanks,

James

0 BHARATH KUMAR NAGAVELLY over 9 years ago in reply to James Jolliffe

Prodigy 10 points

hi
we are using TPD12S016 for imxq .i can able to read EDID information. it is showing in linux is "same EDID"
Why am i getting this kind of massages in linux?

Processors

Processors forum

Using I2C0 bus of AM3358 for HDMI DDC Bus...