TCA9517: Recommended pull-up resistor value in HDMI application

Hi Clemens,

Thank your for your comment. But they said that this issue was cleared by applying 47kohm pull-up resistor on I2C bus. Is there any problems on 47kohm pull-up resistor? 

Regards,

Did they replace the 10 kΩ with 47 kΩ, or add another resistor?

Hi Clemens,

Yes, they replaced 10kohm with 47kohm, and didn't add any additional resistor. Please tell me your opinion.

Apparently, the problem is that some device on the bus is too weak to sink the current through the pull-up without V_OL becoming too high. In that case, you indeed have to use weaker pull-up resistors.

What exactly is that ACK issue? Can you show an oscilloscope trace?

Hi Jeffrey,

Can you provide the scopeshots of both SDA/SCL on both sides (A and B) of the TCA9517 device?

-Bobby

Hi Clemens,

Please check an attached scope shot of both SDA and SCL. In case of 10kohm(A side) and 47kohm(B side), the frequency count of this issue is lower than before, but there is still an abnormal ACK issue depending on the laptops and monitors of HDMI input and output.

Scop shot of both SDA and SCL.pdf

Regards,

Hi Jeffrey,

From the looks of the wavefrom from SDAB to SDAA on the 9th clock cycle, the signal looks like it ACKs on B side but it isn't seen on A side. I think this may be due to the driver on B which (not the 9517) which has a weaker pull down driver and can't drive the signal below ViLc of TCA9517. When you use the weaker 47k pull up resistor, the current IoL is lower so the VoL on B side is able to drive below the TCA9517's ViLc thus allowing B side of TCA9517 to drive A side low from B side. This seems to confirm what Clemen's was stating with his "the problem is that some device on the bus is too weak to sink the current through the pull-up without V_OL becoming too high"

-Bobby

Hi Bobby,

Thank you for your response. They are trying to re-verify this issue in reference to your comment. I will let you know if completing their verification.

Regards,

Sure thing Jeffrey,

You should be able to verify this by looking at the VoL on the B side of TCA9517 and seeing if its larger than ViLc spec'd in the datasheet on the ACK on B side. 

-Bobby