I2C acknowledgement problem and resistor pull-up

[riyazm]

We are using a buffer for I2C interface signal's voltage level translation. Here our I2C host is OMAP3530 which is of 1.8V signal level and device is TVP5150AM1(video ADC) which is 3.3V signal level.So we are using TXS0104E for voltage translation which also supports I2C signals.
When we try to access the TVP5150AM1 from OMAP3530, we are not getting the acknowledgement from the processor side.
The internal pull-up of TXS0104E (at 3.3 V side) is 10K and the I2C lines of video ADC is 2K2.
what is the equivalent pull-up resistor we can put.
please comment on this....

Regards
Riyaz

 

[cks3976]

First of all please check on the OMAP3530 side if the SCL is pulsating when you are attempting a DAC read. If that's correct, check the same line near DAC. if you are getting clock on both sides check the SDA. If either is correct, check the I2C driver you have implemented in OMAP3530 . Another thing- 2K2 seems to be a bit strong pull up to me. 10K or 13K on either sides of the voltage translator would be my choice. If you miss either SCL or SDA, check the input and output pins of voltage translator.

 

[riyazm]

First of all please check on the OMAP3530 side if the SCL is pulsating when you are attempting a DAC read. If that's correct, check the same line near DAC. if you are getting clock on both sides check the SDA. If either is correct, check the I2C driver you have implemented in OMAP3530 . Another thing- 2K2 seems to be a bit strong pull up to me. 10K or 13K on either sides of the voltage translator would be my choice. If you miss either SCL or SDA, check the input and output pins of voltage translator.

hi cks3976
we probed signals at the i/p and o/p of the buffer, the problem we are facing is that the acknowledgement bit is not going to low. the waveforms captured is shown below.

i/p of buffer:

o/p of buffer:


The pull-up resistor at the processor side(i.e., at i/p of buffer) is fixed to 4K7,hence we are not any making changes.
The changes we want to make is at the I2C lines of Video ADC(i.e., at o/p of buffer) which is pulled to 2K2.
The Voltage Translation Buffer IC has an internal pull-up resistor of 10K.
Please,suggest for a equivalent pull-up resistor value for this.....

Regards
Riyaz

 

[zuisti]

In my humble opinion - seeing the waveforms - it's not HW but an SW problem (at the proc side).
The SDA line seems to work well, just not at the end of the communication.

Do you are using I2C library routines? Try to write your own routines (not a big deal), particularly
in the precise handling of the ACK and NACK bites.

 

[cks3976]

Hmm.. what is that big glitch on SDA at the end of communications ? That does not look exactly;y like a Hardware issue to me. Some acknowledgment handling is at the PROC side seems to have taken a hit. As Zuisti have suggested, please check it again. Especially in the last 5 clock cycles the data being read is 00010. Is it correct ? or if that was to be 00000 ? Please check

 

[zeroskj1988]

i am not able to use i2c communucation until i pull up both SDA and SCL line to vcc using resistor less than 10k

 

[cks3976]

i am not able to use i2c communucation until i pull up both SDA and SCL line to vcc using resistor less than 10k

Hi, Pull-up resistors between 1K5 to 68K can be used on I2C bus. the value of resistor depends on how long is the trace length ot I2C bus and what is the total load put on the bus.

I thought of sharing this wonderful article on effect of Pull-up on I2C bus performance. Please check the below link :
**broken link removed**