LSF0108: I2C Signal Integrity with LSF0108

Manoj Chate

Part Number: LSF0108

Hi team,

I'm using I2C from carrier board to add on board with buffer P82B715DR on add-on board. Further to this level translator LSF0108 is used to use I2C at 3.3V.

Please see below SDA waveform with above described system (i.e. Carrier <__> Buffer <__> Level Translator ) :

After observing the SDA line waveform , I disconnected LSF0108 from I2C line by removing series 0 ohms. So the system was like :

(i.e. Carrier <__> Buffer ) . Now SDA line waveform looks like -

Can you please suggest what is causing issue is SDA line waveform with LSF0108 connected.

Also ,why after disconnecting LSF0108 the wave forms got clear. Please suggest.

Thanks,

mANOJ

over 2 years ago

0 Clemens Ladisch over 2 years ago

Guru 243110 points

The LSF is a passive switch, it cannot drive the signal high. I suspect that there is another device behind the translator responsible for that. Please show the entire schematic.

0 Manoj Chate over 2 years ago in reply to Clemens Ladisch

Prodigy 180 points

Hi Clemens,

There is no device behind translator , as the device is not connected and just I2C pull-ups are there.

Here is the schematic attached.

Please make note , in schematic J4 is not mounted. Concerned I2C bus is I2C1_SDA & I2C1_SDA_CARRIER.

6888.schematic.pdf

Thanks,

Manoj

0 Clemens Ladisch over 2 years ago in reply to Manoj Chate

Guru 243110 points

The pull-up resistors R27 and R32 behind the translator are extremely strong, and might prevent other devices from pulling the line low enough. Try 10 kΩ instead, and go lower if the rising edges become too slow.

0 Michael J Schultis over 2 years ago in reply to Clemens Ladisch

TI__Mastermind 24485 points

Hello,

I agree that if the pullups are too strong then other devices that try to pull the line low may have trouble pulling it low. The LSF device itself will not pull the line low as it is a switch based architecture.

For more detail in understanding translation with the LSF Family, I recommend you review the following video series:

https://training.ti.com/translation-basics-lsf-family

https://training.ti.com/translation-lsf-family-0

Also, from your schematic I also see SPI and UART translation.

The LSF devices are capable for supporting these interfaces, but you might also consider translation devices optimized for these applications, such as the following, which have fixed direction channels optimized for these interfaces and buffered outputs for optimal signal integrity.

TXU0204 - UART interface

TXU0304 - SPI interface

Best,
Michael

0 Manoj Chate over 2 years ago in reply to Michael J Schultis

Prodigy 180 points

Hello Clemens and Michael,

I tried making I2C pull-up as weak pull-up (10k) on the other end of LSF0108 i.e R27,R32 in schematic.

Would like to inform that , this lowered the Vol to 0V from earlier 308mV. But , in this case as well the glitch is there, which is shown below in image as highlighted. Do we have any reason or answer for this ?

Also , why this behavior is observed only for SDA line and not for both SDA and SCL lines , as strong pull-up (330 ohm) was same for SCL and SDA on other end of LSF0108 ?

Thanks.

Manoj

+1 Dylan Hubbard over 2 years ago in reply to Manoj Chate

TI__Mastermind 42600 points

Hey Manoj,

Can you zoom in on that? It looks quite negligible in the current view.

Logic

Logic forum

LSF0108: I2C Signal Integrity with LSF0108