TCA9517-Q1: TCA9517-Q1: I2C BUS TRAP

This looks as if some other device on bus A pulls SDA low. You would have to reset that device.

hi Clemens，

I think you want to say" device on BUS A pulls SDA low", right? because there is only Master/MCU omn BUS B side. so I want toaks that why we power off the VDDA, SDA on BUS B side could not dirve high ?

Sorry, I did indeed confuse the sides.

The voltage of 0.5 V at SDAB shows that the TCA9517-Q1 outputs low on the B side because some device on the A side pulls the line low. (A device on the B side would have to pull below 0.5 V to make the TCA9517-Q1 drive low on the A side.)

So one of the A-side devices has a software or hardware error. Try removing devices; but the only way to find out which device sinks currents during operation would be to measure the voltage drop over a series resistor.

hi Clemens,

 so, it means that the I2C BUS was pulled low by slave device, not master side,right? 

BTW, we send command (i2cscan) on B side Master device - MCU, but it could not drive neither the SDA nor the SCL low, so that the I2C bus still tap, do you have any suggestions?

thanks.

hi Clemens,

when abnomal occure, I pulled low then high the EN pin of leveshift, the I2C bus recovered, if device on BUS A drive low the SDA, I don't think it could be recovered in this way, what do you think?

It is hard to say anything specific without knowing which device pulls SDA low.

Hi Ziyang,

Any block diagram or schematic you could share with us? 

Is the EN pin shared among many devices on the I2C bus? Pulling LOW then HIGH on EN pin is affecting only the TCA9517-Q1 or many devices at once? 

Do you have oscilloscope captures of the issue? 

Regards,

Tyler

hi Tyler，

1、NO，EN pin is pull to 3.3V directly, I use fly-line to pull it down to GND and high by removing the fly-line;

2、

on the bus a:

on the bus b: there is delay between the 8th & 9th clock,SCL is drived low by slave device

hiTyler

yes, this is normal operation, not sure why there is abnormal : SCL-B is high while SDA-B is 0.5V, why the slave device pull the SDA low all the time.