I’m using a TOUTPI2 (PR2-3 Rev. E) I2C Shield on a Raspberry Pi 3B. The 5V line measures 5.06V while the SDA and SCL output lines read 3.7V with nothing connected on the I2C bus. These voltages can be pulled higher with the pull-up resistors from other boards (discovered before isolated the problem to the shield), so the bus seems to be getting pulled to the 3.3V rail somewhere (through the level-shifter).
I’ve noticed others discussing similar problems here and here on an older version of the board.
Communication with I2C devices works, but I’d rather not ignore obvious warning signs. Am I correct in thinking that the level-shifter IC chip may be faulty?
I was abnormally curious regarding the voltage drop and so did some more investigating. Looking at the (recently revised) datasheet for the chip and the shield design, it seems Vref2 does not connect to the 5V through the 200k like the EN pin does, which the datasheet warns could cause excessive current drain (to the Vcc1 rail?).
I tried verifying this by cutting Vref2 from the 5V rail and shorting it to the EN pin, but, alas, working with small electronics is not a strength of mine, and I wasn’t able to successfully complete the operation.
Could this be the problem, and if so, is there a newer revision of the board that has resolved the issue?
Hi,
both Vref2 and EN will connect to the 5V through a 200K resistor.
This change was made in rev.F board.
Older version had 5V connected directly to Vref and that’s what the PCA9306 datasheet recommended at that time. We have over 10,000 of these units in field like that.