Hello,
The PCA9512A datasheet includes the following note:
(1) Can someone confirm that the pull-up resistor voltage need not match the supply voltage without enabling the acceleration function? Specifically, is there a hazard in using 2.5V pull-ups with a PCA9512A device powered by 3.3V?
My application has an 2.5V I2C master (whose input may not exceed 3.05V (VDDIO + 0.550V). The PCA9512 has a VCC spec of 2.7V minimum. I experimented with a PCA9515 legally powered at 2.5V, but there's an fixed-offset component (a PCA9615B) downstream that can't propagate the bus. I could create a tiny 2.7V or 3.0V supply; but would prefer not, unless absolutely necessary. Assuming the scenario works as I'd like-- I'd have a pair of pull-up resistors to 2.5V on the primary side of the PCA9512A with VCC = 3.3V; while the secondary side has, in two instances 3.3V pullups with VCC2=3.3V, and in a third, 5V pull-ups with VCC2 = 5V.
(2) Of lesser importance, I'd like to understand what the hazard is in enabling the accelerator in this scenario. Has no bearing on my application; just intellectual curiosity.
Thanks very much in advance!
Heh... I just discovered the PCA9306... Better solution?
Please follow datasheet description about this topic application,otherwise can't promise!
Thanks for your response, G.W.
I ultimately had to pull the trigger on the board respin for schedule reasons. I had used a Pericom (Diodes Inc) PI4ULS5V202 level shifter elsewhere in the design for an MDIO bus (in that case 1.8V to 2.5V) with great success. So I bugged that part in here for this application and it's working reliably, and is datasheet compliant. So I'm up and running.
So ultimately I have three I2C bus translators, two still the PCA9515B for branches that do not cascade static offset devices-- both powered at 2.5V, translating from 2.5V to 5V, and from 2.5V to 3.3V, respectively; and the PI4ULS5V202 powered at 2.5V translating 2.5V to 3.3V.
(I'm still interested in understanding from an intellectual curiosity standpoint, though.)