TCA9517A: ViLc minimum value (difference with TCA9517:null)

Hi Muk-san,

There are a couple different things happening in these E2E hopefully I can clarify all of them. In general the VILC value we give on the datasheet is a typical value that can change from device to device. It gives an overall average value for all the variation of silicon from device to device. The reason the TCA9517A has a higher VILC with seemingly the same values as the TCA9517 is because with updated processes we were able to manufacture a more reliable silicon die that on average operated at a higher VILC. 

Now as Bobby said, in the second E2E post you mentioned, the value for VILC had very little variation for every device we tested. I went through the characterization data and across multiple lots they all preformed at the typical value with a maximum variation of 0.02V.

Theoretically you could say that the worst case scenario for VILC of the TCA9517A could be 0.38V. However this is extremely unlikely and almost always you are going to get a device that has a VILC of 0.45V

Best,

Chris

Hi Chris-san,

Thank you for your answer.

Vol - Vilc = 70mV (typ)

It is max value, right?

(from https://e2e.ti.com/support/interface/f/interface-forum/552807/tca9517---guaranteed-spec-for-vil-max-without-bus-contention----vilc/)

Theoretically, What factors does effect to this value? what kind of parameters does is depend on? (for example, temp, current, etc)

I2C spec (the spec of external device connected to B-side) is Vol=0.4V(max)

if Vol - Vilc can be limited lower than 50mV, customer can use TCA9517.

But, if Vol - Vilc is 70mV(max), they cannot use TCA9517, because it have a risk, from your following comment.

"Theoretically you could say that the worst case scenario for VILC of the TCA9517A could be 0.38V"

Thanks

Muk

Hi Muk-san,

The VOL-VILC=70 mV is a maximum value. From -40C to 80C the VILC does not change. This is a value guaranteed by the design of the device so it won't change unless the VOL of the device changed.

Once again I would like to reiterate that the chance of having 0.38V VILC is extremely unlikely. We ran tests over multiple lots of devices and the lowest VILC we got was 0.43V and even that was unlikely. You can be confident that the VILC of the TCA9517 will be around 0.45 with a variation of 0.02V. If you would still like me to look for another device let me know.

Also how many TCA9517s is the customer looking to use?

Best,

Chris

Thanks Chris-san

Could you please let me know the Test condition for VILC? (included external components)

Thanks

Muk

Muk-san,

I'll look to see if I can find a test setup for you.

Best,

Chris

Hi Muk-san,

We don't have a test set up for this measurement however I talked to my team and they told me how this measurement would potentially be done. Essentially you would just force a voltage onto the B-side and watch the A-side oscillate. Then you would step the B-side voltage down until the oscillations stop and the A side is constantly held low. This is the VILC.

The other approach is to set a 30% VCCA voltage on the A side with a weak current compliance. Step the B side down until the A side gets pulled from 30% to 0V. That would also be VILC.

Best,

Chris

Thanks Chris-san,

I would like to confirm about your second line comment.

"The other approach is to set a 30% VCCA voltage on the A side with a weak current compliance. Step the B side down until the A side gets pulled from 30% to 0V. That would also be VILC."

Could you please let me know more detail of this measurement method?

Is it meaning, I/O pin in Aside should set 30% x VccA, correct? What is meaning "30%" value? 

Thanks

Muk

Muk-san,

Lets say your VCCA = 3.3V. You input a voltage on the A side SDAA pin of 30% of 3.3V. This would be 0.3*3.3= 0.99V. So you would apply 0.99V to SDAA. Then on the B side lets say your VCCB = 5V. You would apply 5V to the B side SDAB pin and slowly step down your voltage until your A side SDAA pin is pulled from 30% VCCA to 0. In this test you have to make sure that you still have pull up resistors and you are still applying a voltage to your supply pins.

Let me know if you have any further questions.

Best,

Chris