ADS131M08: Input impedance of ADS131M08`s REFIN pin

Hello Sunny,

Can we use resistor divider to offer reference for ADS131M08, like 2.5V reference + two1kohm serial resistor to get a 1.25V reference?

We're making an assumption here that the customer has 2.5V reference already and needs a 1.25V input for the ADC, which is why they want to use a resistor divider instead of another active component to provide that voltage. If this is the case, I would like to recommend the REF2025 which has a bias pin that automatically halves the 2.5V out to 1.25V which would solve this issue without the need of resistors. I'm also sure there's a few devices out there that have the same feature.

Two 1kohm on the input might work but you'll need to test it on your end and agree to the error that will be introduced.

The reference is a low impedance output and does expect a low impedance node. Failure to do so will impact the performance of the ADC. By how much? We'll need the input impedance for that. But before going there, please tell us about the tradeoffs you are looking for. High performance? Low power? Low cost? Low board space? These are the things I can tell based on what questions you're asking. I will explain why this is important in the next answer.

From my perspective, swapping out the reference to the REF2025 solves the problem more quickly without needing the reference input impedance but it might not be an acceptable tradeoff.

Could you help to share the input impedance of ADS131M08`s REFIN pin@32kHz data rate? If that works, what`s the maximum output resistance can ADS131M08 accept? and does this decrease performance of ADS131M08?

In general, the reference is sampled along with the input on every clock cycle so the equation mentioned in the datasheet is the same

To abstract it a bit more it would be Z_in = R_equivalent * 4.096MHz / f_mod

However, the equivalent resistance used in the inputs is not the same as the reference so 330-kohm cannot  be used. As you can tell, this number is not specified in the datasheet.

So, if we agree to go down this path, we need help from the design team to get this answer and it may take weeks to months to get an answer. If you want an answer more quickly, we would need to change the question so that it focuses on the tradeoff that you want. So the question might change to "what magnitude of error will I get specifically using the 1-kohm resistor divider?" which is a more narrow investigation that might not need design's help.

Thoughts?

Best,

-Cole