NI-9202 - Delta-Sigma External Clock - Is this possible?

[HFC-DHall]

I am playing around with a NI-9202 in a cDAQ-9174 chassis (Also have a NI-9401 in the chassis). We have an application where we need to continuously take 2-3 simultaneous samples based off an external clock from a digital input signal (PFI) to make sure we sample in time with a 3rd party device also collecting data. Our rate is on the order of 100S/sec. The application is being developed in C#. While playing around with it, I kept getting the -200414 Error (Requested sample clock source is invalid). To simplify testing I switched over to Measurement & Automation Explorer and setup a NI-DAQmx Task there. I get the same error if I set the sample clock to "External" for anything but the "OnboardClock" option.

I believe I am getting this error because the module is a Delta-Sigma style that uses an internal oversample clock technique.. Because of this, only the module's on-board clock can be used for this style of AI sampling. Is this a correct statement?

I've searched the manuals, this site, and the internet in general, but cannot get a definitely answer to this question, so I thought I would post here.

[ZYOng]

What did you select as the Master Timebase Source?

I am not 100% sure, but based on NI-9202 Specifications, the module only accepts an external timebase of 13.1072 MHz

[Kevin_Price]

Don't know the 9202 in particular, but Delta-Sigma converters in general don't let you supply a direct external sample clock.  At least *some* of them allow you to supply an external timebase that it will then divide down to create its own sample clock, but usually there's a fairly small range of frequencies allowed for the timebase.  That's unlikely to help you here.

I'd look to do one of the following:

1. Let the 9202 export its sample clock out to the external device and set up the external device to sample on what is, to it, an external clock.  But I realize that might not be feasible.

2. Run your 9202 at ~10x the sample rate of the external device and capture the external device's clock signal with an unused 9202 channel.  Then you can post-process to keep only the data where the clock was seen to assert.   A nice side effect is that this approach automatically compensates for the filter delays inherent to Delta-Sigma converters.  Your real signals and the clock are both delayed by the same amount, so you get proper correlation without further mucking about.

-Kevin P