Do you need both accuracy *and* precision?
Precision will largely be determined by quantization effects as the sample clock must be an integer divisor of the board's master timebase. For example, the 6259 M-series board that I commonly use has an 80 MHz master timebase which can be used to generate the sample clock. In fact, I'm pretty sure it *is* used by default. The nearest integer divisor when requesting 171 kHz will be 468, producing an actual nominal sample rate of 170.9402... kHz. (A divisor of 467 would produce a nominal sample rate of 171.3062...) The quantization doesn't really scale linearly over wide ranges of target sample rate, but for rates in the vicinity of 171 kHz, quantization steps are a bit less than 0.4 kHz, so you can expect to hit your target to within about +/- 0.2 kHz.
Next you need to consider clock accuracy. The 6259 board is rated at 50 parts per million accuracy at nominal temperature of 25 C. I don't know the temperature effect or whether you need to concern yourself with it in your app. But 50 parts per million on a target of 171 kHz amounts to less than 0.01 kHz accuracy error.
So it would appear that the 6259 would be a candidate. There are likely other cheaper M-series boards that could also work. The older E-series had only a 20 MHz timebase, which would lead to quantization error of about 1.5 kHz. Don't recall their accuracy specs, but quantization already puts you over your error budget.
-Kevin P.
ALERT! LabVIEW's subscription-only policy came to an end (finally!). Unfortunately, pricing favors the captured and committed over new adopters -- so tread carefully.
