Re: Determine the decimation rate (oversampling rate) programmatically

[CodingMonkey]

Hello all,

I am using the NI-DAQmx C API with a cDAQ-9185 controller and various modules with delta-sigma converters.

I can illustrate my problem with the NI-9232 IEPE module: According to the datasheet, the module supports three different "decimation rates" (= oversample rates): 64x, 128x and 256x.

My question: Is it possible to programmatically determine the current decimation rate (oversampling rate)?

As far as I can see, only the value DAQmx_SampClk_TimebaseDiv can be queried, which is the product of 2 times decimation rate times clock divider.

Thanks a lot!

[ZYOng]

Valid Sampling Rates for NI DSA, CM C Series and SC Express Devices

Use DAQmxGetSampClkRate to get the actual sampling rate.

[CodingMonkey]

Thank you, ZYOng, for your answer. Actually I'm not really referring to the sampling rate, but to the oversampling rate of the delta-sigma converter.

The datasheet of the 9232 module describes:

Preliminary solution I implemented yesterday was:

double oversamplingRate = 0.;

// first get base and sampling rate
float64 timeBaseRate = 0.;
float64 samplingRate = 0.;

DAQmxGetSampClkTimebaseRate(taskHandle, &timeBaseRate)); // get time base
DAQmxGetSampClkRate(taskHandle, &samplingRate));         // get sampling rate as set by driver

// compute decimation factor
const int decimationFactor = static_cast<int>(timeBaseRate / samplingRate);

// decimationFactor = 2 x decimatinRate x clockDivider, clockDivider is in [1, 26] (cf. 9232 datasheet)
const int cMinClockDivider = 1;
const int cMaxClockDivider = 26;

// possible decimation rates for Ni-9232 oversampling
const std::set<int> sDecimationRates = { 64, 128, 256 }; // cf. 9232 specs

// now find a decimation rate that fulfills samplingRate = baseRate / (2 x decimationRate x clockDivider)
bool decimationRateFound = false;

for (auto it = sDecimationRates.crbegin(), itEnd = sDecimationRates.crend(); it != itEnd; ++it) {
  const int decimationRate = *it;
  const int potentialClockDivider = decimationFactor / 2 / decimationRate;
  if ((potentialClockDivider >= cMinClockDivider) && (potentialClockDivider <= cMaxClockDivider)) {
    decimationRateFound = true;
    oversamplingRate = decimationRate*samplingRate; // now we got it!
    break;
  }
}

However, this solution has two disadvantages:

1. It is valid only for the 9232 module
2. I do not think that the solution is unique.

As for the second point, I choose the highest possible oversampling rate, but I don't know what NI does internally.

So I've been looking for an API function to query the oversampling rate, but I don't think there is one.

[santo_13]

And why is the decimation rate important to the user if is handled by the driver?

[CodingMonkey]

---

@santo_13 wrote:

And why is the decimation rate important to the user if is handled by the driver?

---

I need the decimation rate to determine the input delay:

Cf. https://www.ni.com/docs/de-DE/bundle/ni-9232-specs/page/overview.html:

Cf. also: https://knowledge.ni.com/KnowledgeArticleDetails?id=kA00Z000000P8toSAC&l=de-DE