min sample rate vs actual sample rate

[Andy Peters]

I'm sure this has an obvious answer, but here goes.

I have a USB-5132 'scope, and using niScope Configure Horizontal Timing.vi I set, among other things, the min sample rate. In my case, I've chosen 20 MHz, which of course gives a sample period of 50 ns.

I use niScope Read (poly).vi with the WDT variant to actually read the waveform data.  I noticed something very odd -- waveform limit tests throwing Error -1802 "waveforms have a different dt values' -- so I put a Get Waveform Components unclusterizer on the waveform wire, and looked at the waveform's dt value. It tells me my dt is 40 ns, which of course is 25 MHz. I am also asking for only 2000 samples.

So what causes this discrepancy? Why doesn't the digitizer simply accept my desired sampling frequency?

[Austin_W]

Andy,

The Minimum Sample Rate parameter is the frequency at which digitized samples are stored, expressed in samples per second. This parameter is rounded up to the next legal sampling rate that your device supports.  The Actual Sample Rate property returns the actual sample rate used for the acquisition.

Unless you specify another clock source, the digitizer is using an internal oscillator as its clock source.  For the 5132, this oscillator is running at 50 Mhz.  When using the internal oscillator as the sample clock source, the digitizer can only use divided down versions of this clock, so some sample rates are not possible.

If your application requires sampling at specific intervals that cannot be achieved by using the internal 50 MHz clock, the NI 5132 can accept an external sample clock. 

[Andy Peters · Accepted solution]

---

Austin Walton wrote:

Andy,

The Minimum Sample Rate parameter is the frequency at which digitized samples are stored, expressed in samples per second. This parameter is rounded up to the next legal sampling rate that your device supports.  The Actual Sample Rate property returns the actual sample rate used for the acquisition.

Unless you specify another clock source, the digitizer is using an internal oscillator as its clock source.  For the 5132, this oscillator is running at 50 Mhz.  When using the internal oscillator as the sample clock source, the digitizer can only use divided down versions of this clock, so some sample rates are not possible.

---

So is there a list of the supported actual sampling rates? Or is it simple power-of-two division, or the usual 1-2-5 sequence?

Perhaps a better way for NI-SCOPE to work is to make the sample-rate control an enumeration that forces only legal values.

---

 

If your application requires sampling at specific intervals that cannot be achieved by using the internal 50 MHz clock, the NI 5132 can accept an external sample clock. 

---

Actually, in my specific case, the sample frequency is somewhat arbitrary, and I actually left it at the default (20 MHz) which is reasonable for my application. I am trying to do limit tests and the limit-compare VI was complaining that the live waveform sample rate did not match my masks' dt parameter, which I thought was odd because I created the masks using the Express VI. So in other words the masks have a 50 ns period and the live waveform has a 40 ns period.

So anyways, now I know how to fix it -- regenerate the masks with a waveform sampled at 25 MHz and not 20 MHz.

Thanks,

-a