The VI depicted in Sineandsquaresignals.jpg is essentially the same as one of the DAQmx Shipping Examples. LabVIEW's Basic Function Generator actually keeps the phase of the signal in a shift register so on each successive loop iteration the waveform will pick up where it left off from the previous iteration.
If you were to write the waveform once and regenerate it, then you would want to make sure it was a complete period of your signal. However, the example you have linked is configured to generate each sample one time--you must continually rewrite samples to keep the buffer filled (failure to do so would result in an underflow error rather than jitter on the output). The phase of your output waveform is saved from loop to loop so writing a complete period at a time shouldn't be necessary.
This scenario could result in period-to-period jitter if your sample rate was not an integer multiple of your output signal since the waveform would not be exactly the same set of samples each time. However, in your case the sample rate is a multiple of your desired frequencies so you should be getting the same output data from cycle to cycle. In any case, the jitter would always be less than one sample (which is 1.25 us) so wouldn't explain what you are seeing in any case.
I think the apparent time jitter you are seeing (shown in the video) is quite possibly coming from noise on the sine wave output (I notice you are trigging the scope from the sine wave). At ±5V and 80 Hz, the maximum slope of the sine wave would be 800 V/s (which is 1.25 ms/Volt). Assuming you are triggering at 0, each mV of error would result in a a 1.25 us time difference on when the trigger is received. So you're looking at an approximate ±6 mV of deviation in your output to account for the 15 us deviation in the trigger time. I realize this is outside the accuracy spec of the analog output, but that spec assumes that the output has fully settled and I don't believe it accounts for noise. <speculation> It's possible that the reason you see less jitter when running at a slower sample rate has to do with how the output settles--it does not transition linearly from voltage to voltage so you might end up with a steeper slope at the point of the trigger when your waveform has less samples in it </speculation>.
To give a better picture of what is actually going on, I'd suggest triggering the scope from the square wave (which has a very fast rise time) so you will have a relatively constant time reference. The sine wave will have some noise on its output (which is probably normal) but I wouldn't expect any time jitter based on how you are configuring your outputs.
Best Regards,
John Passiak
