Histogram of Gaussian White Noise with NI5112

[Julien_BERNU]

I am digitizing some random gaussian white noise with a NI-PXI 5112. The problem is that even with very high number of samples (1 Mega sample) the distribution presents a lot of periodic pikes when making a voltage histogram (either with built in function or by postprocessing data with Mathematica).

I am using two synchronized digitizers and they both have the same behaviour (pikes symmetric around 0 and with periodicity), but give different response from one another.

I already tried to filter the data, change the vertical range of both aquisition and generation devices, but the pikes always appear at the same relative values, and with the same relative number of counts.

How can I overcome this problem ?

[LocalDSP]

Could you post a screenshot of your histogram? How many bins do you have in your histogram?
Remember that the 5112 is an 8-bit boards so you only have (in the order of) 256 possible values so unless you 'hit' the exact values your bins will have un-equal weights. This could result in the periodic behavior you are observing.