Implementing software defined noise generators

EasyChair Preprint no. 998

Abstract

In the last few decades, the research field related to noise and fluctuations became more and more important and the usage of the results is increasingly promising. In addition to simulations, testing and validation of implemented systems should also be performed on real physical systems. These measurements require noise generators with specific distribution and spectral properties. It is obvious, that the most convenient method to generate such noise is by using digital signal processing. The randomness of these sources is provided by pseudo-random number generators, generators that have limited cycle length/time. Of course, signal generators capable of noise generation are commercially accessible, however, these solutions have limitations. In our previous publication, we presented a Digital Signal Processor based solution capable of generating 1/fα shaped noises. As a result of advances in technology these days there are many user-friendly and high-performance solutions that provide us new possibilities in noise generation. To ensure proper period length, first, we implemented different pseudo-random number generators on the FPGA based hardware. Then, we adapted and extended our algorithm to the new hardware, as a result, we could significantly increase the sampling frequency, the available bandwidth, the D/A resolution and the number of channels. The noise generator is developed in the LabVIEW programming environment and can run on any FPGA based NI platforms, but it can be ported to any other FPGA based hardware. The new instrument uses a flexible filter arrangement to combine filters of different frequencies to produce more generic spectra up to five decades. The implemented solution is open-source, thus, anyone can use, customize or make further developments of the design for their personal needs.

Keyphrases: 1/f noise, digital filter, DSP, FPGA, noise generator, software defined instrument

@Booklet{EasyChair:998,
  author = {Robert Mingesz and Dénes Faragó},
  title = {Implementing software defined noise generators},
  howpublished = {EasyChair Preprint no. 998},

  year = {EasyChair, 2019}}