andreas chwala

Hi all,

I have recorded noise data with my PCMCIA card, self made anti alias low
pass (first order only) and amplifier (50). Im measuring on all 4 recorded
channels the same strange signal (the raw data look as expected): for some
frequencies it looks like a whole block has simply added an offset in the
frequency domain with many averages: http://www.chwala.de/fft/99av.tif

Diving deeper, I found the the following picture with only three averages:
http://www.chwala.de/fft/3av.tif

The real part of the fft over the whole time series looks like this:
http://www.chwala.de/fft/real(fft).tif

(All done using matlab.)

Any suggestion what this could mean, how this can be produced?

Regards,

Andreas

--
Andreas Chwala
Email: andreas{bei}chwala{punkt}de

rusty

The card is picking up some periodic digital noise from somewhere I would
guess. Perhaps PSU noise or clock noise

Some more info would be good
Is the horizontal scale 0 to 10KHz or 10 MHz.

rusty

andreas chwala

Sorry for not labelling the axis - horizontal is 0kHz to 10kHz. (Interesting
note: I have observed a similar behaviour with a totally different 24bit AD
now also.)

Andreas

andreas chwala

Hi all,

I have just found that the histogram of the sampled values is also a bit
strange, not Gaussian at all! Have a look:
http://jetzweb.de/chwala/fft/hist.tif
Maybe that explains a little bit more?

Regards,

Andreas