1
\$\begingroup\$

Below time series offset value of a force transducer amplifier which is sampled by a 16-bit single ended earth grounded data-acquisition board at 500Hz (red plot) and 12kHz (green plot). Before I was having aliasing noise so I implemented 200Hz LP filter using 1u cap and a 1k resistor at the output of amplifiers. The amplitude of the noise decreased on scope test. But the noise level still changes with different sampling rates. So below is the time series of the same output at 500Hz and 12kHz sampling rates:

enter image description here

I couldn't figure out whether it is still due to aliasing or SNR being related to sampling rate regardless of aliasing. As you see below for the FFT of the same output, there is like between 12dB to 15dB difference between 500Hz and 12kHz sampling rates:

enter image description here

So now when I sample with 12kHz it seems I get better SNR.

In my previous question someone answered:

Your signal adds coherently, giving you 6dB of gain for doubling the sample rate. Your noise adds incoherently, giving your 3db for the same doubling. You just got 3dB SNR improvement for 2X the samples. Of course you have to sample at 4X to get 6dB improvement, and 8X for 9dB, so there's practical limitations.

I have couple of questions:

  1. How can I verify this dB difference is due to aliasing or just the result of another phenomenon?

  2. Regarding suppressing common-mode 50Hz and its harmonics, unfortunately I cannot use differential ended receiver and I'm also using coaxial cables. If this is still due to aliasing, can I still use common mode chokes to suppress high freq. noise; or will it be useless since the system is not balanced and not diff ended?

\$\endgroup\$
  • \$\begingroup\$ Were you using different signal conditioning (for example, anti-aliasing filters) for the two cases? \$\endgroup\$ – The Photon Jul 16 '18 at 18:35
  • \$\begingroup\$ The conditioning for the plots I presented use exactly the same filters for same outputs. They are taken consecutively as offsets. I just removed the mean values in Python to see the functions clearly. \$\endgroup\$ – HelpMee Jul 16 '18 at 18:37
  • \$\begingroup\$ The only difference is sampling rate in those plots. \$\endgroup\$ – HelpMee Jul 16 '18 at 18:38
  • \$\begingroup\$ Looks like aliasing with PS ripple . Check FFT on DC using AC coupled to 50 Ohm term. What fundamental signal and expected SNR? \$\endgroup\$ – Sunnyskyguy EE75 Jul 16 '18 at 19:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.