Spectrum Analyser Ghost Problem

Post any questions you may have about our current range of oscilloscopes
Post Reply
Arpadpolan
Newbie
Posts: 0
Joined: Thu Apr 01, 2021 1:31 pm

Spectrum Analyser Ghost Problem

Post by Arpadpolan »

My brand new 2205 analyzer shows a 10Khz sinusoidal frequency at near 0dmV at the incorrect frequency when i decrease the scan width under 10 Khz. In other words, frequencies beyond the scan band will show up in full force as phantoms. Something that seems like a gross problem that renders the spectrum view practically unusable. The issue can be reproduced with external or internal source, and at any frequency.

Did anyone else have this issue?

Input signal time domain view:
time domain.jpg
correct spectrum display when input frequency is within scan band:
spectrum at 24khz scan width.jpg
incorrect spectrum display identifying the 10Khz signal as totally different in frequency:
false full magnitude ghost at 6khz scan width .jpg
false full magnitude ghost at 2khz scan width .jpg
another way to reproduce is to increase the input signal freq and watch it as it reaches the end of the spectrum band, and continue increasing and watch it walk back into the band as a phantom.

bennog
Advanced User
Advanced User
Posts: 86
Joined: Mon Nov 26, 2012 9:16 am
Location: Netherlands

Re: Spectrum Analyser Ghost Problem

Post by bennog »

This is a problem with all calculate spectrum analyzers if the input frequency is above the sample frequency.

Tested with Fluke 196C and exactly the same.

Benno

hexamer
Advanced User
Advanced User
Posts: 3
Joined: Tue Aug 12, 2014 10:09 pm

Re: Spectrum Analyser Ghost Problem

Post by hexamer »

Those "ghosts" are to be expected. They are called aliases. As Benno pointed out, your signal frequency is well above the Nyquist frequency of fs/2.

I like to use this tool to show where aliases will occur: https://www.analog.com/en/design-center ... -tool.html

When I plug your parameters into that tool, I get the exact same results you're reporting.

Arpadpolan
Newbie
Posts: 0
Joined: Thu Apr 01, 2021 1:31 pm

Re: Spectrum Analyser Ghost Problem

Post by Arpadpolan »

thank you very much for the explanation, i am lacking the theoretical background to realize this!

Post Reply