Custom EMF probe frequency response calibration

Post any questions you may have about our current range of oscilloscopes
Post Reply
Darko8
Newbie
Posts: 0
Joined: Wed Mar 14, 2012 1:36 pm

Custom EMF probe frequency response calibration

Post by Darko8 »

Hi,

How would be possible to create custom probe in picoscope software with frequency response compensation to show corrected frequency spectrum with given EMF probe (for which I have measured frequency response graph)?

That would be really helpfull because now it is hard to judge strength of the noise I measure, because of attenuation of lower frequencies by 20 to 40dB.

Regards, Darko

p.s. Attached is frequency response of one of the EMF probes I have and example of measured spectrum using CH-A:coaxial cable, CH-B: radiated EMF using EMF probe
Attachments
Blue conducted EMF - RED radiated EMF with PBSH3 probe.png
PBSH3 25mm probe frequency response.png

Gerry
PICO STAFF
PICO STAFF
Posts: 1145
Joined: Mon Aug 11, 2014 11:14 am

Re: Custom EMF probe frequency response calibration

Post by Gerry »

Hi Darko,

You can use the low pass filter in the custom probe to roll-off from 1Hz (as you need a low pas roll-off of 20dB/decade and the filter in the custom probe is a 1-pole low pass filter). However the custom probe wouldn't be able to give you a resonant peak to counteract the notch in the frequency response at 140-150MHz, and we have no other tool, in PicoScope 6, capable of that either.
Regards,

Gerry
Gerry
Technical Specialist

Darko8
Newbie
Posts: 0
Joined: Wed Mar 14, 2012 1:36 pm

Re: Custom EMF probe frequency response calibration

Post by Darko8 »

Thank you Gerry,

Idea of using LP filter is good enough for me (for now), I only need probe in ~100MHz range.

Dont you think it would be worth to consider adding such 'calibration or compensation' feature to picoscope spectrum analyser or even as arbitrary shape time domain filter?

For spectrum view it could be only spectrum 'bin' length lookup table which values could be added to current spectrum display. Solution for time domain would be more CPU demanding because there would have to be more complex solution using filter, but it would be solution for both domains.

It would be useful to me even if it would be available as offline processing only, but now I also depend on long time spectrum averaging to remove white noise from spectrum, so I cant capture time domain signal long enough to simply process it outside of picoscope and get dynamic range/resolution I need.

Best regards, Darko

Gerry
PICO STAFF
PICO STAFF
Posts: 1145
Joined: Mon Aug 11, 2014 11:14 am

Re: Custom EMF probe frequency response calibration

Post by Gerry »

Hi Darko,

Sorry for the delay in answering.

To answer your first question, tools that help you to perform better measurements, such as reducing the impact of the unwanted signals so that the signals of interest can be measured more precisely and/or accurately are a welcome addition. Tools that allow you to just shape the data without any obvious measurement benefit are not really part of a Test and Measurement analysis Toolkit. Instead they belong in a Signal Processing Toolkit, where you can change the data for non measurement reasons.

So, in our Spectrum Mode we have Waveform Averaging which allows you to reduce both the variance of the random, stochastic types of signal (e.g. noise, i.e. increasing the measurable range of signal to noise) and the variance of the signals of interest (increasing signal accuracy), which is clearly a measurement tool.
In our Scope Mode we have Resolution Enhancement, which applies a moving average filter to the data to reduce the random variation of amplitude due to noise (i.e. increasing signal precision).

A frequency compensation feature in Spectrum Mode, or an Arbitrary shape filter in the Scope Mode is not really going to improve the measurement of captured signal (as it will effectively distort the signal) so it is not really a measurement tool.

You mentioned, such a feature as offline processing being useful to you. You could do that yourself, by exporting the Spectrum data as CSV, importing it into Excel and then applying your lookup table (as a column of values) to the data to remove the notch.

Regards,

Gerry
Gerry
Technical Specialist

Post Reply