Poor bandpass filter :-(

Forum for discussing PicoScope version 6 (non-automotive version)
Post Reply
Knud Tangsgaard
Newbie
Posts: 0
Joined: Sun Aug 16, 2015 7:15 am

Poor bandpass filter :-(

Post by Knud Tangsgaard » Sun Dec 30, 2018 1:29 pm

Hi,

Do only find very small attenuation of signals outside specified frequency range of built in bandpass filter :(

Setup:
PicoScope 4824
AWG generate sine signal, connected direct to Channel A
Math channel: BandPass(A;39000;41000)

Sample rate= 80MS/s, gives min. cut off frequency of: 1,25kHz

Amplitude of math channel at different frequency:
20kHz= 792mV
30kHz= 757mV
40kHz= 711mV
60kHz= 594mV
80kHz= 459mV

Questions:
1) Why does the bandpass filter only have very small impact?
2) What is the maximum Q-factor for bandpass filter?

Best regards
Knud Tangsgaard
Attachments
BadPass-40kHz.pssettings
(3.95 KiB) Downloaded 50 times

lab!fyi
Newbie
Posts: 0
Joined: Mon Dec 12, 2016 6:58 pm
Contact:

Re: Poor bandpass filter :-(

Post by lab!fyi » Mon Jan 07, 2019 8:26 pm

If this is of any help I sometimes managed to get little better results by nesting filters with same params eg this for 20MHz: BandPass(BandPass(A,19999990,20000010),19999990,20000010)

Gerry
PICO STAFF
PICO STAFF
Posts: 368
Joined: Mon Aug 11, 2014 11:14 am

Re: Poor bandpass filter :-(

Post by Gerry » Mon Jan 21, 2019 6:56 pm

Hi Knud,

The Maths Channel bandpass filter is effectively a single pole filter, so you only get a gentle roll off of between 3dB and 6dB/octave (there is no resonance). So, there will be a practical limit to how small you can set the bandwidth of the bandpass filter, before the filtering effects start cancelling each other out, and the signal level starts to reduce.

You may be able to use a trick or two to make the slope a bit steeper, but ultimately the Math Channels don't provide a full digital filtering toolkit for amplifying narrow bandwidth signals, with surgical precision. However, they do provide you with, a very effective means to clean up and condition the signal for analysis, when complimented by the other software tools at your disposal, and then give you the capability to transform the signal in a variety of different ways, some of which even make ingenious use of unintentional ways that the Math Channels are implemented (as can be seen from some of the posters on our forum).

Regards,

Gerry
Gerry
Technical Specialist

Post Reply