0
\$\begingroup\$

signal in oscilloscope

I am designing a low pass filter in matlab to remove the noises shown in the image. The original signal of interest lasts only for about 80 ms and rest of the signal are noises from probe tip.

In matlab, i used the following script. d=fdesign.lowpass('Fp,Fst,Ap,Ast',0.001,0.5,60,1);

But I not getting the effect of removing the noisy signal exactly. I dont understand the parameters 'Fp,Fst,Ap,Ast'. I tried playing with those values, But its not working. What values should i give to these parameters to remove the noise exactly and have only the signal of interest ?

Thanks.

\$\endgroup\$
  • \$\begingroup\$ This is probably better suited to dsp.stackexchange.com \$\endgroup\$ – JRE Jul 8 '15 at 14:56
  • 1
    \$\begingroup\$ I suggest you take a look at the documentation for fdesign.lowpass (mathworks.com/help/dsp/ref/fdesign.lowpass.html?nocookie=true) \$\endgroup\$ – JRE Jul 8 '15 at 14:56
  • \$\begingroup\$ Most of the noise resembles powerline hum, so you may need to start by looking into the many questions posed here on improving oscilloscope measurements. \$\endgroup\$ – JRE Jul 8 '15 at 14:59
  • 2
    \$\begingroup\$ As a spike, your signal will be fairly broadband, making it difficult to filter it from the noise. You may be able to use a narrow band filter to remove the powerline hum, and other filters for other identifiable noise sources. \$\endgroup\$ – JRE Jul 8 '15 at 15:01
  • \$\begingroup\$ Hmmm inductive proxity probe signal? \$\endgroup\$ – Andy aka Jul 8 '15 at 17:27
2
\$\begingroup\$

To answer your direct question, the parameters are defined by the following diagram, taken from the Mathworks documentation

diagram

But, as JRE pointed out, your "spike" signal will also be highly distorted by any low-pass filter that has a significant effect on your noise.

This is not a simple problem to solve. If your signal always has the same general form in the time domain, you might consider a "matched filter". You also might want to look at various kinds of nonlinear filters.

\$\endgroup\$
  • \$\begingroup\$ Thanks for the info. The signal always has the same kind of pattern. \$\endgroup\$ – user3396084 Jul 8 '15 at 15:20
2
\$\begingroup\$

It looks like you are still working on this project.
You may want to start with improving your measurements before you go to filtering.

Many oscilloscopes can measure the difference between the two inputs.

So, you could set your scope to difference, then attach one probe to each end of your load resistor. That ought to reduce the powerline hum without using a filter.

Like this:

schematic

simulate this circuit – Schematic created using CircuitLab

With the scope set to subtract probe 2 from probe 1, you should get a much cleaner signal.

\$\endgroup\$
  • \$\begingroup\$ I'm saying you should be using two probes. \$\endgroup\$ – JRE Jul 9 '15 at 12:52
  • \$\begingroup\$ ah okay, where do i connect the two ground leads of the probes then ? \$\endgroup\$ – user3396084 Jul 9 '15 at 13:41
  • \$\begingroup\$ Nowhere. Just leave them unconnected but out of the way. \$\endgroup\$ – JRE Jul 9 '15 at 13:42

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.