Author Topic: Testing a low pass filter with a spectrum analyzer with no tracking generator?  (Read 2249 times)

0 Members and 1 Guest are viewing this topic.

Offline Chris Wilson

  • Frequent Contributor
  • **
  • Posts: 392
  • Country: gb
  • Race car engineer, dog lover, hoarder.
I am building a Softrock Ensemble RXTX SDR transceiver from a kit. It's nearly done and part of the build instructions are to make a low pass filter. The RXTX can be built for one of four pre dined set of bands and I am building mine for 30, 20 and 17 meter operation. For those bands the filter in the attachment is to be built. I have a spectrum analyzer available, but no tracking generator. can I use the incrementing output of my signal generator, which goes 80KHz to 1040MHz to feed the filter input, and somehow follow it with my spectrum analyzer? I don't doubt the instructions, I am just curious to learn and actually "see" it working.  Thanks.
Best regards,

                 Chris Wilson.

Offline jahonen

  • Frequent Contributor
  • **
  • Posts: 983
  • Country: fi
You could try putting spectrum analyzer to "max hold" mode (I assume that it is not a very old analog one), and generator to sweep in frequency range of interest. After a little while, you should get a pretty good result. If your generator does not have sweep function, you could try manually wiggling the frequency around, but that is of course quite tedious approach.

Regards,
Janne

Offline olsenn

  • Frequent Contributor
  • **
  • Posts: 994
I imagine that +1GHz signal generator is quite nice, but it's frequency response is not flat. Without a signal generator, obtaining the frequency response of your filter WON't be quick and painless, but it IS possible. At its most basic implementation, you could just put the signal generator's output to your DUT's input, and measure the DUT's output with your SA at a whole bunch of frequencies and plot it.

If you just want some basic values like the -3db dropoff etc, then you could just measure the amplitude until the voltage gets to be -3db or 0.7071 times the voltage. You can play around with the sweep function (if it's present on your generator) and different settings in the SA, but DO NOT expect your external sweep generator to act as a tracking generator; the signal needs to be synchronized to the scan rate of the SA for that to work.

Offline w2aew

  • Super Contributor
  • ***
  • Posts: 1024
  • Country: us
  • I usTa cuDnt speL enjinere, noW I aR wuN
    • My YouTube Channel
It is unlikely that you'd be able to sync the sweep of you siggen and your SA, However, there is still a way of getting the same sort of result.  What you would do would be to setup the generator to sweep the desired frequency range, and setup the analyzer to span the range of interest.  There are two things you'll need to do to make this work:
1) Ensure that the sweep speed of the generator is much slower than the sweep speed of the spectrum analyzer.  It's likely that the generator has a stepped sweep function instead of a continuous sweep.  If so, set the dwell time (the time it sits at each step in the sweep) to a value longer than the sweep time on the Spectrum Analyzer.  This will ensure that the signal is "stable" during each sweep of the analyzer.
2) Set the analyzer to display a Max Hold trace.  After many sweeps of the analyzer, your max hold trace should show you the Bode type plot that you were looking for.
======================================
YouTube channel: http://www.youtube.com/w2aew

Online Aurora

  • Super Contributor
  • ***
  • Posts: 1980
  • Country: gb
I agree with the previous comments and have used the swept sig gen with SA in Max hold to good effect. You could also consider building a simple RF noise source using a Zener diode and MMIC or Bipolar amplifier. You may not get a flat output response but you do get a nice broad band source of RF that you can 'normalise' against without the filter in circuit, and then view the response with the filter present. It works pretty well and has been used for years by Hams without expensive sig gens available. I have just bought and repaired an Advantest R3361A which includes the tracking generator....  Christmas arrived early for me  :)

I have also used swept signal generators called 'Wobulators' with a simple diode detector providing amplitude detection. These connect to the X and Y inputs of an oscilloscope and plot the RF response on the CRO display. Mine was a Wavetek 1002 but others can be found on ebay quite cheaply.

For RF noise sources look here for starters:

http://www.n5ese.com/noise.htm

http://www.n0ss.net/noise_generator_ac7ac-style.pdf

http://www.ham-radio.com/k6sti/nsrc.htm

http://www.qsl.net/m0ayf/Noise-source.html

http://bytecollector.com/archive/misc/BuildYourOwnNoiseSource.pdf


Have fun

Fraser :)

Offline updatelee

  • Contributor
  • Posts: 43
youve already got a signal generator though, youve got a si570 in your softrock. and it outputs a square wave so its got massive harmonics, perfect for testing a filter :)

Offline vk6zgo

  • Super Contributor
  • ***
  • Posts: 2408
  • Country: au
youve already got a signal generator though, youve got a si570 in your softrock. and it outputs a square wave so its got massive harmonics, perfect for testing a filter :)
Not really,as the level of the harmonics fall off as they increase in frequency,so you have to use a correction factor.

Offline updatelee

  • Contributor
  • Posts: 43
yes of course, but just compare the with filter to without filter on your spectrum analyzer.

Online Aurora

  • Super Contributor
  • ***
  • Posts: 1980
  • Country: gb
"Normalise" is your friend !

If your SA has that function of course  ;)

Fraser


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf