Question

I would like to use the Analog Discovery 2 to look at the carrier signal coming in from an AM broadcast station.  To do this I would need to tune or filter the many signals coming in on my antenna down to one station of interest.  I am thinking that this type of tuning is not something that the AD2 can do.  Is this correct?  Does anyone know of a simple device that can act as a front end to the AD2 to tune and perhaps amplify the signal in this way?  I have looked into SDR's that are very focused on doing just this but I am not seeing how I can extract the carrier signal before it goes to mixing and demodulation.

Share this post


Link to post
Share on other sites

16 answers to this question

Recommended Posts

  • 1

Hi @duncan crane

You could use the Math Filter channel in WaveForms application Scope interface for bandpass filter.

image.thumb.png.641f225a1f6146d3a2f2fb9eda22512c.png

Then you can record a longer acquisition, rectify and apply low pass filter.
Further you can even play the processed data.

image.png.34e11e3f33ceda9e1576fe326ef90d80.png

Share this post


Link to post
Share on other sites
  • 0

Hello @duncan crane,

I have moved your question to the section of the Forum where the Digilent engineer who is most familiar with the AD2 will be able to see and respond to your question, though with timezone differences it may be a couple of days (till Monday) when they are able to respond.

Thank you,
JColvin

Share this post


Link to post
Share on other sites
  • 0

Just a random thought: With 100 MSPS you should have enough bandwidth to capture the whole AM spectrum (almost) all at once.

Save to file, then demodulate or process in Matlab, Octave, whatever. Offline processing is usually the more powerful tool.

Of course, the "simple device" exists - if you can afford $20k+ for a vector signal analyzer with IQ quadrature outputs, then things get fairly systematic. Otherwise we're in amateur radio territory, there the direction is usually "make do with what you've got".

A tuned filter to approximately match the antenna to the high impedance (1MOhm) input of the AD converter can achieve voltage gain without need for amplifiers, if sensitivity is a concern.

Share this post


Link to post
Share on other sites
  • 0

I saw the Matlab option but have to admit I was scared off a little by all that I do not know.  Maybe I should jump into it.

Thanks X.

Share this post


Link to post
Share on other sites
  • 0

Atilla,

What you are showing with the Math Filter in WaveForms looks exactly like what I am looking for!  Very exciting.

Just to be clear, in your example the input comes in on one of the Analog Discovery 2 channels and is fed to a Math Filter routine in WaveForms.  The output of the filter (in brown) is in the time domain and shows the original pre-sampled carrier wave form?  So did the filter do a reverse FFT on the sampled spectrum?  How precise can I get this picture?  Can I see just a couple of cycles of the original AM carrier and be confident of where in time a given cycle starts as compared to a similar AM carrier signal coming in on the second Analog Discovery 2 channel?  That is to say, could I see a few degrees of phase shift between the two original AM signals?  Assume that we are talking about a 1 meg Hz AM broadcast signal coming in on two different antennas to the two Analog Discovery 2 channels.  I assume the two AD2 channels are synced. 

Duncan

 

Share this post


Link to post
Share on other sites
  • 0

Hi @duncan crane

I don't have antenna frontend, just acquired some signal with a wire on Scope Channel 1. The result was just typical radio noise.
I added the FFT view to see where the peaks are.
The Math 1 is configured do bandpass filtering, no reverse FFT but https://en.wikipedia.org/wiki/Butterworth_filter which adds a bit of delay, but this should not create any problems.

The AD2 channels are synchronized, the seen phase is most probably due to antennas and/or cabling.

Share this post


Link to post
Share on other sites
  • 0

atilla,

I purchased my AD2.  Love it. Here is what I am seeing using your suggestion of a Math filter.

image.thumb.png.5939620bd4867af7cdbe618be147861b.png

This is a lovely picture of the 1.4 MHz carrier from my local AM broadcast station WHMP.  However, the triggering is not right.  Every time I do a Single capture I get this same picture but it crosses the 0 sec vertical at a different voltage.  I am guessing that this is because the source of my trigger is Channel 1 which naturally contains all kinds of overlapping signals coming in from my antenna.  So the Rising 0 volt point of Channel 1 will not necessarily line up with the Rising 0 volt point of the output from the filter.  I think what I really need for a trigger source is the output of the Filter not the input.  Can this be done?  I don't see an option for this in the Scope Trigger settings.  I checked out the SDK examples and I didn't see an example of using the output of a Scope Math Filter as the source of the trigger but I am not skilled in these programming languages.  Any thoughts?

Share this post


Link to post
Share on other sites
  • 0

Hi @duncan crane

The trigger source for Scope can be a real input channel, other instrument, manual trigger, external trigger or digital for mixed mode.
The Math channels are processed on PC in the application so the device can not trigger on these.

Share this post


Link to post
Share on other sites
  • 0

Hi @duncan crane

This is what kind of triggers the oscilloscopes are offering... Even the ones with many zeroes in the price tag, support triggering only on input channels or external trigger, eventually on digital signals/protocols. The math and other operations are performed after capture in software, so such data can only be used as qualifier.
If you want to trigger on demodulated signal with oscilloscope you have to use an analog filter setup.

Share this post


Link to post
Share on other sites
  • 0

I know I have a lot to learn about this device and plan to work through the conversations on the Forum and the many YouTube videos out there.  Please point me to any especially informative sources that you think I would benefit from.

You have got me thinking in a slightly different direction.  I have been talking about "triggering" but maybe I don't need triggering in the classical sense.  I like the way I can hit the Single button on the scope and capture a few cycles of my signals on the two channels.  I don't need a repeating sweep, just one capture gives me what I need.  And I don't care exactly when the capture started or at what point on the signal's upward or downward voltage swing the capture started.  What I do care about is that the capture on both channels starts at exactly the same time.  So that when I look at the two signals displayed on the scope I know that I am seeing when in time the two signals occurred in relation to each other down to a precision of a few nanoseconds.  Was channel 1 ahead of channel 2 by 100 nanoseconds or visa versa?   So maybe a Manual trigger is good enough for what I am doing.  But after the two signals are captured I need to apply that Math Filter to each of them to narrow the signal down to the one AM broadcast channel.  Would you expect the filtering process to introduce random delays or phase shifts between the two channels so that the original timing between the two signals is lost?  

I shall study on.

Share this post


Link to post
Share on other sites
  • 0

Hi @duncan crane

You can find tutorials, projects, online classes at: https://learn.digilentinc.com/

I see you only have about 2mV pk2pk filtered signal level. I don't see what you are capturing with the scope inputs, but 2mV is low compared to 5V range and 0.3mV resolution.
You might need to amplify the signal, like with better antenna...

I can see up to 100-200ps delays between scope channels even with normal signal amplitudes, so this seems to be a hardware limitation.
The filter adds delay, but as I read this is "mostly" frequency dependent. Having the same input frequency for two scope channels and the same filter applied it should add the same delay to both.
https://en.wikipedia.org/wiki/Butterworth_filter

image.thumb.png.6afa9fc4c921f94ef54e16ed1aa72286.png

Share this post


Link to post
Share on other sites
  • 0

atilla,

Sorry for the delay in response to your feedback.  I have been out of action with the flu and in fact am still a little light headed.

I think I understand what you are saying and have come up with a way to measure any random delay introduced by the two AD2 channels and subsequent Math Filtering.  I have two 10 ft long straight wire antennas stretched across my living room ceiling, mounted with insulators at each end.  (My wife just loves them.  Not.)  Using these two antennas as inputs to channels 1 and 2 on the AD2, I set the Math Filters on both to 560 KHz (near the low end of the AM dial).  Then with triggering turned off I sampled one shot at a time.  What I got was a consistent pattern of the two channels being fully in step.  No phase shift.  The signals jump to a new location each time I sample but they jump together.  This is good.

Then I did the same thing at a point higher on the AM dial: 1.4 MHz.  I got the same consistent results.  So for my purposes, the AD2 is capturing two signals at the same point in time and then allowing me to apply a Math Filter to each signal to pull out the one carrier frequency of interest, all without losing the relative timing of the signals from when they first hit my antennas.  This is very good.  My confidence in the AD2 is now very high and I feel that I can use it to work out what is happening with my experimental antennas.

Thanks, I will be in touch.

Duncan

image.png.219e7b2cc93632ca0edc270fa4026dab.png

image.png.6f48c8955c21437d143035bfb9cda405.png

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now