Post Go back to editing

Phaser: Calculating FMCW Max Unambiguous Range

Category: Software
Product Number: CN0565
Software Version: Phaser Radar Examples
Jon,
First, thanks very much for posting video 7 in your Build Your Own Radar series. It's great! Adding the MTI information is awesome and super helpful. Also, thanks for the shout out in pointing you in the direction of the X-band horn antenna. These seem to work quite well. The MTI processing looks like it works really well! I haven't had a chance to do field testing with it myself yet but look forward to the improvement.
In looking through your Range_Doppler_Plot.py example I noticed you tweaked some of the key radar parameters since video 6. Specifically you sped up the ramp time from 500 us to 300 us and lowered the Pluto's sample rate a bit from 5 Msps to 4 Msps. My first question was what prompted you to make these tweaks?
My second question has to do with finding the Phaser's FMCW max unambiguous range given these new parameters. I used the new parameters from your Range_Radar_Plot.py:
       c = 3e8
       sample_rate = 4e6
       chirp_BW = 500e6
       ramp_time = 300 us
And then calculated the maximum range, Rmax using what I believe to be the correct calculation:
       Rmax = (fmax *c) / (2 * ramp_slope)        (equation A)
where:
       fmax = sample_rate / 2 = 2e6    (to meet Nyquist max measurable bandwidth frequency using the Pluto)
       ramp_slope = chirp_BW / ramp_time = 1.667e12
which gives the max range (I believe) of:
       Rmax = ( 2e6 * 3e8) / (2 * 1.667e12) = 180 meters
Does this seem correct? And now to the real question that's bugging me. Many references I  find that include the FMCW max range equation will have the sample rate in place of the fmax and are written in some form like this:
       Rmax = (sample_rate * c) / (2 * ramp_slope)    (equation B)
which would give twice the max range, 360 meters in this case. Is the max range equation properly of the form of equation A or equation B (or something else entirely)?
Surprisingly, I couldn't find a clear reference to this is Charvat's book nor in the MIT Lincoln Lab lectures. Here are some other references to FMCW max range it I've found in searching around that only served to confuse this seemingly simple issue:
There are many others however none clear up this confusion between using sample rate and frequency. Can you you guys help clear this up?
Thanks very much,
Robert

Thread Notes

  • Hi Robert, nice to talk with you again.  Those are great questions, and I'm planning to do a video soon where we walk through the range equation and how to optimize those parameters. 

    But to answer your main question:  is the maximum baseband frequency you can generate equal to sample_rate or sample_rate/2?  For a single data converter, it is sample_rate/2, as you point out.  BUT..... remember that Pluto is a complex data converter.  Not complex as in difficult!  But complex meaning it has both an I and a Q data converter.  So it has 2 ADCs for each receive channel (i.e. Rx0 and Rx1), and then two DACs for each transmit channel (i.e Tx0 and Tx1).  

    https://wiki.analog.com/university/tools/pluto/users/non_quad

    So when we send Pluto 20MSPS (for example), we're actually sending 20MSPS of complex (I and Q) data points.  So 20 MSPS to the I converter and also 20 MSPS to the Q converter.  That means each converter can do 10MHz of signal generation on one side of the carrier (negative and positive freq).  Putting that together is what gives us 40MHz of possible baseband frequency creation for Pluto.  

    So the simple answer is if it is a real ADC/DAC, then max freq is sample_rate/2.  But if it is a complex ADC/DAC, then think of it as sample_rate.  

    You can see this in action with an excellent video by Dr. Metzger:

    https://youtu.be/MQi-sQ1GZcY?si=bTHkqvfGrfT8hkdy

    For example, he does a 20MHz chirp BW for a 30MSPS Pluto sampling rate.  He also has some advice on how to set various parameters in there.  But he could say much more, he's a very knowledgeable consultant.  

    Hope that helps!

  • Jon,

    Ah yes, of course! A bit of a head smack moment. Makes perfect sense. I needed to think about this diagram that shows the full IQ bandwidth from -fs/2 to fs/2 just as you reminded me:

    And thanks for the pointer to Dr. Metzger's video. Super interesting. I've already watched his whole series.

    Look forward to seeing more of your YouTube videos.

    Robert