Prototyping the RADARduino

UPDATE: February 3, 2014:  The RADARduino has received a commodity classification from the Department of Commerce as 6A008.b.  Since the design itself will be open source, all of the technology will be freely available.  Check back here for more details soon.

For those of you who read IEEE Spectrum magazine, this month’s edition includes an article called “Coffee-Can Radar” which details author David Schneider’s experiences in building a short range radar based on the work of Dr. Gregory Charvat, who in 2011 taught a short course at MIT on Cantenna Radars.  This course has spread to other institutions including to the University of California Davis, University of Vermont, Michigan State University, MIT Lincoln Labs, etc.

Dr. Charvat’s FMCW (frequency modulated continuous wave) radar project uses a collection of connectorized Mini-Circuits components for the RF section, and several traditional components on a breadboard for the back end processing and VCO (voltage controlled oscillator) control.  Antennas are of the “cantenna” type, and the whole system operates in the 2.4 GHz band with an output power of ~ 10 mW.

I decided that this project called for a kit version built on a single printed circuit board.  So I started thinking about how to modify the project to expand on Dr. Charvat’s work, as well as provide backwards compatibility.

In thinking about ways to expand the usability of the radar, I decided that instead of controlling the VCO with a ramp generator chip, I’d use an Arduino microcontroller with a DAC (digital to analog converter).  This allows for the arbitrary generation of radar waveforms and frequency chirps, in addition to USB computer interface.

The first major component change was to integrate the RF components, which meant finding suitable surface mount parts for the LNA (low noise amplifier), mixer, VCO, buffer amplifier, and power splitter.  At this stage I decided to increase the frequency to the 5.8 GHz band which allows for the use of smaller antennas.

On the back end, I switched from a discrete anti-aliasing filter to an integrated part, and added a boost converter to generate the +12V required to control the VCO over its usable range.

This past week, we started assembly of the first prototypes, which are nearly ready for testing.

The first part of the assembly is the surface mount RF components.  It is planned that this portion of the circuitry will come pre-assembled, as there are too few people with reflow soldering capabilities, and packaging SMT components is time consuming.

Spreading solder paste onto one of the boards using a Kapton stencil

 

Solder paste applied to the SMT pads

 

Alex KJ6WIL placing SMT components on the three prototype boards

 

SMT parts populated and ready for the oven

 

Prototype boards undergoing reflow using an OSPID controller and a hacked toaster oven

 

Nearly complete prototype next to an Arduino (VCO is missing, will be installed soon)

 

The radar stacked on an Arduino

 

 

As was mentioned in the caption, the VCO is missing.  Digikey sent the wrong part, but tomorrow the right ones should arrive and will be installed.  We are looking forward to the next stage which will be to fire it up and see what problems or issues we may run into.  RF is never easy to get right the first time!

-Tony KC6QHP

 

 

 

 

40 thoughts on “Prototyping the RADARduino

  1. I’m very interested in this project. I’ve been searching very hard for hobbyist level radar kits/plans etc. My background is in automation and controls. Analog circuits scare me, but I’ve become more and more interested as time goes on. My home projects usually revolve around the Arduino platform, so this is perfect. When they become available count me in!

  2. Extremely cool. I’m a big fan of Greg Charvat’s radar work, but this has great potential to be smaller, cheaper without the Minicircuits modules and more robust, as well as maybe eliminating the need for a PC if only simple information such as altitude is desired.

    I look forward to seeing more updates with the results of your further testing, as well as schematics and design files if and when you’re hopefully going to be sharing them :)

    • Thanks! I’m hoping to post test results in the next few days, and yes of course the whole thing will be open source (though there are some elements of the windows application that will be closed, there will be fully open Python code to go with it)

  3. Following the article published on hackaday.com, I began looking on PCB creation to get a more compact version of the design. But you did it. Really cool.
    Waiting to see updates of your work.
    If you need help, I may help (specialized in software).

    • Thanks! I’m working on rev. 2 right now. There are a number of small changes which of course mean major layout changes :) At the moment I have two friends working on the software end of things, but once I have hardware available, I’m sure there will be plenty of opportunities to develop interesting software to go along with this.

    • At this moment I’m working on the layout for Rev. 2. In the next day or two or three, I’ll be putting up an entry here showing the full set of results from the first set of experiments (Doppler, RTI, SAR) on the first prototype seen above. So, thing are in the works. I’m hoping within ~1 month to have kits ready in some way.

      • Thanks for the update and quick reply. Looking forward to seeing the results. Im planning on incorporating this into a UAV I am building.

        • Excellent! I just talked to someone today about doing a flight demo on a quadrotor. Hopefully within the next couple weeks that will happen. There may be some image processing challenges and the question of transmit power will be interesting to evaluate.

          • Great work! I was hoping to find exactly this idea implemented. I would also be interested in testing one on a small fixed-wing UAV for potential inclusion in university research.

  4. You might find that you get a significant mismatch from the edge-mounted SMA connectors. I found that I had to narrow down the PCB pad for the centre pin to the right width for 50 Ohm, and also extend the groundplane till it was exposed on the edge of the board. Some PCB houses don’t like that, in which case I get out a file and file back to expose the groundplane myself.

    • That is a great suggestion. I am familiar with the effect you are talking about. On my x-band multiplier boards I use a different connector that uses a GCPW launch and a much smaller diameter pin and dielectric that is better matched to the board stackup. It makes a 2 dB difference in power over the connector used here. The price difference however is more like 6dB so I reserve that for x-band and above. I will try trimming down the pad here and see if that helps, but I need to see how that affects solderability for potential builders. As for the metallization to the edge, you are right that few board houses will do this. Even the high end places dont like it.

      I have not yet had the chance to put these on a VNA to check the match, but it is a worthwhile thing to do and should give me an idea of how bad it is. I may end up trying to put some matching in if it is bad. It may be as simple as adding some inductance beefore the pad to resonate it out. How that works across 5.5-6 GHz will be another issue.

      Tony

      • A 6GHz VNA is something I don’t have access to any more. When I get time I’m hoping to design a 4.4GHz one using LT5560 mixers, a bit like the N2PK one but with at least two s-parameter ports and with 2 receivers per port, and using a square-wave LO for the mixers which should make them quieter and have more stable gain than with sine wave LO. I am hoping to use the differential inputs of the mixers to do a subtraction at RF, to make a return loss bridge with only resistors so that it is broadband and simple. I hope to use an ADF4351 as the LO source, and switch over to a divided version of the ADF4351 signal for low frequencies. For the source generator, I guess I will want another ADF4351 with switched banks of filters to get rid of harmonics, and switch over to a DDS clocked by the ADF4351 for the really low frequencies. If I run the DDS only using tuning words that make the output frequency an integer fraction of the DDS clock, then it should be free of non-harmonic spurs, I believe. If I can figure out a good way to multiply the source ADF4351 then I could add a range above 4.4GHz, as the receive mixers would probably still have some conversion gain at harmonics. Sadly I lack the time to implement this all quickly.

        • That sounds very well thought out. Those very broadband PLL/VCO ICs are quite interesting and aside from the harmonic spur issue, are really great broadband sources. I have yet to play with them myself, but would like to one day. It seems that National/Analog/Hittite have some competition going in that market, and will be interesting to see what they come out with in the near future.

  5. Is there a reason why you’ve used the MCA1-60 mixer instead of the MAC-60 replacement? The latter is slightly cheaper, footprint-compatible and the milspec environmental rating stuff sounds good on paper if nothing else :)

    Have you seen this? Interesting:
    https://github.com/loxodes/mr_radar

    Operates at about 2.4 GHz, and uses components that can mostly all be obtained from DigiKey, rather than all MiniCircuits parts, which is convenient.

    • Interesting, I had not seen the MAC-60. Looks like it should work fine.

      I had not seen the web site for that radar, though I had seen mention of it somewhere before. I wanted to move up in frequency to 5.8 GHz for a couple reasons, mostly for the smaller antennas.

      -Tony

  6. It may even be possible to have a design where the VCO can be changed, and the splitter, mixer etc. can also be swapped for other devices (or the same devices are perhaps capable of broadband operation) allowing users to choose operation over a range of different bands between say 2-6 GHz, for example to meet local RF spectrum regulatory requirements in different countries, or to use 4.3 GHz, which seems to be the standard frequency of professional radar altimeters.

    • The VCO can definitely be changed and there are a number of pin-for-pin compatible parts that will fit, covering 2.05 to 8.7 GHz in several bands 200-500 MHz wide. The LNAs in the HIttite HMC71X family cover a good portion of that bandwidth, again in a few bands. Unfortunately they don’t have a part that is specified to cover 4.3 GHz. Otherwise the parts are pretty broad band and should work fine. The gain may need to be adjusted along the chain, but that can be accomplished with chip attenuators.

      -Tony

  7. dear tony my name is rediet ,from ethiopia ,in the Horn of africa.
    i am final year communication eng.student . my final year project is on ‘highway car collision avoidance system using cantenna radar system’ .
    i do have the idea and am staring to design it , but i can’t extend to simulation parts ..please if you are willing to help me let me know .
    thanks.!
    my email address: redietmom@yahoo.com

  8. I have a few quick questions.

    Is there an ideal spacing distance between the two antennae? From looking at prototype, it looks like you are using around two wave lengths.

    Does it matter if coax feeds from the vivaldi antennae go up or down? I just want to be sure there are no phasing issues.

    Do you have an availability update?

  9. I’m sorry to hear about the need to pause the project. :)

    Anyway, are you using a 4-layer stackup for these boards?
    The boards look like they’re about 62-thou boards, but the thin (20 thou or so) width of the 50Ω microstrip lines suggests that it’s about a quarter of that thickness down to the groundplane.

  10. I’m also very interested in this project. Sorry to hear that it’s been paused.
    Hopefully, it won’t be long before you can resume working on it.
    Thanks

  11. Hi Tony,
    Following you the last 12 hours, cause you are the first I saw in Greg’s Forum toggling with the Arduino succesfully. Any chance to lay hands on your SHIELD, I could be as well a tester for your prototypes (…am R&S Application Engineer and have all devices…).
    I wonder if the firsts shield of yours had also the RF integrated, or did you use the ARDUIONO only for postprocess and to bypass the XR2206 chip.

    Many thanks in advance
    Y3G

  12. Goodafternoon!!
    Any news? plz, we are waiting for this . its really hard to find this kind of wonderful project.
    :) Goodluck

  13. Do you have a status? I would like to buy one of these from you ASAP. I’m not near a savy as you with this stuff and this would help me out a lot. Let me know!

    Great Job btw!
    Brad

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>