Building Your Own Amateur Radar

Where to begin?  I asked the same question nine years ago when I started working on my DARTS radar.  There are no real answers to give.  As far as I can tell, there is no one else in the free world
working on an amateur tracking radar.  This is mainly, I believe, because it is such a hard thing to do, and encompasses so many disciplines.  It sounds so conceptually _easy_, but believe me that is a
self-deception of sinister proportions.

You _could_ build a 10 GHz "police" Doppler radar kit (Ramsey Electronics, www.ramsey.com I think), but I take it that isn't the direction you're going. You _could_ buy a surplus military radar unit, but
in many ways that would be more confusing than developing your own, since much of that technology is so dated.  Also, they are _huge_, heavy, and mostly use 3-phase 440 Hz power (where do you get
that?).

There are _no_ kits for real tracking radars anywhere in the world, as far as I can tell.  Heck, there weren't even decent microwave semiconductor components commonly available until recently.

My advice, such that it is, is to read _everything_ you can get your hands on, and to experiment. Read in all electronics disciplines, but especially in radar, antennas, signal processing, analog, digital, and
RF electronics.  (Check out my Radar Resources page for some ideas) Read up on control systems, and mechanical engineering.

Learn from the people who've done similar things.  Read all the stuff on my web site.  Read W1GHZ/N1BWT's web site, ESPECIALLY the Online Antenna Handbook.  Read _ALL_ the Microwave
Updates from the ARRL, cover-to-cover.

Learn _everything_ you can about radio electronics. Get an amateur radio license, if you don't have one.  Start experimenting with microwave radio. Join North Texas Microwave Society and one of the
ham VHF Societies.  Put together some microwave kits from Downeast Microwave (www.downeastmicrowave.com).  Read and understand everything in the HP RF and Mini Circuits data books.

And did I mention antennas?  Get some antenna software (like MiniNEC from www.arrl.org) and start experimenting with antennas.  You _must_ get a copy of Savatini's "Microstrip Antennas for
Wireless Applications", because antenna design software comes with the book (only $90!)

Oh, yeah, DSP.  Buy a DSP demo board from Texas Instruments or ADI and write some signal processing programs.  Learn all the DSP algorithms you can, from control systems, to filters, to radio
receivers in software.  You'll need 'em. Learn C if you don't know it.

Also, learn metal-working skills.  Learn to operate a drill press, a gas or electric welder and a Dremel tool.  Learn to measure twice, drill straight and cut true.  Heaven knows I've had to learn those
lessons more than once.

In this Tech Note, I'll mainly be talking about tracking radar, because that's primarily what I'm interested in. But, most of this information applies to weather, automobile, and sporting radar as well.
 
 

Semiconductors

Reflector Antennas

Horn and Patch Antennas

Surplus Equipment

Antenna Drives

Sources & Resources

 Semiconductors: MMICs, HEMTs, and Gunns

     Cheaply available stable gain blocks like these allow a change of paradigm in radar, one that allows personal radar to exist. This new paradigm trades off a few tuned, high-gain stages that are highly shielded (say, in a milled aluminum block) for several cheap MMIC gain stages. You stop worrying about 1-dB losses, because you can easily make these up with another MMIC amp. With this gain-centered paradigm, more signal processing can be accomplished on a small home-constructed PCB that in twenty pounds of aluminum-cased 70's-vintage stuff.

     Also, MMICs allow you to do more at RF, where the wavelengths are small, than at low IFs. For instance, signal processing is smaller and easier to implement at 3.5 GHz that at 70 MHz, using MMICs and etched filters, hybrids, power splitters, etc. If your tiny MMIC preamp has a lower noise figure than the conversion loss of a mixer, and will mount right next to your etched microstrip patch antenna, you've lowered your noise figure and overcome cable losses to boot.

MMIC amps from Mini-Circuits Labs, like the MAR series and the new ERA series, provide absolutely stable gain at up to 10 GHz. The new ERA-2, for instance, has at least 13 dB of gain in the amateur 5.7 GHz band, and 16 dB at 1296 MHz. All this at 1.52 each in single-part quantities with a 6 dB noise figure.

Also, check out HP's MGA-86576 and MGA-82563 MMICs. The first is a nice LNA which works very well at 5.7 GHz. The second is a medium-power amp which can output 100 mW when driven into saturation.

     High-electron-mobility transistors (HEMTs) and gallium-arsenide FETs (GaAsFETs) are available that can provide low-noise amplifiers up to 24 GHz. HP's ATF-36077 is a VERY low-noise LNA part that I use. Also, check out California Eastern Labs for datasheets and information.

     I don't get into this much, but for FM/CW and Doppler radar, don't neglect Gunn oscillators. SHF Microwave Parts sells 'em for cheap, and those guys will answer questions for you via E-mail. I've corresponded with Alan Rutz, and he's very nice and helpful. They also sell some parabolic dish antennas and various feeds.

 Microstrip Technology

     Microstrip filters for the microwave ham bands have already been designed by guys like Rick Campbell, KK7B. Also, there are microstrip mixers and hybrids (the latter useful for monopulse feeds!) See the ARRL Handbook and the ARRL UHF/Microwave Projects Manual. Also, check out the Microwave Update, put out by the North Texas Microwave Society, which the ARRL publishes, too.

     All this microstrip circuitry helps you get your project working faster, too. Gone are the days of spending hours to tweak 1 dB out of a waveguide filter. These filters are purposely built with fairly wide passbands to allow for mechanical tolerance variations, and an inexpensive MMIC can overcome whatever insertion loss there may be.

     And speaking of losses, these days, low-loss substrates are cheaply and readily available that blow lossy G-10 out of the water. RO-4003 from Rogers is one of my personal favorites. It can be worked just like FR-4 / G-10, can be cut with scissors, and is usable beyond 10 GHz. The Rogers site has the specs, and a 12" x 18" panel of 0.020 RO-4003 (double-sided) costs about $32. Not too bad, and way cheaper than RT-Duroid, also a Rogers product. By the way, I have gotten samples of RO-4003 from Rogers.

Matjaz Vidmar, S53MV, has come up with techniques to use FR4 at 2.4 and even 5.7 GHz.  It's worth a look at his no-tune transceiver article.

     Duroid and other Teflon boards are more expensive, and are available from places like Down East Microwave and Microwave Components of Michigan. Norm at Microwave Components has Teflon for $8.00 for a 6 x 3.5 x 0.031-in piece, and Duroid 6780 for $10.00 for 6 x 3.5 x 0.031 in. These prices are per his Aug 1995 flyer -- I guess the doesn't change his prices too often.

     Your home PC is your best resource for making PCB patterns. Puff is a neat old IBM-PC/AT-vintage program that can draw hybrids, couplers, etc., and help you get the right line widths for transmission lines. It's $10 from the CalTech Electrical Engineering department. There are other PCB programs that are either free demo versions or low-priced student versions, like ProTel or OrCAD. Check out a local college bookstore. I used to use Microgfrafix Designer to design my own PCB layouts, but now use the pricier (but much better!) Accel. Don Lancaster (www.tinaja.com), Radio-Electronics columnist and engineering-type-guru, writes Postscript directly for PCB patterns ! Use whatever graphics program you're familar with, but make sure the trace widths in the CAD drawing match the actual board widths, to within a few thousandths of an inch.

     To get your layouts on the board, I have found a product which takes the cake. Techniks makes a "PnP Blue" material that you can put in a Postscript (there, Don Lancaster!) or other laser printer and print your pattern directly on it. Then, just iron it on the board with Ye Olde Clothes Iron (carefully!) or other heat-press-type device, and voila! ready-to-etch microwave PCB. It also makes great front panels, too.

If you have a need to make many prototype printed circuit boards, and can afford an investment of around $20K, I highly recommend the T-Tech Quick-Circuit circuit-board milling machines. Since DARTS has been with C2 Technologies, I have been using this machine, with unbelievably good results. I can mill microstrip filters and patch antennas with 0.001" accuracy!

 Reflector Antennas

     MCM Electronics has offset-feed RCA DSS dishes. These little 18" beauties have about 26 dB gain (typ) at 5760 MHz, and about 33 dB at 10 GHz. They're a perfect size for small motors and gearing to drive. In addition, at 5760 MHz, the beamwidth is about 10 degrees, so they're fairly easy to point for initial acquision, and a little more forgiving of mispointing error than a larger dish would be.

Digital Satellite Source also sells the 24" dishes, which are billed as an upgrade to the 18" dishes.

     You'll have to compensate for the polarization distortion caused by the offset feed, but it's possible to use this to your advantage. Offset-fed dishes radiate cross-polarized RF with a linearly polarized feed. The cross coupling increases with shorter focal length and larger offset angle. Cross-polarized radiation is found not on axis of the main beam, but in the 45 degree planes close to the main beam. With circularly-polarized feed energy, there is no cross-polarized radiation but instead a slight misalignment of the main beam.

     See the MIT Radiation Laboratory Series, Volume 12, page 140, and T.S. Chu and R.H. Turrin, "Depolarization Properties of Off-set Reflector Antennas". Also, see W2IMU's Offset-Feed Dish for more details on the properties of offset-feed dishes.

     It's a little more complicated to find the focus of an offset-fed dish than a prime-focus dish. The following was contributed by kip@atl.mindspring.com via E-mail:
 

If you're stranded on an island and don't have all this new fangled electronic stuff, you can [find the focus] with a straight stick and 3 pieces of string.

Do it on the long axis (20 in for DSS 18 in dish) of the ellipse and you'll get the offset angle on the feed (about 24 degrees). Make all three strings the same length. Attach (loosely) the strings, one near the center and the other two near opposite edges of the dish. One end of all three are tied together. That will be the focal point. The other end ties to the stick and you keep all three parallel and at right angles to the stick. Adjust the position to get all three strings tight at the same time while keeping the angle in and out the same.

The distance from the center of the dish to the "point" (three strings common) is the focal distance. The angle between the two outside strings is the usable beam width of the drive antenna ( use about 10 dB down for low side lobes ). And, of course the angle between the "output" strings and the axis of the dish is the offset.

For an 18 " DSS type dish ( 18 x 20 " ellipse), guess about 24 degrees offset and 14 " focus to start.

Of course you can write some math equations to do too, but you need some accurate measurements to get the curvature correct.

     Paul Wade N1BWT has a DOS program called HDLANT that can calculate the feed of an offset-fed dish. He also has an offset-fed dish article in "Microwave Update 95", from the North Texas Microwave Society, and published by the ARRL. This article shows the calculations required, and gives the 18" RCA dish numbers as an example. I used these numbers directly.

     SHF Microwave has small spun-aluminum (center-feed) dishes good to 40 GHz. Also, talk to your local cable supplier or telephone company. Small 24" dishes are taken out of service all the time because of small dings, and these are usually free for the asking. I've gotten a couple from this source. Hamfests and equipment sales are good sources, too. I sure wish I could get hold of some Primestar^® dishes! E-mail me if you know a source.

     Some have asked me about the loss of gain from "dings" in the surface of a parabolic reflector antenna. Darrel Emerson, aa7fv@amsat.org, contributed this section via E-mail:

     The equation for loss of gain against surface roughness is often known as the "Ruze law", dating back to 1952. The antenna efficiency, compared to a perfect dish, is:

where sigma is the rms surface roughness and lambda the wavelength.

     This equation is quoted in most antenna textbooks, and I believe was used to derive the curves in the various ARRL publications. It's interesting that, to test the theory, Ruze actually took a perfect parabola and purposely put a small number of "dings" in it, measuring the antenna gain before and after. There's a picture of this dish, complete with dings, in one of his papers.

     Using this formula, if you want the dish to have 90% of its maximum possible efficiency, the surface of a reflector has to have an rms error of about lambda/40 or less. An rms error of lambda/10 will reduce the gain to about 21% (-6.9 dB) of its theoretical maximum.

     If there is just a small number of dents, it can matter where they are on the dish. Because the feed usually tapers the dish illumination down towards the edge of the dish, a ding at the edge will have less effect than a similar ding near the center of the dish.

     The energy lost from the main beam of course goes into sidelobes. The scale-size of the surface errors affects where these sidelobes (sometimes called the "error beam" or the "error pattern") end up. For example, if most of the errors are dings, say, 10 wavelengths across, then the lost energy will be within about (lambda/10) radians, or about 6 degrees, of the main beam. If the characteristic width of a ding is half as great, the lost energy will be spread out twice as far.

This is all from the standard textbooks.

 Horn and Patch Antennas

Patch antennas are nice for phased or fixed-focus arrays. What I use are the square pseudo-half-wave, circularly polarized elements. With a trim tab so you can adjust the circularity of the polarization.

See the figure below:

(Sorry, but there's small error in the drawing--the feed impedance is not 50 ohms, but 120 ohms...but, a simple 77 ohm quarter-wave microstrip will handle the mismatch.)

Single patch antennas' patterns are a bit broad to use them as feeds for parabolic reflectors. They work better for dishes with low f/D, but then phase center placement is quite critical. A small array can work for a feed, though.

     There are precious few resources on microstrip antennas.  One comprehensive antenna reference is Jasik's Antenna Handbook.   However, most of the patch antenna material in there is old (>10 years), and the most important work in patch antennas has been done since 1988.

I have recently learned that amazon.com, a huge bookstore on the Internet, has several good antenna books. One of my favorites for patch antennas is:

"CAD of Microstrip Antennas for Wireless Applications" Robert A. Sainati (ISBN 0-89006-562-4) (Dewey 621.3'824'0285).

I also highly recommend:

"Advances in Microstrip and Printed Antennas" (edited by) Lee and Chen (ISBN 0-471-04421-0).

 Surplus: Good Olde Junk

To invent, you need a good imagination and a pile of junk.

                -- Thomas Alva Edison (1847-1931)

     Well, everybody might not be that lucky, but go to the hamfests and troll around. Especially the larger ones like Ham-Comm and Dayton. You can find TWTs, magnetrons, and high-voltage pulse power supplies. Sometimes, you can even find antenna drives and even whole radars! In my humble opinion, you should stay away from all surplus radar consoles and displays. You can make a system that's FAR more adaptable by digitizing the range and angle data, then processing and displaying it in whatever format you want on your PC. Old radar displays are very closely tied to the old radar they were designed for, and that limits what you can do.

     More cool and useful surplus comes in the form of "brick" (modular) oscillators and amplifiers. These beauties are available for cheap and can help you get a system up and running in no time.

     "Brick" phase-lock oscillators, from Frequency West, Loral, CTI, and Rockwell, are available from 2 - 18 GHz in various ranges.  I use one for a receive LO in my radar; it has low phase noise and is quite stable, good enough for wideband work. There are several articles in Microwave Update about stabilizing these oscillators and changing out the crystal reference oscillators to move them in frequency. WA6CGR has a good "brick" PLL article on the web.

     Also, "brick" GaAs-FET amps intended as TWT replacements are cool, too. I have one that's spec'd to run at 6.2 GHz that runs just fine in the 5.7 GHz ham band, putting out 2W of RF, using a +/- 12V source. Oh, the specifications say +16V and -20V, but often these power supplies are a reflection of what's available in its intended environment, NOT what it will actually run with.

     Needless to say, to take advantage of the surplus market, you've got to educate yourself to know what you're looking for. They don't teach about radar tubes in college, folks. Read all the old microwave books you can get your hands on, like the entire 1946-vintage Radiation Lab series from MIT, Skolnik's Radar Handbooks, and anything about microwave technology. Go to Radio-Research Instrument Co.'s web site, request a catalog, and familiarize yourself with all the old radars and components there. If you get the chance, talk to hams who have been around awhile and are into microwave. Chances are, they know the "old ways" like transmitting tubes, waveguide, and feedhorns. They can teach you a lot, if you're willing to listen.

 Antenna Drives and Pedestals

     The best solution I've found is to use a surplus az-el drive from some old radar, or build an az-el drive from gears and motors yourself. The latter isn't a perfect solution, but it can often be satisfying to know you "built it yourself", and also you get to control the parameters of your pedestal, rather than having them picked for you.

     Intertec has an article that tells you everything you ever wanted to know about permanent-magnet DC motors and then some. PM motors are good for this application because they have a linear current/torque curve, and that makes the control electronics simpler; you don't have to have a PID loop for torque control.

     Arrick Robotics has a whole list of suppliers he uses for motors, gearing and stuff for robots, and much of this is useful for moving small antennas, too. Edmund Scientific has some motors big enough to move little RCA dishes.

     Wirz Electronics has a dual-channel, 3-A 36-V PWM motor driver board for $45 that is ideal for az-el drive of DSS and other small dishes. Wirz also sells the Hamamatsu sensors used to make cheap incremental encoders.

     And to generate your PWM, the National Semiconductor LM629 is the perfect PID control chip. It will control velocity, acceleration, and position of a motor, driving it with a PWM waveform, and receiving feedback via a quadrature optical encoder. That's a lot of functionality in a 28-pin DIP. Do what I did and set up two interactive PID loops, one on the LM629, and one in your radar's computer. The LM629 loop worries about keeping the antenna on a particular position or velocity setting, and the radar computer sets "goals" for the LM629 based on how far off boresight the target is.

     Don't forget Radio-Research Instrument Co.'s catalog, because they have a lot of antenna drives listed. Even if you don't buy from them, their catalog is a good resource for identifying the old radar antenna drives you find at hamfests and equipment sales.

     Another resource for mechanical stuff that's easily tapped are agricultural / farm & home stores. These people usually carry various gears, bearings, motors, etc. for agricultural implements, some of which are small enough for antenna mount usage.

Note: This Tech Note is constantly being updated, so check back frequently.

 Resources

• Arrick Robotics (www.robotics.com)

• Bayside Controls (www.motionex.com/bayside/bayside.htm)

20-20 Utopia Pkwy.
Whitestone, NY 11357 USA
Ph: 800-343-3353
Gear boxes for stepper and servo motors. Expensive but worth it for accurate positioning.

• Craftech Industries, Inc.

PO Box 636
Hudson, NY 12534 USA
Ph: 518-828-5001
Fax: 518-828-9468
Very broad line of plastic fasteners, fittings, and hardware.

• Digi-Key (www.digikey.com)

701 Brooks Ave. South
Thief River Falls, MN 56701 USA
Ph: 800-344-4539
Fax: 218-681-3380
Mail order electronic components

• Edmund Scientific(www.edsci.com)

101 E. Gloucester Pike
Barrington, NJ 08007-1380 USA
Ph: 609-573-6250
Lots of little gearmotors and miscellany.

• Globe Gears (www.globegear.com)

550 Virginia Drive
Fort Washington PA 19034
Voice: 1-800-523-2576
FAX: 1-800-635-6273
E-mail: globegear@worldnet.att.net
GEARS! 640-page catalog has standard and custom gears of all types.

• Hamamatsu Corporation

360 Foothill Rd.
Bridgewater, NJ 08807
Phone: 908-231-0960
Fax: 908-231-1218
Sensors. The 5587 is the start point of a cheap optical encoder. Wirz sells these.
• MCM Electronics

800-543-4330
18" steel & plastic DSS dishes, feed mounts
Ask for a catalog.
• McMaster-Carr

PO Box 740100
Atlanta, GA 30374 USA
Ph: 404-346-7000
Fax: 404-349-9700
The source for hardware of all kinds. Every antenna builder needs this one.

• Microwave Components of Michigan

Norman Alred WA8EUU
P.O. Box 1697
Taylor, MI 48180
(313) 753-4581
MMICs, PC boards, FETs, HEMTS, parts, etc. Be sure to include at least $2.50 minimum for shipping.

• Mini-Circuits Laboratories (www.minicircuits.com)

13 Neptune Avenue
Brooklyn, NY 11235
(718) 934-4500
(417)335-5935
Fax: (718) 332-4661
MMICs, switches, mixers, filters, etc. $50 minimum order.

• Radio-Research Instrument Co., Inc. (www.techexpo.com/WWW/radiores/radiores.html)

584 N. Main St.
Waterbury, CT 06704-3506
(203) 753-5840
Radars and parts. Be VERY nice to these people. The deal mostly with deep-pockets customers, and they don't have to help out amateurs, but they helped me out with information, and I was very appreciative.

• Rogers Corporation

Microwave and Circuit Materials Division
100 S. Roosevelt Ave.
Chandler, AZ 85226-3415
Tel: 602 961-1382
Fax: 602 961-4533
www.rogers-corp.com
RO-4003: http://www.rogers-corp.com/mwu/ro4000.htm
Duroid 5870: http://www.rogers-corp.com/mwu/5870.htm

• SAVA Industries

70 Riverdale, Rd.
Riverdale, NJ 07457 USA
Ph: 201-835-0882
Cables and pulleys.

• SHF Microwave Parts Company (www.shfmicro.com)

7102 W. 500 S.
LA PORTE, IA 46350
(219) 785-4552 fax
Antennas, Gunnplexers, PIN diodes, and other parts.

• Stock Drive Products (www.sdp-si.com)

2101 Jericho Turnpike
New Hyde Park, NY 11040 USA
Ph: 516-328-0200
Very broad line of precision mechanical components including gears, pulleys, bearings, and hardware. Expensive, however.

• T-Tech, Inc


Makers of Quick-Circuit board milling machines.

• Techniks, Inc

P.O. Box 483
Ringoes, NJ 08551-0463
Voice: (908) 788-8249
Fax: (908) 788-8837
PnP Blue is $30.00 for 1 pack (20 sheets).
•  

US Digital www.usdigital.com
11100 NE 34th Circle
Vancouver, WA 98682 USA
Voice (360) 260-2468
Sales (800) 736-0194
Fax (360) 260-2469
The source for optical encoders. A little pricey, but lots of good info at their site.

• Wirz Electronics(wirz.com)

P.O. Box 457
Littleton, MA 01460-0457
LMD18200-based servo motor driver boards .
Hammamatsu optical encoders.
PIC supplies.

[DARTS Main Page]
Send comments and suggestions to: steve@hamhud.net
This document copyright Steve Bragg, KA9MVA.