A Work In Progress

Remember that the pictures are clickable to bring up larger, bandwidth-eating versions.

To view the AVI clips on this page, you'll need to install one of the Intel CODECs below.   Sorry they're not Mac compatible.

• Original Intel CODEC (OK for Windows 95, but not WinNT)
• New Intel CODEC (Works under NT for me)

Generation III: Pulse Microwave (1996 - )

1998

Early 1998: Large Pedestal

March 18, 1998

The DARTS pedestal upgrade is nearly complete. This was required because the previous design was too flimsy for an accurate track.

Work on the interrogator and transponder electronics is proceeding.
 

Using this dish would have been a HUGE mistake.  I simply had no way of pointing a beam that narrow!  This dish was purchased because, in my effort to commercialize DARTS, I thought people would want far-away tracking capability (ala CATS), and more gain on the ground was one way to achieve that.  The larger 24" dish gave about 2 dB more gain than an 18" dish. (See, again, letting other people call the shots.)

The webbing is laser-cut mild aluminum, and the angle aluminum pieces are 6061-T6 aircraft-grade. Man, is that 6061 ever tough to drill!
 The completed pedestal leg:

Here's a closeup of the jacks we built to level the pedestal.  100% OVERKILL.

The body of the pedestal is heli-arc welded 6061 aluminum plate, built into a hexagonal tube:

Again, ENORMOUS overkill.  But my voice was drowned out by that of my boss, who had made it his personal mission to make the largest, most vulgar pedestal possible.

The legs are attached to the body via bolts and rivet-nuts. There are 6061-T6 brackets on the inside of the tube that carry the riv-nuts.

Here's the finished pedestal, with the azimuth platform on top:

The pedestal was sturdy enough for two people to sit on it, and maybe more.
 
 

Here we see the elevation (left) and azimuth bull gears. To the right is the elevation motor with its pinion gear. The azimuth pinion is the same size, but with a bigger motor.

(note:  we later abandoned this huge pedestal when we went from a dish antenna to a phased array.  Too bad.  Sure was sturdy.  But, it was immensely heavy!)

The CATS Prize

Around this time, someone pointed out DARTS to a group of leading-edge rocketeers chasing the Cheap Access to Space prize

...I've been having a lot of discussions with people who are chasing the $250,000 CATS (commercial access to space) prize, which involves tracking a rocket to 200km. Our official position is that DARTS is not ready at this time for tracking rockets to such altitudes. This means the CATS guys are going to have to find another way to track their rockets. We regret that we won't be able to participate in those historic flights, but we must move DARTS foreward in a manner consistent with its original design goals. Our main audience has always primarily been commercial-solid-motor Class B high-power rocketry. We remain committed to those goals.

I must admit that it was pressure from CATS contestants which forced me to conclude that DARTS' goals do not include commercializing it.  It seems, in my opinion, that the CATS competitors have their own agenda, which they are not willing to pay for.  (Another lesson learned: watch out for people who want to take over your project with their own NON-FUNDED agenda.)

April, 1998

Here's one of the 5.7 GHz horn antennas I built.   I built these using templates produced from Paul Wade's HDLANT program.  I used these horns on several of the field tests.
 
 

 I was so enamored with the clean appearance of these horn  antennas that I  considered building a multi-horn feed out of them.  Below is a geometrical (non-functional) model that I built.

 The pair of horns on the right was to act as transmit and receive for elevation (null-tracking), and the horn on the left would be used for traverse (azimuth) position sensing. The idea behind this feed is to perform null tracking only on elevation, and interpolate the azimuth error signal to keep the receive beam centered on the elevation horns (since most of the target motion will occur in elevation).
 

Medford "Fun Launch"

 I flew the Loc "Lil' Nuke" on a G80, and it seemed to perform satisfactorily.
 

First 5.7 GHz Transponder Hardware

This is the TRANSMITTER only -- but the receiver is only about 1/5 the size of the transmitter. But-- be warned, it's UGLY. I hacked this one together to WORK, not to look good...

It's made, as are my other developmental transponders, of 0.020 Rogers RO-4003 material. I plan to migrate to 0.032 material, because it's stiffer, but I have a lot of the 0.020 material on hand. Also, later I learned to put "rails" on the side of the board, made of 0.062" FR-4 PCB material.  The resulting unit looks like an "I-beam", and is very sturdy.

The above photo shows a prominent view of the "plumber's delight" filter made from a copper plumbing cap. I learned this trick from Paul Wade, N1BWT, who in turn learned it from some European ham radio operators.

We can see the 80 MHz oscillator on the far right side of this board view. Just to the left of that, in the shielded can, is the multiplier section.  This was my attempt to multiply a square-wave at 80 MHz to 5.76 GHz by using a high-rate SRD (Step-Recovery Diode) multiplier.  The SRD is a charge-transfer device that depends on the transit time of carriers in the diode to transform a low input frequency into a higher frequency on output.

An 80 MHz computer clock oscillator is used to drive a step-recovery diode, generating copious odd-harmonic output. The harmonic at 5760 MHz (or thereabout) is quite strong. A 3-pole microstrip filter isolates only the harmonics in that region.

Following the filter is another MMIC amplifier, and a 5-pole microstrip filter. The final filter effectively isolates only the desired harmonic (others are more than 40 dB down). The filter output is both isolated and amplified by a final MMIC stage up to the 0 dBm level, which drives the final amplifier stage.
 

SRD Details

SRD multiplier components

The SRD needed more input power.  All the SRD stuff I've seen says the SRD likes to be driven with between +15 and +26 dBm into 50 Ohms.
 I think the pipcap served as a nice resonator for the SRD.  Also, I think I used three stages of amplification, one of which was  the MGA-86576.

There are two stages of MMIC amplification are to the left of the shield. You can also see the output pad of the pipecap filter. A pin probe through the board into the pipe cap's "cavity" couples to the filter.

Here's another nice view of the pipcap filter. I was pretty enamoured with these little things: very high Q, easily tuneable, little or no radiation (as opposed to microstrip filters), and low insertion loss.

However, they're really too heavy--at high accelerations, calculations showed, the plumbing cap would rip right off the board.  Back to the drawing board...  Also, the SRD multiplier didn't work out.  For the factor of 64 multiplication I was looking for, I needed much more input power.  I did get a detectable signal at 5.76 GHz, but only at about -35 dBm.

 Pulse Modulator

The circuit has three functions:

1. Delay the incoming trigger pulse (RXPULSE) 12.8 uS before transmitting the output pulse.  This function is only significant in the transponder, where it is used to generate a fixed ehco delay.
2. Generate a pulse whose width depends on the state of the TLMDATA line.  If the TLMDATA line is low, a pulse 12.8 uS wide is generated.  If the TLMDATA line is high, a pulse 6.4 uS wide is generated.
3. Modulate the 80 MHz clock with the pulse in a phase-coherent fashion.

In addition, the circuit has a transmit disable input (TXDIS), which disables the MOD80MHZ output.
The BEACONPULS input allows the circuit to be triggered by the transponder's microprocessor instead of the receiver, to create regularly spaced beacon pulses which aid in finding the transponder.
 
 
 

Above is the transponder transmitter board that was the next generation after the above one. After I built this unit up, I weighed it: only about 3 ounces, WITH NiCd rechargeable batteries. Also, it's only 1.2 inches wide (including its enclosure), and about 5-1/2 inches long. With the antennas and receiver, it will end up about 7 inches in length. I'd like to shorten it, though, by using more surface mount components.

 I prototyped some digital logic to pulse modulate the transponder.  This board would generate pulse widths based on the telemetry, and also generate the precise digital delay that is critical to the close-tracking capability of the tranponder.  On the board also is the PIC16C84 microcontroller (near bottom) which oversees all transponder operations.
 
 

Below are some screen shots of the pulse waveform.  The circuitry on this board drove the frequency multiplier chain on the early transponders. It generates a 12.8-uS-wide train of 80 MHz pulses when a "one" is being sent to the ground station, and a 19.6-uS-wide train when a "zero" is being sent. I caught the transmitter on a 12.8-us "one" when this screenshot was taken, even though I set the marker to 19.6 uS.:
 
 

The First 5.7 GHz Interrogator Hardware

And the preamplifier to go with it (this is an Al Ward W5LUA design):
 
 

 The output of the downconverter went into the AD606 log amp (later found to be too slow for pulse operation), and then into this custom IF/ADC board:
 
 

June, 1998 - Turning Point for LDRS 17

After working all year to support the LDRS 17 launch, by early June I hadn't gotten any decent patch antenna stuff working, and I had to admit that we wouldn't be able to support the launch.  Disappointing those hardy rocketeers was one of the hardest things I ever had to do.  (But, this was convincing me all the time that DARTS was not, and couldn't be a commercial product, at least in the forseeable future.)

June 9, 1998

I have always believed that of all its interesting technologies, the microstrip-patch-fed tracking antenna is what sets DARTS apart from older-generation radars. And it's the one that's giving me so much trouble! It is for this reason that I must make the announcement:
 
 

DARTS will not be ready in time to support the LDRS 17 launch.

 I know this is a shock to many of you, especially those of you who wanted to fly a transponder at LDRS. I apologize especially to the Utah Rocket Club, who have been avid DARTS enthusiasts for a long time.

DARTS isn't ready for prime time just yet, though. I need more time to fabricate and experiment with different feed configurations. I had contemplated using horn antennas in an effort to "rush out" something that works. I just couldn't bring myself to take that much of a step backwards just to get something to work.

I have to occasonally be reminded that DARTS is supposed to be a fun hobby project, a showpiece for a variety of interesting technologies. And that DARTS is still experimental, a laboratory of sorts for all kinds of whacky ideas.

(What an idiot.  I should have stayed with horn antennas.)

The First Operational 5.7 GHz System:   Sooner-Boomer 15 (June 26, 1998)

Sooner-Boomer 15 was the first time the 5.7 GHz DARTS transponder has flown on a rocket.  It was very important to me to fly the transponder on a rocket at the earliest date possible, even though most of my goals could have been accomplished using other means.  Amateur rocket flights are _brutal_ on electronics.  From my observations, more than 50% of electronically-controlled ejections and engine firings fail.  Admittedly, some of this could be due to failure of charges and firing mechanisms at altitude, but is probably due mostly to electronics failure.

I had several goals for the Medford outing:

• Get DARTS hardware out in the field, and see how it performs in actual hot-summer, rocket-launch conditions.

 
• See if we can track the transponder during flight, and receive it on the ground, thus validating the transponder's antenna design
• Flight-test transponder hardware, checking to see if rocket acceleration, ejection, or the materials making up the rocket would affect operation of the  transponder.
• Check to see if 10 mW, using a dish for reception, is adequate for a 2-5 mile path. All my calculations told me this would be quite adequate.
• Decide if the antenna pattern of the ground antenna was sufficiently narrow, or whether it was desirable to make any modifications to it.

A closeup shot:

Here's a picture of the entire package that fit inside the nose cone.  To the left, the transponder's emergency backup chute can be seen.
 
 

 The Interrogator antenna was an 18" plastic RCA minidish, and a horn made from tin plate and designed using N1BWT's HDLANT31 program.  The receiver consisted of a surplus Collins 6 GHz satellite ground station downconverter, and a California Microwave "brick" oscillator.  Below is a photo of the setup:
 
 

 The pedestal is the same as I used at last year's Sooner-Boomer launch.

Sitting on the cooler is the Interrogator IF and processor.  Here's a close-up shot of the front panel:
 
 

An AD606 log amp board was used, since I was using CW for this test and didn't care about pulse response time (a faster AD640 is the one being used in the pulse interrogator).  The output of the log amp was fed into the A/D converter of a PIC16C73, which calculated the signal strength and displayed it on the LCD.  Also, the '73 generates the sigal-strength-dependent tone used to point the dish.   Below is a detailed internal shot of the interrogator, with labelled components:
 
 

The rocket was a LOC Precision "Lil' Nuke", a rocket about 28 inches long, powered by a G80-4T composite motor. I went for a low-powered motor for my initial test for two reasons: 1) it reduced the likelihood that the rocket would land far away, and 2) I am not yet certified for high-power (H-engine and higher) rocketry.
 
 

 Below is a shot of me carrying the rocket to the Range Safety table.
 
 

 Below is a picture and video of me explaining what we are about to do.
 
 

All day long we had 40+ MPH winds, which delayed me getting the rocket up until afternoon. The pavillion that I put up to shield us and our equipment from the scorching Oklahoma sun was torn up in the wind, and a couple of the steel tubes that made up the framework were snapped in half.  You can see how far we tilted the rocket into the wind for launch:
 
 

Below are some recordings of the tone as it varies from off-target to on-target:

darts_sb15.wav  - 11.025 kHz 16-bit stereo WAV file (463 kB)
darts_sb15b.wav - 8.000 kHz 8-bit mono WAV file (84 kB)

Although DARTS is designed to be auto-tracking, the signal-strength tone is still a worthwhile diagnostic aid, and will still be there even when I get the auto-track function working.

Things were working well, so in early afternoon, we prepped the transponder and loaded the rocket on the launch stand. Almost immediately, the high winds ripped the launch lugs off the rocket. I had to make emergency repair with 5 minute epoxy (aka The Rocketeer's Friend).

When the rocket went up, I was able to manually track it using the signal-strength tone.

On ejection, the shock cord was ripped in two, seperating the transponder-bearing nosecone from the rocket body.  The body tube fluttered down unharmed, and was recovered.  However, due to the high wind, the nose cone drifted to the north for several minutes.

I was listening to and tracking the nose cone during its entire flight, and even when it touched down, although I could no longer see it with the naked eye.  This is one of the real benefits of DARTS technology.

We believed the transponder and nose cone landed in a field just beyond the Medford golf course, about a mile downrange.  As you can see in the video, the rocket and nose cone parted ways as the elastic shock cord ripped loose. The body tube was recovered intact. We loaded DARTS up in our pickup truck and drove around, listening to the signal.  Although the interrogator batteries were getting weak, we traced it to a field filled with waist-high grass and scrub-brush.  Unfortunately, this field was in a "red-zone" (private property), and no recovery was possible unless the property owner finds it.

It was not very important to me to recover the transponder, since I had learned what I wanted to know--the transponder could be received in-flight and on the ground.  While the distances involved weren't large, I think we proved our point with this flight.
 

Summer 1998

• The transponder must be smaller and lighter.
• Pipe-cap filters are unaccepable in the transponder, because of the uneven mass load.

Note the sturdy "I-beam" construction.

 I've built several versions of transponder transmitter similar to this, and all have been disappointing.  The multiplication using SRD's is hit-and-miss with low input power, and high-factor MMIC multiplication is experimental at best.  I am fairly sure now that real DARTS transponders will use an oscillator at 1440 or 2880 MHz, and use a MMIC for multiplication.  A factor of 2 or 4 isn't hard to do.
 

PIC-based Interrogator

Right now, I've got a 20 MHz PIC16C76 processor designed in.  Here are some photos.

Phased-Array Antenna Experiments

To track a rocket, I needed angular information in two orthogonal axes; in this case, elevation and traverse (azimuth).  With my level of expertise at that time, I didn't see how I could design anything to make the dish scan in two coordinates.

At the time, I considered an offset-fed dish fed by a small patch antenna array to be an optimal combination.

To get angular information, you need offset beams.  Frequency scanning is a technique for offsetting the antenna beams that requires no active components on the antenna.  Simply switch frequencies, and the beam tilts.  Perfect, I though.  I can frequency-scan in the slow-moving coordinate (azimuth), and beam-switch in the other.  The dish antenna would collimate the wide beam produced by the small array.

Series-feeding the array of patch antennas was another design simplification.  The less transmission line on the antenna board, the less the pattern is spoiled.

However, I found that it is difficult to analyze small arrays, for a number of reasons.  To analyze small arrays, one has to strictly evaluate Green's functions for them, a computationally and analytically expensive proposition.  In addition, in series-fed arrays, the assumption that only a forward power wave excites the patches
falls apart for a small array.
 

Now, I was getting desperate (and delusional).  So, I decided to model a large array.  Well, it seemed that a 11-element subarray was about the smallest array I could design without the  equations/assumptions blowing up.  That size array cannot efficiently feed an offset-fed dish, for the beam is too narrow.  In fact, the gain was high enough to use
an array for the entire interrogator antenna.  I combined 11-element subarrays to create a 88-element array.  More elements in each subarray smooths out the sidelobes and increases the gain a little.

Among the many advantages of the array is that it is light and easy to move.  It will allow the interrogator to use smaller motors and gears, and be itself smaller and
lighter.

So, I modeled, built and tested a single subarray, which does the frequency-scanning in azimuth.  This kind of antenna is called a traveling-wave array.  A screen shot of my CAD model is shown below.  Yes, I know it's kind of wide...

Here's a  photo of the prototype.  I removed the "meandered" feed lines because I did not then know how to model their interaction with the patch phase shifts.

To test it, I decided to put the source antenna (a horn) high up on a pole, and twist the subarray in elevation.  This, I thought, would reduce ground reflections, since the antenna has a wide beam in the horizontal axis, when mounted as seen below:

By late summer, I knew that the dish would fall by the wayside.  DARTS was destined to have a smaller, lighter antenna.  I was patently delusional at this point, thinking I could do what Hughes Radar does, in my backyard.
This is the Space Age!

End of Commercialization Focus  (September, 1998)

I finally just sent out a rather harsh E-mail to all the DARTS E-mail list recipients, stating that DARTS would not be commercial in the foreseeable future and to stop bothering me.  Several sent E-mail in response, wanting to be taken off the list.

Having to do this hurt a lot, but I felt had to do something to stop the 'snowball effect'. I can only accomplish so much.  I am subject to the winds of change within my company, and within my life.  There are only so many hours in the day, and there are, believe it or not, many more important things to be working on than DARTS.  It is sometimes a back-burner project, and sometimes a hot one.  It just depends on availability time and resources.

Work goes on, albeit at a slower pace.  I expect to be able to spend more time on DARTS this winter. All things considered, I just don't think I can quit doing DARTS.  I'm addicted :) and I'm a big supporter of people doing experimental things just for their own sake.  As you know, many fine things in the world got their start that way.

Fall 1998

Tripoli Oklahoma Fun Launch:  October 24, 1998

To loft the transponder, I used my venerable Loc "Lil' Nuke" firing an EconoJet G35-4W.  Due to the light weight of this rocket and of course of the itty-bitty transponder, the flight peaked out at about 2000 ft (an estimate based on height of surrounding gauge objects...REAL DARTS altitude soon!).

I bundled the transponder up in bubble wrap, and placed it in the little cutout I'd made at the rear of the Lil' Nuke's nose cone.  More bubble wrap was placed in the cone both above and below the transponder (I actually think this was overkill).  The 9V battery that powers the transponder was taped to the "I-beam" frame of the transponder.

The TX5B worked part way during ascent, but the signal was lost abruptly either late in the boost phase, or at ejection...it all happened so fast!  I didn't think the transponder passing out of range would cause such a "cliff" in the signal strength, so I figured there had to be a failure somewhere, maybe a loose battery.

I discovered that the 5V voltage regulator (a TO-92 78L05 for the electronics types) has pulled loose from the board.  This shut down the VCO, and the amplifiers just had no input signal!  The 9V battery clip held just fine, which to be expected since it had no stress on it.  I will redesign this board to incorporate a surface-mount voltage regulator.

TX5B

I used the optimizing feature of the Eagleware SuperStar Pro simulator to design the hairpin filter for this one.  Here's a link to the circuit file:  5el_1.ckt
Here's the theoretical S21 and S11 curves:.

Tuning it around, I think the shape is pretty close to theoretical.

The TX5 transponder transmitter was quite successful.  I flew the third version (TX5B), on a rocket in September.  I'd done some modifications to it to reduce the current draw, which is now 131 mA (was 210 mA).  The new HP MGA-86576 MMIC pulls very little current (16 mA), but has good gain at 5.7 GHz (around 16 dB).  This is an ideal part to follow the TX5's bandpass filter, building up the signal for final amplification by an ERA MMIC.

Also, I'd been experimenting with anti-parallel PIN diode multipliers.  While they were cool, they were mostly a waste of time, requiring FAR too much power to drive them.

By December 1999,  I had all but dropped the idea of using frequency-scanned antennas.  There are just so many problems with them:  keeping the pattern free of sidelobes over the entire scan range, loss of power due to the resistive termination, and the required high-impedance transmission lines, to name a few.  T

Generation 3: 1999
History Top

This document copyright Steve Bragg, KA9MVA. Updated: July 12, 2001