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 - )

1996

1. Microwave frequencies must be used for both uplink and downlink.
2. Pulse modulation must be used for the most accurate range measurement
3. A PC must be the primary radar display and storage device.
4. The ground antenna must automatically track the rocket, and give accurate angular information.
5. The ground antenna must be smaller, preferably a 2-ft. or smaller parabola.
6. Both uplink commands and downlink telemetry must be supported.
7. The transponder must be simpler, smaller, and cheaper than before.

November 18, 1996

• Where has time gone? I've been working on the microwave upgrade. I have built some S-band hardware (an experimental transponder), but recently I've been wondering if Ku-band wouldn't be better suited to the goals of DARTS: small, portable, etc. To this end, have been weighing the pros and cons of S-band and Ku-band:

Band	Pro	Con
S-band	Cheap, small available components; low receiver noise figure; high power available	Large antenna aperture for sufficent gain->large, expensive mechanical eq; large size of feed (aperture blocking); Low gain from reasonable size antenna
Ku-band	Cheap, small available components; high antenna gain from small aperture; highly directional; small mechanical components->small, portable package	Component selection limited; noise figure not too good; problem fitting monopulse components at feedpoint without aperture blocking; cost of obtaining experimental equipment

1997

January 9, 1997

• I've decided to use the 5.7 GHz ham band, mainly because components for that band are relatively easy to get, and Rogers RO4003 (great PCB--try it!) has low loss there. I'm right now working on the design of the monopulse feed for the ground unit's parabolic antenna.

I took some hints from the Westingouse paper I mentioned earlier, and planned the feed to look like this:

Offset-Fed Prototype

Dish, Front View

In the above picture, you notice that there is no feed at the center of the dish. That's right, it's an offset-fed dish, which is much more efficient than any center-fed dish. Also, the feed assembly has been removed for these photos.

Dish, Side View

In the side view, you can see the elevation motor and gearing. There's a 9:1 reduction ratio, and by my moment-of-inertia and angular acceleration analysis, there's a 2:1 overdesign for elevation tracking of a rocket accelerating upward at 22g a mere 500 feet away. The azimuth system is less critical, as the rocket doesn't move quickly horizontally unless there's a real problem!

By the way, to save weight and gearing, the optical encoders are built right onto the motor shafts.
 

Here's a picture of the dish on an early test tripod.

Below is a section of early prototype feed switch board that used Mini Circuits YSWA GaAs switches. These switches proved to have too much loss at 5.7 GHz. The thumbnail below is a rather grainy blowup of the actual. Click the thumbnail photo to see the full-size version.

Here's a block diagram of the interrogator as I understood it at the time.

Sooner-Boomer 14

From my logs:

April 30, 1997

• The new tracking antenna, using an 18" DSS dish, is nearly complete. Using a small dish lowered available gain a bit, but is MUCH easier to articulate and presents a lower windload than the 24" prime-focus dish I started with. Also, the offset feed eliminates feed blockage and makes the feed easier to design mechanically, since it can now be almost any shape.
• The receiver conceptual design is complete, and I am currently building the IF A/D and threshold detector board.
• Mods to the DSP board from the previous DARTS incarnation are complete, and I'm writing fimware to control the antenna pedestal.
• The transponder conceptual design is complete. I am in the process of collecting components for it. I plan to use surplus modular ("tiny brick") DRO oscillators to keep the first few transponders cheap. After that, a DRO will be built directly on the transponder PCB.

However, this slowed down development on the radar, as I worked mostly on other projects at work and home that fall, including setting up house.

In September, I put together a prototype of the antenna servo control system, and used it to move the dish around.  At this time, I learned just how front-heavy the offset-fed dishes were.  I was wondering how I would solve that particular problem.

In November, Chris Schuermann and I began planning to build in infamous "large and dangerous" aluminum pedestal.

December 1, 1997

When I get time, I am working on the interrogator pedestal and antenna. We have worked out most of the mechanical features of the production antenna pedestal. It will be short, light (made of aluminum!), and easily disassembled.

December 8, 1997

I've been having a lot of discussions about monpulse vs. sequential lobing with some of you (most notably the very knowledgable Jayme Henderson).

I also began working on the patch antennas for the transponder and interrogator.  To test some theories I was working on, I built some models.  The following is from my logs:

Here is a transponder test antenna I constructed. The photo thumbnail (click for large version) is on the left, and the PCB layout is on the right. The PCB layout is not to scale.

These are the antennas that I've been using in my transponder antenna tests. They are 1/2-wave-by-1/2-wave patch antennas (linear-polarized in this test; later they'll be circular-pol, since only the feed needs to be changed to accomplish that). Also, each has a single ERA-1 MMIC, pointed in the appropriate direction for transmit or receive.

Since the transponder both transmits and receives, I've built test antennas for both directions.

The boards are made out of Rogers RO-4003 ($34 for a 12x18 sheet of 0.020 thick w/ 1oz copper plating). The board design was done using Accel Technologies EDA, using my HP 48SX for calculation of the microstripline dimensions.

Here are both the transmit and receive antenna boards together. The antique 6" slide rule provides a size reference. The transmit board is on the left, and the receive on the right.

Transponder Antenna Test Boards

Below is the transmit board. At the top is a 1/2-wave-by-1/2-wave patch, connected via a 70-ohm, 1/4-wave transformer line to the wider, 50-ohm trace going to the ERA-1 MMIC amplifier. The matching section is necessary since the patch feed impedance is 120 ohms. A biasing resistor for the MMIC can be seen on the left, as well as a RED power LED and its associated dropping resistor. Below the MMIC (black round package) is a 10-pF coupling capacitor, and another 50-ohm line going to an SMA connector.

Transponder Transmit Antenna - Test Board

Transponder Receive Antenna - Test Board

Here's the receive board. At the top, the same 1/2-wave-by-1/2-wave patch and matching section. A biasing resistor for the MMIC can be seen on the left, as well as a GREEN power LED and its associated dropping resistor. Below the MMIC (black round package) is a 10-pF coupling capacitor, and another 50-ohm line going to an SMA connector. Note that the bias resistor connects to the LOWER 50-ohm line as opposed to the UPPER 50-ohm line as in the transmit board; this is because the MMIC 's input is connected to the antenna in the receive board, and its OUTPUT is connected to the antenna in the transmit board.

The ERA-1 has about a 5-db noise figure when used as a receive preamp; this isn't exactly "maser-quiet", but I'm evaluating the antennas, and in a strong-signal environment, so noise-limited performance is less important here.

November 12, 1997

• The new transponder antenna design is currently in test, and it works much better that even I expected! There are still some modifications to be made, though, and I want to run another flight test before I call it "good". I plan to fly the new design in a model airplane _very_ soon.

• The new interrogator IF board, incorporating changes from things I've learned from testing the previous design this summer, has been "buzzed" (milled on the T-Tech board milling machine), and I'm populating it right now. I expect at least a 6 dB improvement in log amp linearity and a 300% increase in pulse edge response. The faster edge response will increase range measurement reliability at low received signal levels.


The IF board uses an Analog Devices AD606 log amp. I can't believe the performance of this chip! I've seen 9-stage, tuned logarithmic amplifiers built into milled aluminum blocks that didn't have the performance of the AD606. Highly recommended for the experimenters out there.

Here's another bit for the experiementers, especially fellow T-Tech users: Rogers' RO4003 microwave circuit board mills so _nicely_ on our T-Tech 5000! After some experiementation, I settled on a 0.015" end-mill, because it doesn't break as often as a 0.010 mill, and it doesn't cut a frequency-shifting "trench" around the outside of my patch antennas like the T1 V-mill does.

• From the E-mail I have been receiving, I can tell there's two main sets of competing ideas developing around the transponder: the "basics" approach and the "minimalist" approach.


One school of thought (the "minimalist" approach) says the transponder should be as _passive_ as possible, and all control/telemetry should be handled by flight computers or other add-on boards. One advantage of this approach is that it represents the least-cost way to get into a transponder.

The other school (the "basics" approach) says that the transponder should include a basic set of control and telemetry "ports" tailored to the habits of most high-power rocket flights, with add-on boards handling the out-of-the-ordinary stuff. This group believes that the DARTS transponder, since it has spare computer horsepower, can replace the array of flight computers, jerry-rigged timers, thermalite, etc., that is now used to control the flight and parachute deployment of rockets.
 

The IF board used an Analog Devices AD606, which turned out to have too slow a response for pulse modulation.  This actually can be seen in the scope shot below.  The input to the log amp is a 120 MHz pulse-modulated signal generated by my ancient HP Model 608 VHF Signal Generator.

History Top
Generation III: Pulse Microwave (1998)