Going Supersonic on a G-Motor

I ordered an 868 MHz transmitter module from Radiocontrolli Italy, and etched a small PCB for a PIC12F508 microcontroller and RF phasing lines for feeding the tripole antenna. The device transmits occasional beeps, and prior to launch it is switched to constant transmission mode (as it takes ten seconds or so to stabilize). After launch it still transmits constantly for a dozen seconds or so, before resuming occasional beeps to comply with the duty cycle limitations on the frequency. Remote control of the transmission mode, and detection of launch, is accomplished through a three pin header at the bottom end of the rocket, which gets yanked out on launch. After landing, the ongoing beeps from the radio enabled finding the rocket.

I made a compartment for the radio and its tiny LiPo battery out of fiberglass. The compartment takes up half of the rocket's body tube's cross-section in the forward part of the tube, and is closed at its lower end. This allows the ejection charge to pass the radio without damaging it. Also the shock cord is protected by the same compartment. (No recovery wadding was used anywhere.)

You read the above right, I did say tripole antenna. I wanted a circularly polarized signal, since (1) a linearly polarized signal and receiving antenna would cause occasional nulls as the rocket rotates, and (2) a linearly polarized signal with a circularly polarized receiving antenna would lose about 3 dB of signal. I already had the RX antenna left over from the Iso-Haisu project. In it, the onboard TX antenna was a QHA (Quadrifilar Helical Antenna), which produces a beautiful hemispherical radiation pattern with perfectly circular polarization, but the necessarily narrow body tube of Iso-Joonappi couldn't take such a large antenna. Besides, I wanted to receive the signal from directly underneath (of course, for the Doppler measurement), where any antenna inside the body tube might be hidden behind the rocket motor's flame, its ionization attenuating the signal too much.

So I wanted to place the radiating elements in the fins of the rocket. A dipole antenna at 868 MHz will just fit in the fins in "sloper" configuration. But a dipole will be linearly polarized (plus it radiates broadside all over the place). To get circular polarization, I considered a pair of crossed dipoles, fed 90° out of phase. That's the traditional method of producing a circularly polarized signal (and it's also more directional, radiating mostly upwards and downwards). But that would need four fins, and that caused too much air resistance, according to my simulations. Luckily there had been lots of discussion on tripole antennas on the Finnish ham radio email lists. I basically fed three monopole antennas, one directly, one through a 1/3 λ phasing coax, and one through a 2/3 λ coax. The coax braids were connected together at all ends, and tied to the radio module's ground pin. Simple enough, though I have no idea what the impedance of the feed point is! I just hoped the TX module wouldn't mind the SWR too much. And it didn't; that's one benefit of a QRPP TX.

The photo shows a cardboard mock-up of the rocket's rear end, with the tripole elements fitted to the fins. Attached is the PCB with its coax phasing lines, with the TX module and its controlling PIC microcontroller still on breadboard. I used this awful prototype to verify that the antenna does produce a (sort-of) circular polarization, and that it does have (at least some) directivity. Ok, it's nothing spectacular on either front, since I just don't have the space or weight budget to do things properly... But it seems to work (and even the polarization direction is correct), so this design is going on the rocket!

The glass fiber body tube was tedious to make, as I wanted a maximally smooth outer surface. I made a tube form out of several layers of paper glued together, lined its inside with baking parchment (epoxy does not stick to that), and applied epoxy-soaked glass fiber fabric (about 150 g/m² I think, two layers everywhere) inside it, piece by piece, allowing it to cure between additions. I squished the laminate between the baking parchment on the outside, and the release and absorbent fabrics on the inside, by stuffing wads of absorbent fabric inside the tube to apply pressure. It took ages to build up the approximately 300 mm long tube, but it turned out rather good!

I did also try to simply roll up a sheet of epoxy-soaked fiberglass fabric, and press it against the tube form from the inside using an inflated bicycle inner tube. That did not turn out well at all.

I made the nose cone in a similar fashion, adding fiberglass bit by bit. This also took ages. Again, I wanted a smooth outer surface, so I made a paper form, which I lined with baking parchment, and began building the nose cone inside that. I used the old Vaisala rocket's nose cone as my initial mold to make the paper form (which I reinforced with a suitably sized metal keyring to ensure its cross-section stays circular). You can see the Vaisala nose cone, the paper form, and the completed fiberglass cone in this photo.

Then I fabricated a short segment of fiberglass tube to fit inside the body tube, and epoxied that to the cleaned-up nose cone. You can see the final result here.

This was a bit less of a hassle to make. I wrapped the actual motor in a suitable thickness of paper, then wrapped baking parchment over that, and finally the epoxy-soaked fiberglass fabric over that, followed by release and absorbent fabrics, and squeezed it all tight with a couple of meters of elastic band which I wrapped around it. After curing, I cleaned it up, cut it to length, attached a motor retainer clip (made of piano wire) with a strip of fiberglass fabric, and attached the forward end stoppers (made of flat fiberglass composite) with epoxy. The forward end stoppers also protrude outside the motor mount, centering it inside the body tube. When assembling the rocket later, I glued those to the body tube's inside wall.

The fins are made of 1 mm thick balsa sheet, laminated twice with extremely light fiberglass fabric (50 g/m², I think). Before laminating them, I milled grooves for the tripole antenna wires, and epoxied the wires in place, filling the grooves at the same time. I also beveled the edges of the fins with sandpaper. The first layer of fiberglass was open at the forward edge, though I left about a millimeter protruding forward from the balsa's edge, and the layers are nicely stuck together with no gap visible. The second layer was similarly open (but protruding a millimeter, and stuck together) at the tailing edge. The tip chord also has a millimeter of fiberglass protruding, and is nicely closed as well. I intentionally left a bit of extra balsa at the root chord, which also got laminated, and which I trimmed away last (using my Proxxon setup; there's a photo here, but I don't know if you can make out what's happening in that mess...), to leave its edge as square as possible.

Then I repeated the process for the other two fins. I made another separate jig with a 60° angle on it to place the following fins at their correct angles respective to the first (as shown in this photo—a scrap of plywood with a 60° slope, glued perpendicular to another plywood scrap). My fins have never been this straight and even!

Finally, I cemented the tripole antenna's wires, as well as a 3-wire control lead for the radio transmitter, to the outside of the motor mount. I placed short lengths of heat resistant fiberglass sleeve over the wires forward of the engine mount, to protect them from the ejection charge (as seen in this photo).

I wanted the launch lugs to be as low profile as possible to cut down on drag, and I had barely any room inside the rocket, especially between the motor mount and the body tube. So I machined custom launch lugs out of Nylon rod with my toy Proxxon mini drill. The two lugs stick out of the body tube through 10 mm holes, and are held in place from the inside, the lower one by the motor mount, the upper one by a piece of fiberglass cloth epoxied in place. The lugs fit into a 20×20 mm rail.

I made a small parachute out of silk, which I managed to dye an awful underwear pink. The parachute lines and shock cord are all 0.75 mm Kevlar. There's an additional flap of silk that wraps around the packed parachute.

I've seen steel thread of crazy strength snap when the ejection charge shoots out the nose cone. Therefore I made a lengthy zigzag of Kevlar cord onto a piece of sticky tape, which I then folded in half. This will "give" about a meter, reducing even the most violent yank to safe levels. (In the end, this wasn't even necessary—none of the cord was pulled out from the tape sandwich.)

The packed parachute was wrapped into a burrito inside the extra flap I mentioned above. The burrito and the sticky tape thing fit inside the hollow nose cone.

Slots were cut at the bottom end of the body tube, and the fin can was inserted. Fillets were made on both sides of each fin, and for good measure, an additional strip of the extremely light fiberglass (which the fins were laminated with) was laminated over each fillet.

Finally, the rocket and its nose cone were sanded, gooped up sparingly with epoxy, sanded again, and painted RAL 2004 orange.

Roope (the furry cow) can't wait for the launch event. He's going to ride onbaord my HPR L1 certification rocket The Big Cheese! (Not the supersonic Iso-Joonappi, though.)