The Joey-M flight computer was soldered up this week, all going fairly well apart from some minor rework required on some of the finer pitched components.
Joey-M uses the uBlox NEO-6Q GPS with a Sarantel SL1202 passive antenna, along with a small I2C EEPROM such that configuration can be retained when Joey is powered off. USB is broken out to a Micro USB connector such that configuration and debugging can be achieved via the excellent u-Center software from uBlox, and the configuration saved to EEPROM.
Joey-M uses the Micrel MICRF112 434MHz ISM band 10mW FSK transmitter with a twin varactor diode crystal pulling arrangement. Its 13.56MHz crystal has an 18pF capacitor on one side, and the varactor arrangement designed such that under no bias voltage, it also rests at 18pF.
One varactor has a large range of capacitance that allows the center radio frequency to be moved around by about 30kHz. The other varactor has a much smaller range, such that it can be used for very small shifts for FSK. Voltage is applied to the varactors by a twin 16 bit DAC from Linear Technology.
The radio stage of Joey-M is shown below:
To ensure a clean spectrum and thereby increasing the ability to decode, the modulator input waveform is digitally filtered in a Gaussian shape to provide GFSK. The step response of the order-50 Gaussian filter is stored in EEPROM on the AVR, and is written out to the DAC on each change of modulator input voltage. A trace of the modulator input at 300 baud is shown below:
A working firmware is currently running on Joey-M in preparation for a first flight this week. Later, temperature compensation for crystal frequency drift and more complex radio modes (MFSK) will be explored using the Joey-M platform.
Finally, thanks again to Cambridge Circuit Company for manufacturing these boards for us.
We’ve put together a concept design for the launch rail for the Martlet 1 rocket. The rail concept has a 6m truss tower, standing on a base made from standard scaffolding tubes. The tower stands on a hinged plate, that is raise and lowered using guy ropes and winches. The base is held into the ground with ground screws and has adjustable height feet to set the launch angle. The tower will have additional guy ropes to keep it straight when the rocket motor ignites. Along the side of the truss runs a rexroth extrusion that will guide aerodynamically profiled lugs on the rocket.
Today we received the Martlet 1 centering rings from our machining sponsor, Cambridge Precision. The rings are turned from Nylon 6/6. The first sets of rings are to centre the rocket motors in the motor sections. The rings also have 3 radial holes to locate the fins into, and one axial hole for a wiring tunnel, to allow wires from in front of the motor to go to the igniters.
The second set of centering rings is for the parachute sections. The large hole in the centre is for a cardboard tube that stows the two parachutes. This is because the parachutes are ejected through an orifice that is narrower than the rocket diameter. The parachutes are therefore stored in a tube of the ejection diameter, as the parachute pack could jam if it went through a contraction. The rings have various axial holes for wiring tunnels, and slots for lightweighting.
The rings are exactly how we designed them, thanks CP !
Today we headed to Cambridge Precision, our local machining sponsors. They are making all of the coupling rings for the 3-stage Martlet 1 rocket. We visited their R&D department today, to see the first version of the parts. After the anodisation process, some of the parts were no longer a sliding fit, so sadly we had to remove the anodising layer in some places.
Tomorrow we’re heading back to Cambridge Precision to finish the match-drilling and match-tapping operations on the coupling assembly and bulkhead assembly, which links the motor and parachute sections (as shown below).
More photos will come in the next few days, as the parts are completed. Many thanks to Colin, Pete and Steve over at Cambridge Precision R&D, who have put these parts together so accurately for us.
The carbon fibre airframe arrived today for the airframe. It is 118ID 122OD, made from 50% axial 50% circumferential uni directional per-preg, with a working temperature of 180 degrees C. The manufacturing process uses spiral wound heat shrink trape, so the surface has been ground smooth to get rid of the spiral ridges. They’re still there, but should polish out nicely.
Roughly putting the 6 sections of the Martlet 1 rocket end to end (3 x motor sections, 3 x parachute sections), we’re finally able to see quite how large it is……
When the fins are bonded into the carbon fibre airframe, some extra carbon fibre will be laid up tip-to-tip to strengthen the fins. We wanted to practice doing this lay-up and make sure that the C-F fabric that we cut is the right shape, so we made a mockup of the fin can.
A quick lay-up with some unidirectional C-F fabric and some fibreglass on the inside gave us a tube of the correct diameter for the airframe (quite large!).
A quick effort with a hacksaw gave us so mock fins made from plywood, with the edges near the body tube filed to the same angle as the machined aluminium edge will be. An alignment jig was knocked up on the CNC router to keep the mock fins at 120 degress for them to be bonded to the C-F tube. The mock-up will be tested with a lay-up later this week.
Last weekend saw the 20th launch in the CUSF Nova balloon project, carrying two payloads – the experimental Wombat flight computer, and the Squirrel Android flight computer.
We also used for the first time the excellent F-203AV mass flow controller (kindly on loan from Bronkhorst who support CU Spaceflight) for regulating the helium filling of the balloon.
Radio transmission from Squirrel failed on launch due to what turned out to be physical damage to the connection between the phone and the NTX2 FM transmitter, but fortunately the payload was still able to be located with the SMS location tracking (landing in a field near Bishop Stortford – Google Maps). The balloon reached a maximum altitude of 22km, and a large number of photos and videos were captured – see the Squirrel 3 Flickr set for them all!
The Wombat flight computer was a test payload, and unfortunately had some difficulty keeping PLL lock during the flight, making it very difficult to track. We’re now refining the radio design for the next launch.
Here’s a video of the payload tumbling at around 22km altitude just after the balloon burst:
We got the PCBs for the Joey-M flight computer! These are experimental flight computers using the Micrel crystal pulling radio (see an earlier blog post), along with an Atmel AVR and a uBlox NEO-6Q GPS.
Thanks very much to Cambridge Circuit Company for manufacturing these PCBs for us!
Martlet 1 has 3 stages, each stage having a motor section and a parachute section. Each parachute section holds two parachutes: a small spring pilot parachute that is deployed first, to allow the rocket to descend quickly so it doesn’t drift far in the wind, and a large parachute which is deployed near sea level, to allow the rocket to land softly.
In each parachute section, there is an aluminium plate to retain the parachutes when they are packed, and the spring is compressed. The plates have a large hole in the centre to allow the main parachute to connect to an eyebolt in the bulkhead, and they have three holes near the outside for wiring and riser tunnels.
We machined the plates out of 3mm aluminium plate, using our CNC router.
The plates were then cut out and the inside edges hand smoothed, so that they are not abraisive on the kevlar parachute risers.