MISSION COMPLETE Pegasus III was launched at 7.38 (BST) from Cambridge and reached an altitude of 19,495m (63,959 ft) after 134 minutes of flight and then landed at 10.29 (BST) travelling 35 miles. Took 200 pictures from a side mounted camera
More complicated then Pegasus I:
/--(ttyS2)--< GPS (ublox LEA-4H) | |Gumstix (connex 200xm and GPSstix) | |--(ttyS0)--->Nokia 5110 --SMS-->Base | |--(ttyS1)--->Aerocomm 868mhz radio modem --->Receiver | |-->Camera shutter | \-->Cutdown circuit
After the loss of Pegasus II Gumstix were very kind to send me a new gumstix connex and also one of the first GPSstix with a ublox LEA-4H gps onboard (16 channel, 4hz update) which is communicated with through /dev/ttyS2 (STUART). All that is required is for the channel speed to be set:
stty -F /dev/ttyS2 speed 9600
Then:
cat /dev/ttyS2
will show a stream of data. However without an external gps antenna you'll struggle to get an sort of lock. I therefore went with an active antenna which attaches to the SMA connector on the GPSstix. However the default set up of the gps module is for a passive antenna so a little hack is required (see here).
The GPSstix board also breaks out lots of different lines including a further 3 TTL serial ports and lots of lcd pins (which can also be used as GPIOs). Instead, like in Pegasus I, of having to convert up to rs232 voltage levels only for the circuitry in the phone cable to the drop it back down to TTL levels I removed the DB9 serial connector and directly wired the phone connector to the GPSstix. All that is required is Rxd, Txd and Gnd and while gnokii, when initialised, attempts to set RTS and CTS to power the usual conversion circuitry it happily communicates with the phone (The Nokia 5110 from Pegasus I). Actually it is more reliable probably as it doesn't need to power the conversion circuitry. I attached the phone to ttyS0 (FFUART) (after setting up usbnet and disabling getty in /etc/inittab. The only issue here is that on turning on the gumstix if the phone is attached it is necessary to have the phone on. I believe this is due to uboot receiving a “keystroke” from the phone resulting in it to drop in to the uboot shell instead of booting.
The radio modems were attached to /dev/ttyS1 (BTUART) and while I expected to need to link up RTS and CTS in the end all that was required was Txd Rxd and Gnd. I used the modems as a serial port extender and used a script to push data directly through the serial port with a console on the other end to receive it (such as kermit) both running at a speed of 57600.
stty -F /dev/ttyS1 speed 57600 tail -n 1 /gps.log >/dev/ttyS1
On the base side (my laptop (Mac OS X on a 800mhz G4 iBook)) I used kermit:
kermit -l /dev/cu.usbserial
set speed 57600
set carrier-watch off
connect
To trigger the camera to take pictures I again hacked the shutter button and by turning a GPIO on and off I was able to make the camera take a picture. I was able to turn off power saving and also the cameras LCD to save power (and managed to get 206 photos off it.)
The cutdown system followed the one described on the UKHAS wiki with the actual cutdown device being providing by Steve (rocketboy). The power supply came off the flight computers battery pack in parallel with the gumstix. The cutdown device itself was placed above the parachute which when activated will break its plastic tube releasing the balloon allowing the parachute to open and descend.
This was built by Steve and consisted of a power supply (6v photo lithium battery), a pic and a 434mhz 10mW radio transmitter with a 1/4 wave vertical antenna. It was set up to produce a Morse code signal of “Callsign - Pegasus III” and then 10 seconds of a constant tone. The antenna was mounted on the bottom of the payload spread over a 250mm x 250mm bit of cardboard (4 layers thick) with the centre section hanging down.
Went for a different approach for Pegasus III, carving polystyrene is a difficult job with out a hot wire (which i still don't have) so instead we went for layers of polystyrene which have areas cut out for the different sections and then are glued together and then the edges are sealed with duct tape. We used ceiling polystyrene tiles from a local DIY store.
The payload case had a bottom section which contained most of the components and also a lid which had the gps antenna, 868 radio antenna and the cutdown circuitry. This was sealed with duct tape and also large Velcro straps at right angles to each other which was to what the parachute and balloon were attached. Parachute
A 36” parachute was used like Pegasus I with an identical setup of the parachute being pre-deployed and resting on a bottle top on the balloon/payload cord.
I changed from using GPSD (as it struggled with the 4hz refresh rate of the gps module especially not showing the altitude) to a perl gpsparser and logger from defy and also used his logrotate script which every 10 minutes changed the log and gzipped it to save space. The reason for not creating a large logfile is that the jffs2 filesystem chokes on large files (over 500kb) if not correctly powered off. While the filesystem is fine the boot up check thinks it is corrupt and drops into the uboot shell. I couldn't risk this occurring if mid-flight there was a reset so settled for the logrotate script and since then haven't had a problem. Phone
Identically to Pegasus I I used gnokii with a simple script which took the last line of the /gps.log and sent it via sms every 2 minutes (instead of 1 minute like on I)
tail -n 3 /gps.log | /root/gnokii –sendsms xxxxxxxxxx
As the 868mhz module is only allowed to transmit on the ground I set up a script that would transmit for 10minutes on being turned on (allow me to check everything was working) and then sleep for 3 hours and then turn back on again and carry on transmitting
See construction section for more details
I wrote a few perl scripts (see here) with the conditions of:
The cutdown circuitry is activated by turning a GPIO on for 2 seconds.
After learning from my mistakes during the launch of Pegasus II, the launch day was relatively easy. As long as I turned everything on in the right order setting up the payload was easy. We (Carl, Henry, Laurence, Steve and myself) all met at around 6.00am at the launch site and began setting everything up. While I prepared the payload and sealed it up Steve put together the parachute, balloon cord and cutdown device while the other three filled a 500gm balloon from out T tank of helium from Party Days. Once everything was set up, text messages were coming through (with varying gps co-ordinates (not like Pegasus II where they were exactly the same), the 868mhz was transmitting data, we had heard the camera take a picture (it produces a fake click) and the 434mhz was beeping reassuringly we let it go! We had calculated that the balloon need to lift 1.2kg + 1lb of lift to give us a good ascent rate. However the ascent rate was much lower at about 100m per minute (according to the three text messages we received). We watched Pegasus III drift into the distance till we couldn't see it then did some directional checks with Steve's yagi antenna. Afterwards we set about launching MiHAB 1.
At about 9.30 during our retrieval of MiHAB we received the first text message give us the location of Pegasus III. After relaying the information to Steve who had been chasing with his radio equipment we returned to Cambridge ate breakfast and then went south on the M11. We meet Steve and my parents at the field we knew it was in (we had previously looked up the position on google earth) and it took very little time to find it.
It had landed upside down below some very large pylons (it might have bounced off the wires!) in a barley field. The cutdown hadn't activated and instead of the balloon shredding as it should have it had torn and so we retrieved most of the balloon as well.
Pegasus III turned out to be a very successful flight and we learnt some very good lessons:
Possible improvements from the flight: