Author Archives: Anthony Stirk - Page 5

Launch UAVA & PAVA with BUZZ + PIE1

Had a very interesting day yesterday where we launched 2 balloons with no less than four payloads. Dave Akerman and I had been planning to do a launch in 2 weeks, however permission came through and the weather and predictions all looked good for this weekend so we decided to go for it.

The launch permission was for 2 balloons and having chatted with Dave we decided to do one “pico” (light weight payload) and a second balloon would have a number of payloads on it.

The Pico balloon would run just a tracker (the board that transmits its current location) only with no camera, which seems a bit odd but I wanted to test my super light weight board “picoava” out and have a shot at the world altitude record, powered by 2 AA batteries (it will run on 1 x AAA but we thought we’d give it a little more incase it floated). It was encased in foam and weighed in at 65g :

The second flight was going to be the GoPro Hero 2 equipped µAVA payload. Dave Akerman was working on his “Pi in the Sky” payload which was designed to transmit live images back to the ground during the flight. As this was entirely untested in flight and as µAVA has never flown we agreed it would be prudent to fit a 3rd, backup tracker in the form of the proven BUZZ payload.

On Thursday night the predictions were still looking good for Saturday so we made the call to go for it, with the permission to launch appearing just after lunch time on Friday everything was set. Hell even the weather was looking promising so I drove down to Oxford and stayed over so we could get an early start.

Saturday morning didn’t look promising dense cloud cover and dull. I picked up Nick from north of Oxford who was coming along to learn the ropes/help out. On arrival at Daves we threw together a quick plan of action whilst Julie made us bacon butties + cups of tea (thanks Julie!). We were going to go with the pico payload first so we powered it up and glued the packaging shut only to find it jammed the key on the car so we had to put it in the barbecue to shield the signal to be able to unlock the car :/

As 10am came around the rain started coming down much to our annoyance, however the radar indicated it should pass within the hour which it sort of did. We headed up to the launch site about 10:50 and set up. The drizzle didn’t really stop at any point but we decided to go for it anyway. We aimed for a neck lift of 860g on the balloon which is quite difficult to accurately measure in the wind.

At just before 11:30 everything was ready and we released pAVA into the sky. We checked the spacenear.us/tracker site and within about 2 minutes local listeners had started to pick it up and were updating the location, the ascent rate seemed on track so we set about the next launch.

As the next one had three payloads on we started getting them ready, there was alot of string (in the end the distance from the top of the “train” to the bottom was about 60 meters). All the payloads were functioning and at bang on 13:00 µAVA, Buzz and PIE lurched skywards, it was a really impressive sight :


(Video from Daves YouTube Channel)
For once the predictions of landing weren’t too far away from the launch site so we headed back to Dave’s house to follow the progress :

It was quite exciting as pAVA steadily climbed and entered the top 10 highest flights in the world and carried on going :

Meanwhile PIE was returning some really great live shots via the distributed network of listeners :

Finally as pAVA crossed into the 43km “woo woo” zone (woo woo is a term used to denote we don’t quite understand it!) the ascent rate started to slow down considerably and at 43337 meters just 384meters below the current world record set by XABEN-25 the balloon gave way and it started its return to earth. This currently puts pAVA the 4th highest amateur payload in the world.

Turning our attention back to the µAVA, PIE, BUZZ train it began to approach the predicted burst altitude of 34km, which it went through and carried on going. Someone forgot to tell this balloon it was meant to burst  as it carried on climbing until eventually at a quite unbelievable height of 39994 meters (unbelievable for a 1200g balloon with 1kg of payload on it) the balloon finally gave way to the forces placed on it and our payloads began their descent back to earth.

The burst and initial decent were extremely violent with some serious vibrations going through the payloads :

Note Buzz flying past the camera :

Finally at 15:36 the payloads touched down on the edge of Milton United football clubs field :

Recovery was fairly swift as the pink parachute stuck out like a sore thumb on the football pitch :

That would have been an extremely successful day, first Raspberry Pi in “space” (no where near space but lets not get technical here), some misting on the GoPro but otherwise all payloads worked as intended.

We turned our attention back to pAVA which for some bizzare reason was coming down at an extremely slow rate. Normally the payloads descend at 5m/s (double that at 10km) yet at 20km pAVA was doing 2m/s descent. It was a very light payload but that was lower than expected.

We sat about in the nice, now sunny, afternoon as pAVA made its slow way down, the predictions had it landing not too far from Nicks house so we headed up that way. Suddenly just above the ground the descent rate increased and the payload disappeared from our radios.

Using the Yagi antennas to direction find it we could hear it but not decode it. We moved a little close to a local high point before the Yagi pulled a decodable signal in. The payload was in a field on the other side of the A34 :

With seemingly no access to the field and with the rain rolling in I decided to call it a day and headed off back up North. However Dave had other ideas and had noticed a small layby just next to the field. Ignoring the fact it was raining very heavily and the field was full of human height triffids he made two attempts to locate the payload. On the second attempt he found the cord and followed it to the payload :

Probably a little beyond the call of duty but with all four payloads recovered a great days HABing was had by all. Dave’s Pi in the Sky payload (which I’m sure will be written up shortly on his site) was a great success, interestingly those who think you need hand warmers in payloads think again, the Pi payload was incredibly hot, some of the plastic packaging had actually melted and the whole thing smelt of burnt electronics.

Obviously I’d like to say a big thanks to both Dave and Julie for letting me come down and launch with them, to Nick for coming along and helping and of course all the members of UKHAS who come out to help track these flights without whom this would be extremely hard / impossible.

Some pictures :

Just after launch note Dave and I in the field (Click for larger)

This shows the burst note the parachute top right (click for larger).

Even stills from a GoPro HD are pretty impressive (Click for larger)

Edit :

Inflation stop motion video here : http://www.youtube.com/watch?v=gloJM3O2DNo&feature=plcp

Recovery video here : http://www.youtube.com/watch?v=Ui_wejYGVu0&feature=plcp

Daves excellent write up here : http://www.daveakerman.com/?p=592

Possible launch in the next week or so

Been spending time helping the wife develop a website for her Veterinary practice ( vets in bradford, Clayton and Heckmondwike ) as well as developing a few new payloads. Keeping with the mantra small is better I have developed µAvaNut, a small NTX2/Sarantel based “workhorse” tracker. Designed to be reliable this module uses very proven configurations and components in a small form factor :

The second payload is the previously discussed µAva which is a super light weight board with an RFM22B and a chip antenna. This board has two versions, one with a linear regulator and one with a boost regulator for pico use.

Having heavily revised the code for both I’ve developed µAva code for a floater with heavy power saving code in it (Between 23:00 and 06:00 if still in the air the payload pretty much shuts down apart from small “peeps” every now and then)

Hoping to launch soon!

Original AVA Failure Mystery Solved !

My original flight AVA suffered a GPS failure which we put down to a dodgy ground on the antenna however thanks to information from the GPSL mailing list it appears the cause was actually the AEE MD91 Camera that so spectacularly captured the decent into the North Sea.

I’ve tested this evening and can confirm categorically the AEE MD91 does jam GPS at close range even on the more advanced uBLOX chips.

With the camera off :

$$AVA,6,18:42:37,53.7525356,-1.8179630,259,8,6,3,5364*C6B5
The important bit is 8,6,3 = sats, navmode , lock

Turned the camera on and placed near the tracker :

$$AVA,17,18:46:00,53.7562549,-1.8242525,1019,0,6,0,5364*C5BE

All satellites gone and the lock indicator reading 0. Yikes. I quickly wrapped the camera in foil and this does actually fix the issue but I probably would err on the side of caution and not fly this again.

So in summary beware AEE MD91 Camera does jam GPS when in close proximity.

Shiny new boards need shiny new code..

My coding is crap. The code at the core of the AVA1 & AVA2 boards was old, nasty and had some buffer overflows in it I could never track down. I’m sure the issues related to the way I was reading the NMEA data and parsing the GPGGA, occasionally and unreproducibly it would just give overflows usually when it was getting locks.

I started the code again and with the assitance of Jcoxon’s Project EURUS code I changed the way it gets the data from the GPS by switching to polling the GPS for its location. Its much more efficient, also I’ve put in checks and timesouts that were missing on the original code and moved everything to procedures.

All the complex flight status stuff has been removed and I’ve redone the temperature parts. Finally on review I’ve amended the default RTTY from ASCII-8 1 stop bit to ASCII-7 2 stop bits. Additionally I’m sending 10 nulls before commencing transmission to aid the trackers and AFC in dl-fldigi. Hopefully these changes will make a more decodable signal.

I’m really pleased how its worked out, in fact now I’ve procedurised everything it was a very trivial 10 minute job to amend the code to work with the RFM22B making both a RTTY and Hellschreiber beacon.

The code for all variants is now about 10kB as opposed to the 18kB it was previously.

Arduino compatible uBlox MAX 6 breakout board

As there seems to be some demand for an Arduino compatible uBlox MAX 6 breakout board I’ve made one up with on board level convertors and regulator to ensure your Arduino can talk to it and it gets the power it needs :

Arduino Compatible uBLOX MAX 6 breakout

These are available from HAB Supplies here and there are code samples for using with Software serial and hardware serial on the HAB Supplies Wiki.

More new hardware µAva

Following closely behind µAvaNut is µAva, a test board that replaces the NTX2 with the RFM22B also I’ve dropped the temperature sensor on this PCB. This board is extremely light at 5g and has the pads for an optional DC-DC boost converter which should be able to power the board from 2 x AAA cells. I’ve also used 0402 SMD components on this board which as advise really isn’t necessary.

JCoxon’s Project EURUS is currently testing this board to ascertain its suitability.

 

New hardware µAvaNut

Ava2 PCB worked great although it did have a few design flaws, the temperature sensor for one. Having taken inspiration from Darkside I wanted a robust but small tracker that encompassed everything I’ve learnt about making PCBs. The design criteria was as small as possible, NTX2 based with a Sarantel antenna. This unit isn’t for pico flights (although its entirely viable to use it as such as it is very light), I’ve designed it as a robust workhorse and backup tracker.

Enter µAvaNut :

As usual at its core is a Arduino Pro design with a number of changes, I have a lower drop out (250mV) and beefier regulator (250mA), Abracon 8Mhz Crystal , uBlox MAX 6 GPS and the ubiquitous NTX2 on the rear.

Hope to fly this soon

 

Ava2 Launch!

About time I updated this!

Dave Akerman had kindly said I could launch from his location in Brightwalton, the plan was to do 2 launches. The first was a “pico” flight of Dave’s “BUZZ”, a pink sputnik style ball that was very light and a secondary launch comprising of AVA and CLOUD, Dave’s photo payload.

Dave has done a great write up here : http://www.daveakerman.com/?p=441

There are ALOT of photos here : https://picasaweb.google.com/118244444241111963790

 

Project Swift – Board is Complete

As a side project to Ava I’ve been working on Project Swift with fsphil and NigeyS. Project Swift is a flight computer/tracker designed to be floated across Europe. To comply with legislation the board has 2 transmitters , the first is the license exempt 10mW 70cms transmitter for UK operation. The second which kicks in once it has left UK airspace is a 2 meter APRS transmitter.

The initial revision 1 board had a number of errors which have been corrected resulting in revision 2. Revision 2 was tested remotely by fsphil checking each component. When he was happy it was all working another board was made up and dispatched to Nigeyuk.

The boards are now with fsphil and nigeyuk for programming :

Ava2 New Box and Camera System

Been a bit busy lately and haven’t had chance to update the blog. Happy to announce the Ava2 flight computer had its maiden flight on Saturday the 31st of March. I’ll update the blog with details and pictures shortly.

Initially I was going to reuse the AVA1 container which , despite being in the north sea for 18 hours, was still servicable. However this container weighed 300g and was quite large due to it needing to accomodate the older, larger board. After a few hours of hacking in the garage Ava2 had a new packaging :

As you can see it was alot smaller than the original one. Another noticable change was the inclusion of a heated U/V filter for the camera :

Using some enamelled copper wire from the inside of an old Antec PC PSU I had about 2.5Ω with 3Vs through it. This should have resulted in about 3.6 watts of power. The camera was a Canon A710 with my previous script set to take a picture every 5 seconds.

The board is now complete (less the cut down circuitry). I’ve made up 2 boards one with a 434.075Mhz transmitter on it and another with a 434.650Mhz one.