FHALP-3

Summary

FHALP-3 was launched on 26/04/08 from EARs in Cambridgeshire, shortly after launch the flight computer died but we are able to track visually and also by the radio beacon for approximately 50 minutes until all contact was lost. However on landing the flight computer turned back on and transmitted its co-ordinates and were able to retrieve it from 52.332,0.888

Picture downloaded from the camera can be found on flickr

Plan

Theory

Flight Computer

  • FHALP-3 will be based upon the 'success' of FHALP-2, it will continue to use the GM862 module as the flight computer however it'll be the basic GM862 module rather then the module with the inbuilt GPS. Instead a ttl GPS unit that isn't altitude limited will be worked and will be attached to the serial port that is freed up (the GM862 has 2 serial ports).
  • The flight computer will log

Communications

  • GSM this time will work - if it had worked then we would have certainly have found FHALP-2. The code will be pretty much the same, an sms every loop and the flight computer will log the reception throughout the flight, if it loses reception and regains it on landing then this could be used to trigger sending sms of coordinates instead of continuously attempting to send the sms.
  • Radio using a radiometrix module while will transmit coords as numbers via a GPIO, will be powered from the main power supply.
  • Radio beacon on a separate power supply, the radio from FHALP-2 but it won't be connected to the flight computer and instead will transmit a 1hz pulse on 434.075mHZ

Cutdown

  • Same as on FHALP-2 but will review cutdown code, will also be intergrated into the actual payload

Camera

  • Seperate container, will check the light sensor to try and increase the shutter speed and therefore have less blurry pictures.

Parts

Got

  • Radio beacon - FHALP-2
  • 2000mAh LiPo - FHALP-2
  • Cutdown circuit - FHALP-2
  • TTL Serial GPS - from ebay £16.00
  • GM862-QUAD-PY (non-GPS) - Sequoia Tech £77.30
  • Sparkfun GM862 breakout board - robosavvy £18.00
  • Radiometrix NTX2 Radio Module - farnell £19.00
  • GSM antenna (ANT-GSMQB-MMCX) - farnell £5.50

Construction

GPS

Instead of using a GM862-GPS module which has a SIRF III chipset and therefore a 24km altitude limit the plan is to use a GM862 module with a separate GPS module. There are a variety of GPS modules available which run at TTL (the same voltage levels as the GM862 therefore removing the need for additional voltage shifting) however they are either SIRF III or ublox and so require a PCB to be made. Instead the plan is to use an older GPS module which doesn't have this altitude restriction but there are very few off the shelf TTL modules.

Luckly a ublox RCB-LJ gps module and board came up on ebay, its ripped out of a dev system but has ttl inputs and is happy to run off the LiPo. One thing that needs to be done is that you need to provide a voltage source for the active antenna or it won't run, this again came from the LiPo power source.

  • Pin 1 - Active antenna voltage - connected to LiPo via a 10ohm resistor
  • Pin 2 - Vcc - directly attached to the LiPo
  • Pin 11 - Serial Port 1 Tx - connected to the GM862
  • Pin 20 - Ground - LiPo

Radio Beacon

Have finished the radio beacon, it uses the radio module from FHALP-2 (Circuit Design CDP-TX-04S-R 434.075mHz) and is attached to a CMOS 555 timer which has been set up to work at approximately 1Hz. It runs off 2 AAA lithium batteries (separate power supply so that if the main flight computer fails it'll continue to work) and the circuit is mounted on the back of the battery holder.

Parts

  • 1 x CDP-TX-04S-R 434.075mHz Radio Module
  • 1 x CMOS 555
  • 2 x 2200 ohm resistors
  • 1 x 220uF capacitor
  • 1 x Red LED
  • 1 x Double AAA battery holder (with solderable tabs)
  • Some veriboard

Pictures soon

Flight Computer

After returning my original GM862-QUAD-Py module I have finally managed to set up the replacement module with my development system. The GM862-QUAD-PY modules are not just GM862-GPS modules without the GPS modules but mine have different hardware and therefore different firmware. You can check to see whether you have the old or new versions by checking the firmware using the AT command 'AT+CGMR', the old version is 0.6xx while the newer version is 0.7xx.

While the modules are claimed to be pin backward compatible there are still some differences:

  • The GM862-QUAD-PY serial ports are more sensitive to voltage then the GM862-GPS, you have to run the voltage shifting off a 3.3v supply, 3.8v from the LiPo is too high and the 2.8v supplied from the module itself (PWR_ON) isn't sufficient.
  • To use the commands AT#RSCRIPT and AT#WSCRIPT it is necessary to use hardware flow control (unlike on the GM862-GPS) however it is possible to cheat a little, by connecting the CTS and RTS pins on the module side it tricks it into thinking it has flow control and will happily send and accept data.
  • Upload also will only work at a baud rate of 115200.
  • The module has real issues with uploading scripts of more then 3000 bytes and doesn't seem to accept them. therefore I've put nearly all the functions into separate libraries that can be pre-compiled and loaded.
  • If you run a script that crashes the GM862-GPS drops back into AT command function while the GM862-QUAD-PY just stops - its really annoying if you are uploading and testing lots of scripts!
  • There doesn't appear to be a SER2 module onboard - my basic script crashes if I import the SER2 modules, no idea why (I've got the latest firmware). Therefore to access the GPS I'm going to use the Rx line of the main serial port - its just a matter of swtiching between 115200 and 4800 for the script and AT commands respectively.

Radio Module

The radio module i'm using is a Radiometrix NTX2 which is being used to transmit Morse. The radiometrix module can be used for long range RTTY tranmission with the correctly programmed circuit so is future proof. The radio module can be turned on and off using the 'En' pin allowing it to be continously attached to a power source. The module transmits at a frequency of 434.650MHz.

The module has 7 pins:

  1. RF GND → Aerial GND
  2. RF OUT → Aerial
  3. RF GND
  4. En ← GPIO 5
  5. Vcc ← Main power
  6. GND → Main power
  7. Tx ← GPIO 3

The module has been placed at the back of the board and directly wired in, with the wires travelling underneath the breakout board. Its been tested and works.

Cutdown

The cutdown is the same charger circuit as was used in FHALP-2 but this time is being mounted internally. It has 3 wire, Vcc, GND and a GPIO trigger input. For more information see the ukhas wiki.

Launch

Right so the launch…

Actually the payload itself was suprisingly easy to set up, on arriving I attached the freshly charged LiPo by its PP3 connector which powered everything up including the GPS but didn't switch on the gm862 module. This allowed the GPS to sort itself out, to check everything worked I turned on the module, it transmitted some radio and sent a text. I then just turned it off and set up the flight train while the others inflated the balloon. After inflating one balloon to find it had a hole we moved sites to get out of the wind, inflated the balloons attached the payload and rather chaotically let go.

The payload was transmitting coords for about the first 3 minutes of the flight and then it died but the radio beacon continued though this stopped after about 30 minutes. I got one text at 347m altitude which is normal for a flight. We were able to track it visually for a good 50minutes until it disappeared into clouds.

After returning for brunch we were just going to find an internet connection to run the forecast modules when sms started to arrive reporting the correct position so after a bit of google maps we jumped in the car drove over to the landing site and didn't have too much difficulty finding the payload though it did require a bit of rally driving down a bumpy track. It was as a whole a pretty successful flight though certainly not perfect.

Once I got back I powered the module up and extracted the logs - there was nothing for the flight, it had died shortly after take off and restarted on landing, also some data that we collected at launch wasn't there suggesting that the logging to flash overwrote it. The pictures again were blurry and cutout probably at apogee.

Flight Info

  • There are no useful logs - however we recieved a text 1 minute after launch at an altitude of 347m and we launched from 48m altitude so the ascent rate with 2 balloons was approximately 300m per min.
  • The camera took 167 photos, the launch was at photo 25, at a rate of about 1 every 17 seconds the top picture was after 58 minutes.
  • Therefore 300m per min * 40 mins = 12000m - so that the rough altitude of the final picture. This was when the 1st stage balloon was set to burst and the shockwave probably knocked out the camera.
  • The reason the parachute appears to hang down so much could be due to the 1st stage balloon having enough lift that the 2nd stage balloon is dragged behind, the line must have been long enough for the 2nd stage balloon to be below the payload.

Lessons

* GM862 - While the concept is good in practice the Telit modules drive me crazy,

  • having to upload each script by serial and then either having no debugging info or having to shift to the second serial port makes development really slow.
  • The QUAD-PY version is really bad, if anything go with the GM862-GPS as at least you get the newer firmware and software libraries.
  • They don't really work - I'm sure in AT command GSM version they are fine but I've become increasingly frustrated with them as a system, they haven't really worked at all…
  • Camera - The camera is a disappointment, on the ground its nice with good optics and an easy system to hack however it doesn't seem to be able to cope with the swing around - I'm going back to the cheapy Premiers which get the job done.
  • A Radio beacon is a must, if not we would have lost contact from the beginning, instead we could listen to it for quite a while. That said it fell off at one point and wasn't at the landing site.
  • On retrieval the payload's main antenna had broken off, it gets weakened as its on the bottom of the payload and when the payload is left on the ground gets bent. Either it needs to be strengthened or the design needs to be changed.
  • The good news is that we got the payload back which means that there are lots of spare parts to be used again!.

Old Info

SIRF II GPS

After a bit of research (Deconstruction of a Pretec CF GPS and that the drivers required are just for the CF→serial adapter so that it appears as an additional serial port (an old hack I did for Peg II on the gumstix)) I found out that Compact Flash GPS units are just TTL serial GPS modules with a CF → Serial adapter in between.

  • Get a unit (£16 off ebay),
  • take it apart,
  • separate the GPS module from the adapter
  • Identify the wires - on my module:
    1. 3.3v
    2. Tx
    3. Not sure
    4. Not sure
    5. Not sure
    6. (Grey) GND
  • Power up and start reading the NMEA (Remeber that you can only put 3.3v, and more and it might go bang!)

I'll add pictures soon (once I get my xD card back.)

Further investigation has suggested that SIRF II GPS units don't work above the dreaded 18km altitude limit which means that using this gps is as good as using the built in SIRF III module on the GM862-GPS. One GPS unit that does work above 18km is the Lassen IQ module which can be bought from Sparkfun and Diamond Point - will investigate further…

 
missions/firefly/fhalp-3.txt · Last modified: 2008/05/03 20:15 by jamescoxon
 
Except where otherwise noted, content on this wiki is licensed under the following license:CC Attribution-Noncommercial-Share Alike 3.0 Unported
Recent changes RSS feed Donate Powered by PHP Valid XHTML 1.0 Valid CSS Driven by DokuWiki