Aussie RocketCam Model Rocket Onboard Inflight First Person Video Camera link via 2.4GHz Microwave Video Transmission | Construction Details |
Intrigued by the possibility of transmitting live video and audio from a model rocket to the ground?
Me too! Scroll directly to a list of inflight first-person FPV video clips
or read on for the full Aussie RocketCam story.
The construction page has detailed descriptions and photographs of the construction of the onboard inflight video downlink system.
March 2009 I have updated the technical specifications of the CCD camera. This is the second camera I have flown as I loaned the first one to David Boyd when he had a last minute camera failure just prior to launch. Unfortunately David did not recover his rocket for some months after launch and upon testing the camera the focus seemed a little soft to me so he purchased it off me and continues to fly with it. I always suspected that the new camera I bought did not yield the same quality pictures as the original and I have just noted that the new camera only uses a 1/4 inch CCD instead of the 1/3 inch of the first camera. I'll have to try and track down the specs of the original camera and see what the claimed resolution is.
May 2007 At the invitation of the Wireless Institute of Australia (WIA) I gave an after-dinner presentation on my model rocket video work at the WIA National Conference held in Parkes, NSW over the weekend of May 4 and 5 2007. The weekend included a tour of the CSIRO Parkes Observatory, also known as "The Dish". The Parkes radiotelescope participated in the relay of live television coverage of the 1969 NASA moon landing and is an active radioastronomy research instrument today. The Aussie Rocketcam transmits in the S-band and so a photo of me posing with the S-band feed for the Galileo deep-space probe to Jupiter that was fitted to the 64m Parkes antenna was irresistable. Galileo transmitted using 15W into a 7dBi antenna at 2.3 GHz. CCD camera technology (as used on the Rocketcam) was one of the spin-offs of the Galileo mission. Galileo's CCD camera had a resolution of 800x800 pixels.
February 2007 Estes have jumped on the rocket video bandwagon with their Estes Oracle Digital Video E2X Rocket. It records on-board video from a tiny onboard CMOS camera that is integrated into the plastic nosecone. The resolution is 320x240 pixels ie QVGA and the .AVI video files are retrieved via USB. The reviews on Amazon are reasonably positive but the video resolution of the Estes Oracle camera is lower than my CCD bullet camera (512 x 582 pixels in PAL video mode) but you don't have the losses associated with the analog transmission of the video back to the ground (earth) receiving station. The worst as aspect is probably the lack of onboard sound recording and the need to take your laptop out to the field. Check out the Estes Oracle at Estes Oracle at Amazon.
March 2005 I've added a discussion (below) on the degree of the doppler shift in the frequency of the transmitted 2.4GHz video signal as received on the ground.
Oct 2004 The archive.org freecache system has been deactivated and I am now using NYU's Coral CDN (Content Distribution Network) system to handle all the video traffic from this site. Freecache did a great job, handling the Slashdotting of the Aussie Rocketcam site surprisingly well, although many users were unable to download the videos right away.
April 2004Altium have licensed Aussie RocketCam inflight footage for use in tradshow presentation and promotional DVD for their Nexar FPGA development system.
October 2003 Rex Ridenoure from Ecliptic Enterprises Corporation politely let me know that "RocketCam" is actually an Ecliptic Enterprises trademark. So for information regarding the Ecliptic Enterprises RocketCam(TM) systems for rockets, spacecraft and other remote platforms please visit Ecliptic's website at www.eclipticenterprises.com.
The Video System
In September 2000 I purchased a microwave video transmitter / receiver
system and a CCD bullet (or lipstick) camera from local electronics shop Radio Parts. These AV (audio-visual)
transmitters are generally used for domestic video extension or security applications. The lipstick camera itself,
which is housed in a stout cylindrical alloy case, would most commonly be employed in a security monitoring application.
Various cameras were auditioned and the original camera, with a 1/3 inch Panasonic CCD chip, was selected for its image quality, dynamic range, high resolution and effective auto-exposure system. It is heavier, larger and less convenient to mount than the
alternative "board camera" designs, but I am prepared to wear those disadvantages in order to obtain superior pictures. There is a lot of time and money involved in a rocket launch so you may as well go for the highest possible result.
When choosing a CCD camera compare the image quality, paying particular attention to video noise, level of detail (resolution) in the image and whether high contrast lighting conditions create an exposure problem for the camera.
The current video camera has the following specifications.
Pick up Element
1/4" Panasonic Color CCD image sensor
512(H) x 492(V) NTSC
512(h) x 582(V) PAL
380 TV lines
2 lux / F2.0
More than 48dB (AGC off)
1/50 to 1/100,00 sec
Auto white balance
Standard Board Lens
f3.6mm / F2.0
Inflight rocket video downlinks are often accomplished using mirror systems (I believe the original Super-8 film format Estes Astrocam did this) however it was a design requirement for my Rocket Camera system that the camera have a direct view of the ground for maximum image quality and mechanical simplicity. The construction page Rocket Camera construction page has many photos of the video camera, mounting arrangement and other electronics in the payload.
The video transmitter is branded "Copland" but is made by Racewood. It can transmit on one of the four frequencies listed below. These frequencies fall within the 2.4GHz to 2.5GHz (or 13cm wavelength) ISM (Industrial Scientific and Medical) band in Australia as defined in the Australian Radiofrequency Spectrum Plan. Note that many imported video sender transmitters will exceed the maximum legal EIRP authorised under LIPD Class License 2000:
Ch1 = 2413 MHz
Ch2 = 2432 MHz
CH3 = 2451 MHz
CH4 = 2470 MHz.
There appear to be no standards for positioning these video channels within the ISM band and other 2.4GHz transmitters available in Australia use slightly different frequencies that might be + or - 2MHz on each channel. Interoperability between systems is further complicated by the use of different frequencies for the audio-subcarrier and different numbers of audio channels: ie 1 for mono and 2 for stereo.
Australian ACMA regulations limit the maximum power output of all transmitters between 2400MHz and 2.4835MHz to 10mW. "video sender"/transmitters are not(as per 2008) specifically mentioned in this band and so therefore are captured by the 10mW EIRP limit see www.acma.gov.au/WEB/STANDARD/pc=PC_1278#video. The allowed power outputs are even lower at lower frequencies where video senders are specifically mentioned. This is regulated under Radiocommunications (Low Interference Potential Devices) LIPD Class License 2000.
Because higher power outputs are legal in this band for (narrowband) telemetry and "telecommand" ACMA is intending (via a variation to the class license) to specifically clarify the meaning of telemetry and telecommand control to ensure that people don't act outside the spirit of the legislation by simply overlaying low datarate on-screen telemetry (ie GPS data) over a high power wide bandwidth FM video transmission which would travel very significant distances from a rocket at altitude. A typical WiFi card puts out 30mw with some cards offering up to 200mW. Your mobile phone, cellphone or handphone (if you're in Malaysia) develops several Watts.
Geek note: 2.4GHz AV senders use FM modulation for the vision carrier unlike "terrestrial" (ie not from a satellite) analog broadcast television which is AM modulated. Both systems use FM modulation for the audio.
Proof of Concept
I decided to use off-the-shelf video hardware so that I could quickly establish whether the concept of live video transmission
from a model rocket was viable. At the time I didn't know a lot of radio theory and was concerned about issues such as:
the limited range of this very low powered system (10mW output power)
the effect that the rapid motion of the rocket might have on the signal ie
doppler shift or other effects. (Update 25 March 2005). The magnitude of the doppler shift would cause at most a 2KHz drop in the received frequency. Obviously not a problem for a wideband mode like FM TV (could it be cause of problems for the narrower bandwidth AM audio-subcarrier?). This estimate of 2KHz of doppler shift is based on the antenna being situated at the launch location and the rocket travelling at 1000Km/h which is somewhat faster than the video rocket travels with its relatively heavy payload and low impulse mid-power rocket motors).
For a discussion of doppler effects see Tony Langdon VK3JED's site.
the attenuation of the signal cased by the airframe when it was between the
transmitter and receiver as the rocket rolled during flight
the impact of
the changing orientation of the transmit antenna with respect to the receive antenna as the rocket
arced over at apogee. (I was particularly keen to capture parachute ejection and was not disappointed once this
was finally accomplished. I love watching the parachute unrolling and inflating in slow motion as the
horizon drifts into view behind the rocket.) The deep blue of the sky suggests that the camera might
have a polarising filter.
The Launch Vehicle
The mass and bulk of the components selected for the video link required the use of a larger rocket than could
be safely launched with standard Estes A to D black powder motors.
Fortunately at around the same time a friend imported several Aerotech ARCAS
mid-powered rocket kits from an American online hobby retailer. The diameter of
the ARCAS was just sufficient to house the transmitter circuit board if it was removed from its case (see the construction details page). The bullet camera is mounted on the
outside of the airframe, parallel with the body tube, pointed directly downward and
protected via a streamlined nacelle carved from a block of balsa.
The original video signals obtained during early flights were of high qualty but contained extended dropouts and are not currently available on the site.
The Research Goals section further down this page discusses the problem and some solutions.
Watching the inflight model rocket video clips
I suggest the launch to apogee clip in slow motion as a starting point ie: serpentine5_slo (available in either QuickTime .mov or Windows Media .wmv format) as you can enjoy the flight details including
the graceful parachute deployment. If you watch carefully you can also see the engine cap flying off as the motor fires.
Extended footage of the rocket under canopy after parachute deployment is available in the long versions of each clip (ie serpentine5_slo_long.wmv) but may not be worth the extra download time for those with limited bandwidth. Under canopy the upper section of
the rocket is spinning almost horizontally. However enthusiasts might enjoy stepping through the footage
frame-by-frame for nice glimpses of the paddocks of country Victoria.
If you have time to download more than one clip then the normal speed clips give a realistic sense of the incredible acceleration of a rocket flight and the
howl as the rocket accelerates is amazing.
Scrubbing backwards and forwards through the clip reveals interesting things about the flight
of the rocket that you don't notice from the ground - for example that the direction of rotation reverses. Stepping frame-by-frame through the footage gives one time to
linger over the details on the ground or see the engine cap.
Windows Media Player doesn't seem to allow stepping or scrubbing so I recommend QuickTime as the media playback client, so download
QuickTime and then grab the QuickTime (.mov) versions of the clips because if you only ever play the footage straight
through you're not getting the most out of it.
An added bonus is that the slow-motion QuickTime movies are no larger than the normal versions. To create slow motion WMV's I've had to pad the movies out with duplicate frames.
For the first time I have successfully captured continuous footage of an entire flight including boost, coast
and 'chute deployment and some recovery - a very pleasing result accomplished by using a helical
antenna to boost the signal strenth at the ground station. To cover momentary dropouts footage from
a completely separate secondary ground
station using a yagi antenna was intercut after the flight and interestingly, appears somewhat brighter
and contains artifacts including reflections which create the opportunity for further research.
The normal speed clips show a complete flight including recovery - I make a guest appearance as the rocket is returned to the primary downlink station. It took three attempts to ignite this old White Lighting reload and the slow motion clips are particularly cool as we see one of Troy's "special" igniters finally do the trick.
The primary downlink station was running a Mini-kits EME103 20dB 2.4GHz pre-amplifier. This did not make any obvious difference to the range of the system - it has been suggested this is because I am already on the receiver's "noise floor".
Email Address (Comments welcome)
Capture continuous footage by experimentation with transmit and receive antennas, ensuring there are no intermittent connections, looking at
transmit amplifiers or receive preamps and investigating whether the camcorder being used to record the output from the
receiver is taking longer than one might expect to recover from a momentary loss of signal.
Note June 2002: This goal has been attained through higher gain helical and yagi receive antennas.
Although the signal loss in earlier flights appeared to coincide with burnout, suggesting an intermittent electrical connection associated with negative g this does not, in fact, appear to
be the case. Instead it would seem that the signal was simply lost as the pattern from the two omni-directional antennae were no longer coincident.
Once the rocket had attained a certain height the vertically polarised receive antenna was effectively in a null underneath the vertically polarised transmit antenna.
This may also explain why the signal was regained at apogee.
Other things to sort out
Why the audio is still dropping out occasionally - perhaps the problem is not electrical at all but the mechanical effect of acceleration on the electret microphone.
Why no audio at all was recorded by the second ground station - does it use a different audio subcarrier ?
Why the transmitter stops transmitting when sitting on the launch pad and need a thump to get it working again (does it actually drop out or does the A/D converter or DV codec in the camcorder go to sleep ? (I need to build a VDA to check this or incorporate an S-meter into the receiver).
Thanks to Julian for his invaluable help including aiming the helical at the rocket during flight.
(Darren, Jason and Julian and myself are members of Melbourne Wireless, a group who, like many others around the world, are building a community broadband wireless network using
802.11b IEEE standard WiFi equipment which shares the same unlicensed region of the spectrum as this particular low-powered video transmitter).
Additional inflight footage was recorded independently by David Boyd of Tripoli Australia member using a 12dBi yagi antenna and their
own microwave band video receiver and camcorder. David's recordings were intercut with mine using Adobe Premiere and the resulting clips were
compressed in Media Cleaner Pro 5.
The Aerotech Arcas launch vehicle was supplied and constructed by Dave and Cath.