Decoding Digital (406 MHz) ELT Data Bursts

[KB9JOZ]

I'm wondering if anyone with more technical knowledge with me knows if it would be possible to decode ELT data bursts using a computer sound card.  The specs say it is a phase modulated signal.  Specs are at

http://www.cospas-sarsat.org/DocumentsTSeries/t1Oct03.pdf

Is this the same as BPSK modulation?  If not, could it be decoded with a simple sound card interface or would a modified receiver be needed?  Just wondering if this is a feasible project to consider embarking on or if it's too complicated.  I think a program that decodes these transmissions could be extremely useful for SAR agencies.  Any comments would be great.

Thanks!

Andy
KB9JOZ

[N3ZKP]

I would venture to guess that the overwhelming majority of professional SAR agencies/groups already have the tools to do this - probably from the same people who manufacture the ELTs in the first place.

[KB9JOZ]

Actually most don't, at least not directly.  Normally the information is decoded by LEO satellites and relayed to the appropriate agency by the AFRCC.  The SAR agencies themselves usually just have equipment to home in on the 121.5 MHz analog signal that the ELTs/PLBs/EPIRBs transmit.  The systems that do decode the digital data that CAP uses (they also DF the digital and analog signals and are aircraft-mounted) are prohibitively expensive ($15000+) for most small SAR groups with limited government funding.  

Andy
KB9JOZ

[N3ZKP]

Interesting ...

You just taught me something I didn't know.

Thanks,

Lon

[AA4PB]

If the data burst is being received and decoded by a satellite can't the position information simply be relayed to the search personnel rather than having them carry equipment to decode it directly? It would seem that search personnel on the ground would have to be pretty close the low power ELT before they would be able to receive the signal well enough to decode the data.

[KB9JOZ]

That is exactly how it is done...the information is relayed to the SAR agency.  Being able to decode the information directly is actually not entirely necessary, but might be helpful in some situations, such as if the person with the beacon is moving.  That's why I'm trying to get a feel for how technically feasible it would be.  It's probably not worth putting a ton of effort into...indeed if it were I'm sure commercial vendors would be doing so by the dozens.  But if it's not too hard to do it would be a fun and useful project to undertake.  

What would be even more useful to the SAR community is a RDF system that could DF on the 406 MHz digital signal.  This is difficult because the signal is only transmitted for about a half of a second every fifty seconds.  That means that traditional DF techniques like directional antennas and TDOA systems are pretty much out of the question.  Doppler is one of the few systems that might work on such an instantaneous signal.  Of course there is the continous analog signal on 121.5 MHz to DF on, but it is only a 25mW signal.  The digital signal on 406 MHz is a 5W signal, allowing you to detect and DF the signal from much farther away.  Being able to decode the information from the signal you are DFing allows you to make sure that you're DFing the right beacon in the unlikely event that two or more beacons are transmitting in the same area.

Andy
KB9JOZ

[AA4PB]

My guess would be that it could probably be decoded with a sound card. It may not be too difficult for someone with Windows programming experience, perhaps one of the guys who has done the early PSK31 work. As I understand it, the sound card simply works as a analog to digital converter. All the actual DSP decoding algorythms are done by the main processor.

I've used DF systems that work with TDOA to DF on short pulses. The problem with the ELT is that a pulse only occurs every 50 seconds. It is not impossible, but multipath makes it very difficult to DF on short signals that only occur once a minute. If you get a reflection during the pulse, that wrong information stays until the next pulse and a minute can be a very long time to wait. For this reason I would think that the continuous analog signal would perform much better for DF, even though the data burst signal is higher power. You can also do a lot of averaging to minimize the effects of multipath when you have a continuous signal to work with.