[lug] Wireless router question (OT)

[Nate Duehr]

Stephen Queen wrote:

>>This may sound silly but 30m is well within Ethernet specs -- why not
>>just run a cable between these three nodes?
>>    
>>
>
>As the helicopter lifts off, the system is not moving
>forward, so it has no aerodynamic stability. What usually
>happens then is that the receiver and transmitter assemblies
>start spinning. This twists the cable. This in effect causes
>the cable to become a lot less reliable.
>  
>

These guys deal with twisting cables all the time... www.eoss.org.  
Might look at some of their photos of cable assemblies for ideas.  Some 
of their payloads also get to spinning on the way up and induce 
centrifugal force at the end of the payload string -- meaning that their 
interconnecting cables/wires must be able to withstand quite a bit of 
torture.  Especially on the way down in free-fall after the balloon bursts.

>Actually, plugging in 3 AP's, 2 of which are setup as
>infrastructure clients is simple compared to rigging an
>ethernet cable between everything. It has made the system a
>  
>
Yeah, but it's not working.  ;-)  (Always the pragmatist in the crowd...)

Seriously, 30m is child's play for 802.11b.  You have some other source 
of interference or you've completely goofed up RF connectors or 
something.  I could do the math, but at 100mW, 30M away even with a 
zero-gain antenna system you should have rediculous amounts of S/N 
margin.  Probably in the range of 60dB or better.

Are the AP's below the helicopter talking just to the AP in the 
helicopter?  If so, small panel (flat) antennas laid horizontally on top 
of your payload (if it's big enough - I haven't seen your payload 
strings) would direct the vast majority of the RF signal in an upward 
direction and also hear better because they wouldn't "see" 802.11b 
stations on the ground nearby. 

Are you flying in densely populated city areas or close enough to be 
line of sight to one?  If so - your problem is probably on-channel 
interference from existing AP's. 

You could probably easily find a young RF engineer with access to a 
spectrum analyzer and a kludged together 2.4 GHz antenna that'd be more 
than happy to go for a helicopter ride, if you want to be scientific 
about it and really get REAL data as to why your AP's aren't working. 

>As far as the economics, the AP's are negligible compared to
>the cost of helicopter time.
>  
>
No kidding!  You might want to "fly" the system from a tall building or 
balloon system a couple of times before the next expensive helicopter 
flight to work the bugs out of it. 

Oh yeah... one other (duh) thought... if the antenna for the AP on board 
the Llama is inside, just get some quality connectors and coax (cheap 
crap doesn't work at 2.4 GHz when you're doing REAL work with it... look 
at the loss numbers for whatever you use and build accordingly... coax 
in general is bad at GHz frequencies, but hardline is a pain to use in 
mobile applications)... and dangle the antenna out the door so it can 
"see" the payload string.  Something omnidirectional or as close to it 
as you can get would be good so it can "see" no matter what direction 
the wind dangles and twists it.  That or temporarily mount a small yagi 
pointing down at the payload string - but at these close ranges, that 
seems to be HUGE overkill. 

I think you're getting hammered by other signals in the 2.4 GHz range.  
Cordless phones, 802.11b, everyone and their grandmother is there now.  
You might have more luck with 802.11a cards at 5.8 GHz...?

How much data really needs to go between the payload and the helo?  
Could you switch to something like a low-speed data system that uses the 
300 MHz band ISM non-licensed radios? 

Also did you ever consider using thinnet ethernet (via co-ax) for the 
cable and not UTP?  Certain co-ax (depending on your weight of your 
payloads) might be able to much better withstand twisting and weight 
loads than Cat 5 UTP.  It would help resist the twisting more also.

Nate