[CLUE-Tech] wireless problem

[Nate Duehr]

On Feb 11, 2004, at 11:08 AM, Eric Jorgensen wrote:
> I have a Dell Latitude C810 running RH9. I have added
> a Truemobile 1150 pcmcia 802.11b card running the
> orinoco driver, connecting to a Netgear WG602 access
> point. I have had this combination for about a year,
> and thought that even though I do everything over ssh,
> I should
> tie things down a bit. I changed the default essid and
> gave the Netgear a static IP address, and I upgraded
> the firmware to 1.5.67. Everything worked fine for
> several days after these changes, but since
> a week ago, I've been seeing very poor performance
> (lag while typing, etc.) as well as the following in
> /var/adm/messages:
>
> Feb 11 10:56:20 martin kernel: eth1: New link status:
> Disconnected (0002)
>


> Running kismet shows another couple of stations in my
> neighborhood, one even running on channel 11.  I moved
> to channel 7, but the problem remains.
>
> Any help would be appreciated, as well as pointers to
> other resources (mailing lists, etc.) that could also
> be helpful.

Unfortunately 2.4 GHz low-power devices abound.  Baby monitors, 
Cordless Phones, even your microwave oven's center frequency is around 
2.4 GHz (just happens to be the right frequency for warming water 
molecules -- thus how the food gets cooked).  The bands used to be 
licensed, but the FCC's latest "grand experiment" has been to open up 
bands and let them become a free-for-all.  It has brought us great 
gains in devices like 802.11b and 802.11g wireless cards but it also 
comes with it the joy of interference.

First a theory:

 From your description, I think you're struggling with a new broadbanded 
2.4 GHz transmitter in your neighborhood.  The only sure-fire way to 
find it would be to do a site survey with a radio service monitor and a 
spectrum analyzer.  However you might be able to pick up one of these 
cheap $20 "find 802.11b signals" type gadgets... they're just a cheesy 
2.4 GHz receiver and a light, in many cases.  In your case, the cheaper 
the better -- you don't want one with enough intelligence to tell the 
difference between 802.11b signals and other transmitters.

It could also be an 802.11g system if you live in close proximity to 
your neighbors (apartments, multi-family "dwellings"...I love that 
phrase, always makes me think of the cliff dwellings in the southwest.) 
  802.11g just spreads the signal out across the entire set of channels 
802.11b uses... to get more bandwidth through. Someone could have fired 
up an 802.11g system close-by if you live in an apartment or tight 
quarters.  In that case, distance away from the system is the best way 
to alleviate the problem.

The joys of RF ... bit-rate is a linear math function to bandwidth.  To 
get a higher bit-rate you have to up your bandwidth utilized.  However 
RF signal decreases 2 X distance^2 as you move away from a 
transmitter... an exponential function... so moving away from an 
offending transmitter or noise source is usually more effective than 
anything, especially at the low power levels allowed in 2.4 GHz 
devices.

Of course... geeks constantly publish "hacks" to add external antennas 
to 802.11b and other Part 15 devices.  Most of these are thankfully 
very very bad designs (the infamous "Pringles Can" antennas come to 
mind... complete crap from an RF design standpoint) and the geeks don't 
feed them with hardline (coax is extremely lossy at 2.4 GHz... most of 
your signal leave the coax and radiates or turns into heat before it 
reaches the antenna) so they aren't very effective -- but some people 
do figure out the "right" tools to use and exceed the Effective 
Radiated Power (ERP) limits imposed by law on Part 15 devices by tens 
to hundreds of watts, depending on the gain of their antenna system.  
Luckily to get those levels of gain the signal has to be very focused 
-- high gain antennas are very directional by design -- so the 
interference problem is localized.

The best advice for help:

Experiment -- you should just try different channels until you find the 
best performer.  Or a switch to 802.11a and move up a few GHz in 
frequency would work too.  At least until those become popular.  ;-)  
Once they all become overly-popular in a densely populated area -- 
we're all basically hosed.  (GRIN)

And the real reason for the problem:

In general the FCC's stance right now is "give away bandwidth to 
whoever wants it... and may the original licensees be damned."  This 
shows clearly in the decisions made over the last five to ten years by 
the Commission -- EXCEPT in the traditional broadcast bands.  Those are 
locked up forever by the money and powerbase of the owners of the radio 
and TV companies.

Broadband-over-powerlines (or "BPL" as it's called) is turning into the 
next huge RF battle.  BIG money behind it.  Huge interests who donate 
to Congressional representatives and Senators.

Most engineers agree, HF-band communications the world over will be 
virtually wiped out by BPL technology -- it's been seen in the test 
areas in the real world and also in the labs -- BPL noise will be 
broadbanded and hideous.

Ham Radio, Shortwave Broadcasts, and Emergency long-range 
communications systems like FEMA and others have -- all will be 
destroyed so people can have web pages and spam.  And yet, your elected 
officials have appointed people to the FCC Commission Board (and this 
is a direct quote from a Commissioner) who claim BPL is "Broadband 
Nirvana."  Hmmm... another free-for-all coming, I think.

Wonder who's paying them off... follow the money!  That's what good 
investigators do...

Sorry -- I wandered off-topic here, but anyway... that's where it's 
headed... RF spectrum used to be completely unregulated, then 
government stepped in, now government is backing out and going to let 
"the markets" decide who uses what spectrum... at least over the 
long-term that seems to be where it's all headed.  So like our legal 
system -- your RF devices will work if you can afford the latest and 
greatest gadgets with built in filtering for all the RF pollution 
someday.  That level of RF pollution is coming -- probably sometime in 
the next 30 years.

Back to your issue:

You'll probably have a hard time finding your noise source.  And 
devices licensed under Part 15 of the FCC regulations and other 
regulations like those that cover the ISM bands (Industrial, 
Scientific, and Medical) like 2.4 GHz effectively wash clean the hands 
of the only governing body capable of policing the RF spectrum.  It's 
bad policy.  (900 MHz was the first to go... the licensed users of 900 
MHz including Amateur Radio Operators and the Federal Government find 
the band all but useless for many applications today because of baby 
monitors, cordless phones, and other RF crud... oh yeah, I think 
Ricochet is there too, and one other local wireless ISP -- Suburban 
Broadband's gear.)

Good luck with it... yes, 802.11b/g are a great application of 
technology that has spurred on lots of business, but it comes at a 
price... spectrum management is impossible on the 2.4 GHz bands now... 
thus, certain limited incidents like yours where your device simply 
won't work will happen more and more often.

On the bright side, there's a big future in RF spectrum analysis tools. 
  A friend works for a company that specializes in DSP-based portable 
spectrum analysis gear.  So far the largest purchasers are government 
agencies battling interference from commercial entities (cell phone 
companies, some 2-way systems) who don't maintain their licensed 
systems well -- but many investigations have led to malfunctioning 
consumer gear.

He related one story of asking a department store to turn off their 
wireless music distribution system because the transmitter was 
malfunctioning and jamming a local Police dispatch frequency.  And they 
DIDN'T WANT TO.  They also attempted to turn it back on as soon as the 
investigators went to leave, so they went in and just confiscated the 
thing.  Great priorities, there guys... the customer MUST have MUSAK!!! 
  Who cares if the cops can't talk on their radios?  ;-)

Nate Duehr, nate at natetech.com