Towertalk
[Top] [All Lists]

Re: [TowerTalk] Antenna Separation to eliminate Computer RFI

To: "Pat Barthelow" <aa6eg@hotmail.com>, towertalk@contesting.com
Subject: Re: [TowerTalk] Antenna Separation to eliminate Computer RFI
From: Jim Lux <jimlux@earthlink.net>
Date: Wed, 14 Sep 2005 19:35:27 -0700
List-post: <mailto:towertalk@contesting.com>
At 05:04 PM 9/14/2005, Pat Barthelow wrote:
>Folks,
>
>I am helping a local university  bring up an HF station equipped with
>Winlink Modems and softare to use as a medium for teaching computer
>networking and security.   The station antenna  is set up in the main
>student and classroom computing center, on the roof, within 100 ft of maybe
>150-200 networked computers.  Surprise Surprise, S-9 white noise hash on
>most bands.
>My question is, We need to have the Transmitter and winlink setup located at
>a classroom teaching position, along side the HF station, but it may be
>possible to move the antenna some distance away from the building.  Is it
>likely that a movement of the antenna from directly over the computing
>rooms, to, say 150 ft horizontally away will reduce the current noise
>sufficiently to be usable?  How rapidly does the signal intensity decrease,
>as I move the antenna away from the building?

A lot depends on how the noise is getting to your victim rig.  Have you 
tried running on battery power for the rig and PC (make sure the AC adaptor 
for the laptop is disconnected from the wall, mine is a fine noise source).

You might be getting noise carried on the outside of the coax, too.  Some 
significant chokes at both ends might help.

The point is, if the noise is conducted into the victim, then moving the 
antenna won't help.

If you're sure the noise is "radiated emissions", then you're into finding 
out what the RF properties of the building are, which can be somewhat 
directional.  You might be able to find a "quiet spot" or quiet 
polarization. A battery powered SW receiver with a whip antenna is handy.

In particular, you might want to consider a two element array to create a 
null in the direction of the building. Two antennas fed with equal length 
feedlines, and wired so they're 180 degrees out of phase at the antenna (or 
at the T connector) oriented with the line between the antennas at right 
angles to the line to the building might work.


----

Just as an aside, why you wouldn't use a nice open source HF networking 
protocol like MIL-STD-188 which has a much more modern protocol stack than 
Winlink/Pactor/AX.25...  There's a variety of PC sound card implementations 
for the -188 modem (technically -188-110A or FS1045 or some STANAG spec 
number), and there's a nice free program called PCALE 
(http://www.hflink.com/) that supports most all of -188, including the 
automatic link establishment (ALE), the various voice and data modes, 
email, file transfers, etc.

More to the point from a pedagogical standpoint, MIL-STD-188 is very well 
documented, there's a lot of theoretical analysis available for ambitious 
students to go track down, and all the details of both good and bad Golay 
encoder/decoder and the 8-tone FSK modulator/demodulators are 
available.  There's also Matlab code for much (if not all) of the DSP. 
While there are some Pactor I modem implementations out there for PCs, 
anything more sophisticated, in particular, Pactor III, upon which Winlink 
depends in a practical sense, has most of the details hidden by SCS, who 
own the patents and have been very aggressive in going after folks who have 
reverse engineered it.

When it comes to the higher levels of the protocol stack, Winlink is a sort 
of modified HF packet AX.25 implementation with scripting and so forth to 
implement the sounding and data transfer functions. It also relies on a lot 
of operator skill in setting up all the schedules (essentially, it uses 
human skill to put in the knowledge of propagation 
predictions).  Mil-Std-188-141 is much better thought out.  It does the 
real time sounding and builds up matrices of which frequencies to use to 
call which stations at which times.  Overall, a much cleaner and more 
elegant and functional implementation. Probably because it's the work of 
people who do it for a living, and they are very aware that people live or 
die depending on if it works, so little icky details like fault recovery 
are very well specified and handled.

Winlink/Airmail/etc does have a tighter connection to the internet, if 
that's what you're interested in. And, it used to have a lower entry cost. 
These days, though, with PCs to do the coding and modem work, it's cheaper 
to get on the air with -188.

If you're interested in developing a capability which will interoperate 
with officialdom (i.e. disaster services, DoD, UN, etc.) then Mil-Std-188 
is the way to go.  The Winlink/SCS world is pretty much restricted in a 
practical sense to the sailing/RVing/ham community.  Until a few years ago, 
Winlink/SCS was the only low budget entry point (although SCS modems do 
cost a kilobuck), so there wasn't much interest in -188 compatibility.

We've been looking at implementing some form of digital messaging over HF 
for our SHARES station at JPL, and the kilobuck for a proprietary modem 
manufactured by only one foreign company is a significant sticky point. 
(The "Buy American Act", desires for second sources, and all that)  SCS 
might be fine for sailors to get email and digital weather (if they're not 
in the INMARSAT or satphone bracket), but not something you want to rely on 
for critical emergency infrastructure. (It's widely agreed that PACTOR I 
isn't all that wonderful for HF links.)

However, you're just putting up what's essentially a classroom 
demonstration system with non-critical traffic, and Winlink/Airmail/Pactor 
I might be just fine for that.


Jim, W6RMK  

_______________________________________________

See: http://www.mscomputer.com  for "Self Supporting Towers", "Wireless Weather 
Stations", and lot's more.  Call Toll Free, 1-800-333-9041 with any questions 
and ask for Sherman, W2FLA.

_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk

<Prev in Thread] Current Thread [Next in Thread>