| |
PA2AGA > HDDIG 07.10.99 22:53l 173 Lines 7009 Bytes #-9727 (0) @ EU
BID : HD_99_251C
Read: DL6KCF GUEST
Subj: HamDigitalDigest 99/251C
Path: DB0AAB<DB0KFB<DB0CZ<DB0LX<DB0RBS<DB0SWR<DB0FP<DB0SRS<DB0SIF<DB0HSK<
PI8DRS<PI8DAZ<PI8GCB<PI8HGL<PE1NMB<EA7URC<PE0MAR<PI8VNW
Sent: 991007/1610Z @:PI8VNW.#ZH2.NLD.EU #:5851 [HvHolland] FBB7.00g $:HD_99_251
From: PA2AGA@PI8VNW.#ZH2.NLD.EU
To : HDDIG@EU
Received: from pa2aga by pi1hvh with SMTP
id AA20663 ; Tue, 05 Oct 99 23:52:31 UTC
Received: from pa2aga by pa2aga (NET/Mac 2.3.67/7.5.3) with SMTP
id AA00016257 ; Tue, 05 Oct 99 14:05:01 MET
Date: Tue, 05 Oct 99 13:56:32 MET
Message-Id: <hd_99_251C>
From: pa2aga
To: hd_broadcast@pa2aga
Subject: HamDigitalDigest 99/251C
X-BBS-Msg-Type: B
"Link Availability" - How much of the time the link functions.
"If the link is down, how long can we wait for it to reappear?"
"Link BER" - The error rate in data flowing across the link in the
short term. "Can we afford 1 missed frame in each 10 frames sent?"
"Power Spread" - The difference between the ERP and noise floor at a
Station.
I'll use the example of creating a link between W0RLI (Station A) and N7QDN
(Station B). The path is roughly 50 km.over very rough terrain. The link
will run
at 9600 bps and 9600 baud. Desired Link Availability is 99% or better.
Desired
Link BER is better than one in 10,000. Others can provide that actual
numbers
for this type of link across this path if they are interested. Note that
this is one of
several links I am working with. This has been another confusion.
I will ignore modem effects such as coding efficiencies, data compression,
error
recovery, missing data fill, code spreading, etc. These are techniques that
can be
used to improve the link availability but simply confuse the issue at hand,
since
they can be applied to any link and just change the way you compute BER
and Availability.
When you design this link there are a number of steps involved.
1. Determine the path.
2. Determine the desired link availability.
3. Determine the Link Margin required to obtain the desired availability.
4. Determine the Power Spread required at each Station.
5. Choose radios, feedlines and antennas.
6. Decide if it is worth building the link.
What did we think about when we designed this link?
1. In this case the path is fixed: from my house to Duane's house. The use
of intermediate
relay points only confuses the issue: one would need to design THOSE links
instead
of the link we (Duane and I) wish to build. We already have a two-hop link
in place
between our two houses. What we wish to do is provide a single-hop link.
2. See previous paragraph. This will not be a "super link". It is ok if
there are
times when the link is down, or when the retransmission rate is high.
3. Someone else can do the calculation. It is complex. I'm lazy.
4. From the Link Margin, we obtain the needed Power Spread.
5. We both have several different radios, antennas and feedline available.
We might be willing to buy better antennas and/or feedline but probably
not better radios. We choose 10W radios, 6 element yagis and 9913 coax
because those are already installed and available for testing at both ends
of the path. We have both 2 meters and 70 cm available but prefer to
use 70 cm if possible, thus the choice of radios, antennas, and feedline.
6. We run some tests. First on 2M to verify signal levels because this
is simple to do: turn a knob on an existing radio at each Station. There
is plenty of signal. There is no Link Margin. This is not good. Since the
antenna at N7QDN is an omni, we are not surprised. There are many
large reflectors (mountains) in view in all directions. The scatter sounds
typical for this area: reflections from multiple mountains. Next we try
70 cm, using the Stations mentioned above. No signal detected in
either direction. Time to see what is wrong. My radio is badly mis-tuned.
The receiver local oscillator is almost 10 Khz. off frequency. I correct
that problem. No signal. I check some other paths to several nodes
using the same radio, antenna and feedline. Things work fine. The next
step will be to check Duane's radio and (probably) put it back on frequency.
What about this "Power Spread" thing? Let's say we discover that Station A
has a good deal more noise than Station B due to external noise sources near
it's site. The simple way to compensate for this is to increase the ERP of
Station B.
What does this mean? It means that Station A has a smaller Power Spread than
Station B, but the Link Margin can be made the same in both directions. It
might
be true, for example, that Station A can only have a small antenna due to
site limitations. Station B could have a larger antenna, and thus make up
the desired Link Margin in both directions. One way to think about "Power
Spread" is that it determines the cost of the Station. If you desire more
Power Spread, you need a more expensive antenna, radio and/or feedline.
Note that all the above comments are about a particular pair of sites.
A response of "Choose different sites" simply misses the point.
If different sites are chosen, one still has to do the same work
to determine if THOSE sites are worthwhile.
--
... Hank
http://horedson.home.att.net
>.
------------------------------
Date: Mon, 04 Oct 1999 14:14:11 -0700
From: Mark VandeWettering <markv@pixar.com>
Subject: Let's look at real numbers for TNC software sales
Hank Oredson wrote:
> Note that all the above comments are about a particular pair of sites.
> A response of "Choose different sites" simply misses the point.
> If different sites are chosen, one still has to do the same work
> to determine if THOSE sites are worthwhile.
After all, we know that the first solution which jumps into your head is
the best one!
Honestly, you've done this long exposition about how you desired to
build a single hop link between these two stations, with absolutely no
justification that it is better or even desireable to do so. The
shortest distance between two points isn't always the straight line that
you imagine.
Your "solution" to overcoming the large amounts of local noise at
station B by upping the ERP at station A is mighty un-neighborly.
Spraying all that extra power around limits the abilities of others to
use that spectrum space that you're polluting.
As hams, we are supposed to operate within the bounds of good amateur
practice. This governs not only what we should do, but also what we
should not do. We should not needlessly pollute spectrum with excessive
QRM just to satisfy our largely egotistical need to generate a single
hop link.
When a network WORKS like a network should, multi-hop links aren't the
bug-a-boo that people seem to think they are. Heck, a network without
any multi-hop links isn't a network at all, it is just a bunch of people
yelling at each other.
I've seen the argument made here that "any network is better than no
network". In fact, I know of at least one reason why that isn't true:
investing your time in networking strategies and equipment which provide
only marginal utility keeps you from devoting your time, money and
energies to strategies which will work. You can't convert the 1200
baud BBS packet radio infrastructure into a real network that services
To be continued in digest: hd_99_251D
Read previous mail | Read next mail
| |
