RadioBanter - View Single Post

Jeff Liebermann[_2_]

On Mon, 8 Dec 2008 14:18:49 -0800, (Dave Platt)
wrote:

In article ,
Jeff Liebermann wrote:

Good plan but there are problems. Most hams these daze use walkie
talkies with perhaps 1 watt of TX RF. The typical repeater is running
perhaps 10 to 40 watts out (after the duplexer). The walkie can hear
the repeater almost anywhere, but when trying to talk, they drop in
and out all the time. The mobiles, which run more power, are usually
well matched to the repeater's tx power. I've suggested adaptive tx
power control (to preserve battery power) on our 2 meter repeater, but
nobody wants it.

Seems to me that would introduce another set of incompatibilities.
People out on the fringe area of the repeater's coverage would be able
to hear repeated signals from weak transmitters (e.g. HTs), but if the
repeater saw a strong input signal from a mobile (or an HT near the
site) and dropped its transmitter power, the repeater coverage area
would shrink abruptly and those users out around the edges could lose
coverage. This could re-create the "hidden node" problem in a new way!

Good point, but that's not what I had in mind. I was aware of the
problem and suggested that the power control be run by a micro. The
algorithm needs to be worked out, but basically, the weakest signal
sets the power level. The range of power adjustment also will not be
huge or large enough to have much of an effect on distant stations.
Our current 60 watt amplifier might have the power output reduced to
about 20 watts or 4.5dB power reduction. That's less than one S-unit
and probably wouldn't be noticeable even by distant stations. However,
it will make a huge difference in power consumption and battery
operation run time. There are a few other economies that can be
thrown in, such as having the repeater always ID in low power. Of
course, there are complications, such as the mobiles and handhelds not
reporting their RX signal quality, which makes deciding the repeater
power output somewhat problematic when dealing with the traditional
broken radios. (Same people, same radios, same problems, same club
net, every time). Incidentally, this was suppose to be a minor
feature of a proposed repeater voting system, which is another can of
worms.

I don't think adaptive power management can work reliably in the
absense of a signal-quality feedback from each station which is
accessing the repeater/AP.

Yes it can. I intentionally reduced the repeater TX power during a
club net from 60 watts (before the duplexer) to about 20 watts. Nobody
said anything, there were no complaints, and nobody noticed, even
after I mentioned it after the close of the net. As long as I don't
run the power output to near zero, it can me made to work, even if the
power level is set manually with a touch tone command.

We had a co-channel user that was running carrier squelch. It took
only about 7 years of constantly pounding on the trustees before they
would install PL. They lied on their NARCC application claiming they
had a functional PL system.

Grrr. Worf "Romulans have *no* sense of honor." /Worf

Actually, it was far more stupid than it appears. We had several of
the club politicians pounding on the alleged trustee of the repeater,
who literally didn't care, didn't communicate with the rest of the
organization, broke numerous promises to take it up with the board,
and so on. Complaints to NARCC went to the trustee with predictable
results. We were simply talking to the wrong person. One day, we
caught the right person on the air working on the repeater. He was
informed of the situation, after a few emails and the usual
miscommunications, the PL decoder magically appeared. As would be
expected, the initial transistion was slow. First there was no-PL
during their club nets. Later, it became full time. There was also
some work done on placing a notch in the antenna pattern in the
direction of Santa Cruz. That has worked quite nicely.

Our system (W6ASH, SPECS, at El Camino Hospital in Mountain View)
switched over to PL-tone-only years ago, and it helps in numerous
ways... the absence of noise-initiated kerchunking is a real blessing.
I've preserved the ability to switch the repeaters over to
carrier-sense-only in an emergency but really don't expect to ever
need to use this feature.

That alleged benefit is that no PL allows visitors and hams with
antique hardware to use the repeater. We have a control code that is
available to anyone that will temporarily disable the PL for a
specified time period. It actually does get used by some hams who
apparently plan on being buried with their ancient radios and have no
intention of ever selling them.

The experience taught me a few things,
one of which is that hams generally make lousy RF neighbors.

Some are lousy neighbors, lazy and self-centered. Others hold
themselves and their equipment and operation to very high standards. I
think it depends very much on the individuals involved, on their
attitudes, and on their level of prior experience in supporting a
real-world user community.

It varies. I consider myself a good example of one of the lousy RF
neighbors, so I know they exist.

My impression is that a limited set of mesh repeaters, and a plentiful
set of direct backhaul links on a different (non-interfering) band,
works out rather better.

Sure, but that's not the way most municipal wireless networks are
umm.... engineered. Most are single channel, store and forward
repeaters. The problem is that if you're perhaps 3 nodes away from
where it hits the wired internet, you will have 4 identical duplicated
packets flying through the air. If everyone can see everyone else (as
is the case with omni antennas on office rooftops), only one of those
radios can be transmitting at a time. That means that you would be
hogging at least 3 times the shared bandwidth (airtime is shared) to
deliver one lousy packet. Add the problem of trying to create the
ultimate geographic routing algorithm, and mesh networks have a
propensity to fill the airtime with duplicates, retransmissions, and
junk. The worst of the worst are broadcasts, noteably ARP requests,
which go everywhere. I was monitoring what was coming out of Google's
Mtn View network and finding about 60% of the airtime occupied by ARP
requests. Of course, when the proponents of municipal wireless find
that there are holes in their coverage, they just add yet another
store and forward mesh repeater. Eventually, the density is high
enough for the mesh repeaters to start interfering with each other.
Since this is difficult to diagnose, the usual solution is to increase
the TX power of the mesh node. That does the opposite of what might
be expected, but until the numbers are in for availability testing,
everyone just assumes that more power is more better. It took quite a
bit of arguing to convince one mesh operator to reduce their power and
even more to demonstrate that it worked. As an example, see the data
and conclusions from the old MIT RoofNet at:
http://pdos.csail.mit.edu/roofnet/doku.php?id=interesting
Note the probabilies of actually delivering a packet through just one
hop. Yech.

I couldn't resist and just had to rant. But, it was worth it.

Onward to another thrilling day of doing battle with QuickBooks and my
traditional end of the year billing nightmare...

--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831-336-2558
# http://802.11junk.com
# http://www.LearnByDestroying.com AE6KS