[aprssig] APRS low-power-local ALT input channel
bruninga at usna.edu
Mon Sep 27 14:04:14 CDT 2004
>>> HamLists at ametx.com 9/27/04 11:50:55 AM >>>
>Bob, you just can't help yourself, can you.
>Ok, time to put this flawed argument to rest.
>> Not really, every single user of ANY voice repeater does it
>> all the time. It is such a standard pratice that most people
>> appear to not even think about it.
>Not true. They are monitoring the frequency via the
>output of the repeater...
We disagree. As I enumerated, there are many reasons
why they may THINK they are avoiding each other, but
in actuality (on the basis of the typical 1 second of a
packet) they cannot use what they hear to reliably
avoid a collision.
>> Ah, one might claim that a repeater is full duplex real time
>> and so a user can tell if the channel is busy or not. Well
>> this simply is not so. Here are 4 examples of why not:
>> 1) If it was, we would not hear people *doubling*
>> 2) ALL modern repeaters have 1 second KERCHUNK
>> delays, that delay the repeater coming up by
>Not true. Maybe your repeater has a 1 second kerchunk
>delay, but most in this area do not.
Wow. That is amazing. Almost all voice repeaters back
East here require about a 1 second key down BEFORE the
repeater will come up. Or they require PL which also
takes about a second to lock on...
>The delays introduced by your alt-input channel digip,
>if using CD on 144.39, can be extensive AND if the
>stations using your alternate channel are trying to use
>CD on 144.39 for collision avoidance, then they will see
>little to no benefit...
I think that missunderstands the reason for CD on
144.39. It never has been possible to avoid a collision
with other users on the input by listening to the output
of a digi. The purpose of CD on 144.39 most of the time
is to simply avoid a wasted transmission while the digi is
already transmitting (since it cannot also hear at the same
time). It has nothing to do with avoiding other users (except
for the 4% of the time when you might hear a neighbor
>> 3) Many repeaters have delay lines that further delay
>> the output again, making the output not a true
>> copy of what is on the input.
>... This is why it is good practice to delay speaking after
>keying your microphone to allow the repeater to activate.
Glad to see we agree here. And during that one second
that you are waiting for the repeater to start to repeat you
is exactly the one second when a collision can occur
despite your most well intentions to avoid it.
>What is transmitted by the repeater is a true analog copy
>of what it receives on input.
Yes, but only after a second or so delay, and only if the
PL's match and only if everyone waits the full courtesy
period to transmit and only if everyone hits the PTT
switch slower than 1 second between them. Yes,
then we avoid collisions. But this shows exactly
my point. That it is NOT a poor "design" nor "bad
practice" as you claim to TX with a TX offset on a
system designed for such use!
All I am proposing is that APRS consider splitting the
user uplinks to the digis from the out of area QRM
to avoid the very collision driven poor reliability we
currently have on 144.39 by using a TX offset in the
>> 4) Many repeaters use PL on the input. Thus
>> guaranteeing that they are NOT hearing other users
>> on the input channel.
>This is the biggest fallacy that you promote.... repeaters
>on the same frequency use different CTCSS tones to
>ensure that the users of one repeater do not interfere
>with the users of another.
Which is EXACLTY what we are trying to do with split
operation on APRS!!! That is to give the LOCAL users
ACCESS to the input of the digi WITHOUT interference
from the DX QRM that is coming in on 144.39!
>As I have shown, split frequency with a FM repeater is
>far different than the albatross that you have proposed
So far, I have found very little that you have "shown" to
be valid to the realistic issue...
> You have made the same mistake many first year
>mathematics students make when told "prove this
>theorem". They set out to prove the theorem correct,
>creating invalid premises for the proof without realizing
>that there is also a second part... "or disprove [it]".
Yes, I am very much open to a statistical analysis by
anyone using sound logic. So far in this discussion,
we have modified my original assertions and so the
current debated value of the probability of a collision
on the input channel is sitting at around 4 to 7%.
You claim that is *big* and I claim that it is small
compared to the probability of a collision on 144.39.
I can assure you, that I would choose the 4 to 7% odds
on the alternate channel anyday over my odds on
>Bob, you won't change your way of thinking...
What? I just did. I modified my original claims of around
1 to 3% and upped it to 4% and can even accept 7%
based on your feed back and on other comments from
thoughtful respondents. I will always accept correction
and modification of the numbers based on realistic
reasoning. We all want to find the best solution for APRS.
>But I hope that people reading this thread do get a
>better comprehension of how far off base your arguments
>really are concerning this idea.
I hope they are reading it too. Because then hopefully they
can see the same numbers and same data and draw the
same conclusions. You notice that the 4% number is
based ***entirely on your numbers and your example***.
I'm not making this stuff up. Anyone can look at who they
hear direct, how often they transmit and what would be
their probabilitites of a collision on an alt-input channel.
If that probability is smaller than what they experience on
144.39 then the alt-channel will benefit them. If that
probability is larger than what they have to contend with
on 144.39 then they should stay there. We have shown
that in MY case and in YOUR case, that probability is
down in the 4% range.
I agree. Let the users decide whether to use the alt-channel
or not... but based on real numbers, not well debunked
platitudes like the power of "CSMA" in a mostly ALOHA
More information about the aprssig