###AA6YQ responses below
-----Original Message-----
From: [email protected] [mailto:[email protected]]on
Behalf Of Alan Barrow
Sent: Wednesday, November 25, 2009 8:34 AM
To: [email protected]
Subject: Re: [digitalradio] Digital busy detect
Dave AA6YQ wrote:
> +++The rules to be honored by all stations are:
> 1. if you're not yet in QSO, don't transmit on a frequency that is
> already in use (meaning that signals have been detected during the
> past 5 minutes)
> 2. if you're in QSO and signal other than that of your QSO partner
> appears (the busy frequency detector indicates the presence of signal,
> but you aren't decoding your QSO partner), wait for that signal to
> disappear, send "QRL QRL" in CW, and resume your QSO
OK so far
### Progress!
> +++Amateur radio operators have been trusting each other to mutually
> obey these rules since the service began. On what possible basis can
> you claim exemption?
Here's where it falls apart..... many, many digi ops neither copy CW even to
understand QRL, or would not hear it.
### They need not copy it: they need only understand that CW in response to
a CQ in any mode means "frequency in use, please move elsewhere".
### There will be cases where asymmetry in equipment or propagation results
in a station sending CQ not being able to hear either of the stations in an
on-going QSO that are sending "QRL" in response, but this a fortunately
infrequent occurrence that cannot be addressed by any technology. The fact
that we can't prevent this is no excuse for not addresses the more common
scenario that we can mitigate; as Voltaire said, "the perfect is the enemy
of the good".
And another large percentage would not honor a QRL request, they don't in
other situations for sure.
### I don't agree that this is a large percentage now, and believe that the
amount of negative behavior would decrease as we eliminated the QRM.
> Kindof like asking all cellphone users to install a device that allows
> anyone to disable their ringtone. Just what do you think the compliance
> on that would be?
>
> +++No, its not remotely like asking cellphone users to install such a
> device; there is no parallel whatsoever.
I'd be OK if all mfg's had such a device. But to selectively enforce it is
unworkable. IE: Multipsk, others should have similar detect & honor a QRL
request. Recioprocity is part of being a good neighbor.
>>>MultiPSK only needs a busy frequency detector when its operating as an
unattended automatic station. Attended stations can use their ears.
+++ Only attended stations need detect the QRL; if automatic stations never
transmit on a frequency that is in use, then they will rarely QRM an ongoing
QSO, and so have no need of automatic QRL detection
This does not deal with hidden terminal,
### I disagree. If an attended station obeys rule 1, the probability that
the frequency was clear for 5 minutes before transmission and yet the
attended station's transmission will QRM an on-going QSO is very low. Within
that 5 minutes, the attended station's busy frequency detector would have
heard one of the two participants in that QSO. A collision would only occur
in the asymmetric equipment or propagation scenario.
nor does it address the cases where attended mode ops interfere with
non-attended
### I disagree. Rule 2 says that an unattended station in QSO that detects a
signal not sent by its QSO partner should send "QRL QRL" in CW. The operator
sending that signal would be governed by the "if you hear CW in response to
a CQ, the frequency is in use" principle. Barring an asymmetric scenario,
the unattended QSO would be preserved.
> +++When not in QSO, automatic stations can easily monitor the
> frequency to determine whether a QSO is in progress, even if they are
> only hearing one of the stations involved; this is easily implemented.
> If an automatic station receives a connection request and its busy
> frequency detector has seen no activity for the past 5 minutes, it can
> respond to the request without compunction. If its busy frequency
> detector has been intermittently reporting signals over the past 5
> minutes, it should not respond.
Unworkable on most auto sub-bands, there is just that much traffic. If you
held off 5 minutes for many parts of the day you'd never, ever be able
transmit.
### I have two reactions to this statement:
1. I'd like to see the statistics that back it up
2. if its true, you are acknowledging that unattended stations are QRMing a
lot of on-going QSOs
### If what you say is true, the proper solution would be to widen the auto
sub-bands; but this will only happen after its been demonstrated that
unattended automatic stations cause no more QSO breakage than good human
operators.
Again, unequal standard, do CW/RTTY ops wait 5 minutes? They sure don't when
interfering with PSK.
### Attended stations can listen for 30 seconds, send "QRL?", interpret the
result, and if negative proceed to call CQ. An unattended station with this
same capability -- specifically, "interpret the result" -- could similarly
reduce its wait time.
> +++I didn't say it rarely occurs, I said its rarely a problem --
> because attended stations can communicate with each other and resolve
> the conflict, thereby preserving the QSO in progress. Unattended
> automatic stations are incapable of doing this.
While voice ops do tend to honor this if "enforced", there are many, many
cases where they do not. Net's being one, and ops with unequal power being
another. The world is not as clean as you allude.
### In my experience, CW and RTTY ops honor QRL as well. The number of cases
where QRL heard but ignored is non-zero, but small; as you say, asymmetry is
also problematic. However, our goal here is to dramatically reduce the
incidence of QRM, not eliminate it. 100% solutions are rarely practical; if
we achieve 80%, we'll have reduced QRM incidence by a factor of 5.
Regarding your monitoring test with Pactor- clear "selection bias" (Google
it).... you could not hear the cases of interference that you claim do not
exist.
### I was not measuring the fraction of QSOs QRMed by a particular PMBO; I
was measuring the number QSOs QRMed by a particular PMBO within a particular
time interval. This is a valid measurement, devoid of selection bias.
>>>Rick KN6KB is already on the second iteration of his busy frequency
detector; there is no need to re-invent this wheel.
May be a wonderful thing, but does not address the need to be able to be
leveraged by other programs. You propose a technical solution is not that
hard.
### In particular, the Winmor busy frequency detector could be added to
existing WinLink PMBOs. This single step would yield a sign cant reduction
in QRM from automatic stations.
I asked for one, committed test & deploy it. That would be a big win.
Simple DLL, common standard. All major digi programs to implement & honor
it. Not just modes which you do not care for.
### The criteria is not "modes which you do not care for". The criteria is
"automatic stations running unattended -- independent of the mode being
used".
Since we all hope winmor will be a leveragable DLL, it would be great if
it's busy detector was also leveragable by other modes.
### An opaque DLL would be of limited applicability unless it turns out that
most automatic station control applications have similar decoding
architectures, which I doubt. An architectural description that developers
of automatic station control applications could incorporate into their
implementations would be effective. Source code might be helpful, but I
suspect that most of these applications have already defined the requisite
DSP classes.
> No changes are required to MMTTY, WinWarbler, DM-780, MixW, Digipan,
> FLdigi, or MultiPSK, as these are exclusively used in attended
> operation. Only applications that manage unattended automatic stations
> require modification -- specifically, the implementation of the two
> rules described above.
I respectfully disagree, as users of these programs interfere (perhaps
unintentionally) with each other and auto stations all the time.
### Attended stations already have the means to prevent the kind of QSO
breakage we've been discussing: their operators' ears (and eyes, if using
panoramic reception). Syndicating the CW QRL scheme discussed above should
minimize QRM between attended stations operating in different modes, as well
as between attended and unattended stations.
### Including busy frequency detectors in unattended automatic stations is
not a punitive step aimed at "modes we don't like", as you allege above. Its
goal is to give unattended automatic stations a "QRM avoidance" capability
that is comparable to that of good human operators - in whatever mode they
are using.
### I will of course comply with the moderator's time limit on this
discussion.
73,
Dave, AA6YQ
