| |
ZL3AI > APRDIG 04.10.03 11:17l 203 Lines 9471 Bytes #999 (0) @ WW
BID : 2080-ZL3AI
Read: GUEST
Subj: TAPR Digest, Oct 01, 5/11
Path: DB0FHN<DB0RGB<DB0MRW<OK0PKL<OK0PHL<OK0PBB<OK0PHK<OK0NAG<OK0PPL<DB0RES<
ON0AR<N2BQF<W4JAX<VK4TUB<ZL2BAU<ZL3VML
Sent: 031003/1354Z @:ZL3VML.#80.NZL.OC #:9640 [Chch-NZ] FBB7.00i $:2080-ZL3AI
From: ZL3AI@ZL3VML.#80.NZL.OC
To : APRDIG@WW
Subject: Re: DigiNed, was: Re: Take Control of your DIGI part 4
From: Henk de Groot <henk.de.groot@hetnet.nl>
Date: Wed, 01 Oct 2003 16:51:17 +0200
X-Message-Number: 29
Hello Sean,
At 21:16 30-9-03 -0500, Sean \\The RIMBoy\\" <sean@rimboy.com> wrote:
>I doubt that digined has much in the way of memory requirements. Looking
>over Henk's pages last night, it will run on a 286 w/ DOS. It can't be
>all that bad. For comparison, I crammed aprsd onto a floppy running
>linux, the minimum memory requirement being 12 megs of ram.
I even tested it on a 286 Commodore laptop with 1 MB. 1 MB is needed when a
ramdisk is used. The digi_ned bootfloppy in with which you can create
standard digipeater functionality within minutes uses this. The only thing
that is a bit slow is sattellite tracking. Each calculation takes about 2
seconds (packets received in that time are queued in the AX25_MAC TSR).
>The question becomes, what interface to use. Henk, can digined use the
>soundcard TNC under linux or do we need an external device. If we look at
Under linux the kernel AX.25 interface is used, so yes, of course.
>it from the floppy angle then we're talking one, maybe two soundcards
>being supported at most (kernel bloat with each soundcard). What about an
>external TNC, maybe a KISS type? What about something like this?
>
>http://www.john.hansen.net/PICTNC.htm
Should be no problem for either the DOS or the Linux version. I think it
will run fine also with the above mentioned bootfloppy, its just KISS!
>I'm open to suggestions and willing to do some development on this. The
>only bad part about floppies is that they're not reliable, especially at
>some of the repeater sites I've been to (dust, heat). However, they're
>dirt cheap. CF stuff is getting pretty cheap and the nice part is an
>upgrade is only a CF switch away.
Both our two digis boot of a bootfloppy and all active stuff is loaded to a
ramdisk. Only at boottime the floppy drive is active, we did not have any
problem to reboot the station after one year of uninterrupted service, it
survived the summer and the winter with no problems. I know it is nothing
compared to the conditions on a mountain top, but a good enclosure can save
you from a lot of trouble. By the way, both run on HP 486 Omnibooks; on
PI1APK the LCD screen did not survive the year (summer or winter, don't
know when it broke) but it still runs fine... If you find some
"solid-state" way to boot then that will be okay too.
Kind regards,
Henk.
(the bootfloppy is available on my own web page:
http://www.homepages.hetnet.nl/~pe1dnn)
----------------------------------------------------------------------
Subject: Re: [ Bob Bruninga ] Re: Digipeater Duplicate Checking is broke
From: Henk de Groot <henk.de.groot@hetnet.nl>
Date: Wed, 01 Oct 2003 17:47:51 +0200
X-Message-Number: 30
Hello Gregg,
At 22:42 30-9-03 -0500, Gregg G. Wonderly wrote:
>Last time we talked about fratricide, I volunteered a diagram. The important
>detail is that if both Digis are DWAIT=0, then the area that has zero fade is
>the place that is the RF middle between the digis. If they are equal power
>and gain and haat, then it is smack in the middle. If I am closer to B than
And then it only works if the two digis are exacly on frequency, if they
are say 10 Hz off then you have a big 10 Hz fading component in the middle
and if the carriers are 180 degrees out of phase it will be completely
silent...
This overlapping transmission is done in CDMA and WCDMA systems as used for
CDMAone, UMTS and CDMA2000. It works but only with extreem good power
control regulation and very, very accurate center carriers. This is so
accurate that in WCDMA the signals for individual stations are compensated
for doppler shift because as you move, you move toward one of the
basestations and away from the other; the first raises the observed
frequency while the second lowers it. When the mobile needs to harvest
energy from both basestations (in so called "softer handover") these
frequency shifts have to be compensated. I know this is way beyond amateur
use, I just use this to illustrate that overlapping signals is more than
just putting two unsynchronised signals on air at more or less the same time.
Overlapping is also used by FM broadcasters who only have a license for one
freqency. It can work but only if the FM transmitters are synchronized to
eachother and the audio signal is transmitted in the same phase. Neither is
the case with our simple digis. If we want that then we need digis that can
synchronise to GPS signals; there are industrial GPS recievers that can act
as very accurate timebase (in fact that is what is used for WCDMA/CDMA2000
and perhaps also CDMAone). Besides that they also need fixed points in time
to start the transmission of the same packet at exactly the same time on
air (i.e. we need to take the audio delays in the TNC and TRX into account
to make it to the antenna at exactly the same time).
For the rest I agree your comments, these are exactly the issues for
carefull planning of UMTS statons (= WCDMA technology). GSM and (D)AMPS are
so much simpler in this respect because different stations use different
frequencies (there frequency planning is the big problem). As you mentioned
paging systems have the same problems with synchronisation if simultaneous
transmissions are allowed, I don't know enough about POCSAG/ERMES/FLEX
deployment.
Anyway, I don't think it will work at all for our unsynchronised APRS
hardware and only creates a undetectable mess in the overlapping area. Our
1200 baud discriminators are also not that good to filter out both AFSK
carrier signals from a bunch of interferrence noice, too cheap.
Kind regards,
Henk.
----------------------------------------------------------------------
Subject: Re: DigiNed in dedicated flash devices ??
From: Henk de Groot <henk.de.groot@hetnet.nl>
Date: Wed, 01 Oct 2003 18:02:51 +0200
X-Message-Number: 31
At 01:49 1-10-03 -0400, WA8LMF2@aol.com wrote:
>I suspect that since Digi-ned listens to and talks to a TNC connected to a
>serial port and doesn't do anything fancy with the video console, it might
>actually be usable as-is in a Pocket PC under a DOS emulator.
I currently use borland's "cprintf" that write direcly to Mono/CGA/EGA/VGA
memory, bypassing the BIOS. Thats easy to change (under Linux "printf" is
used).
For serial I/O the software talks directly to the UART since serial support
in the BIOS was never any good. Even APRSdos does not use the BIOS for
serial I/O. I guess the DOS emulator has to emulate this hardware.
DIGI_NED os pritty clean and should not give many problems to get it to
run. The MAC layer driver that drives the different modems and TNCs is
another matter. The MAC layer driver use all kinds of tricks, for example
to use a BayCom modem, one of the PC timers is reprogrammed to 3600 Hz
instead of 18 Hz. So it steals some interrupts and changes this 3600 back
to 18 Hz to satisfy the BIOS and the DOS clock. Also the RTC chip is accessed.
So I think it is tricky to get this thing to run, although it may work if
the emulation is good enough. I think its easier to rewrite a simple new
one that only knows how to drive a KISS TNC without support for all the
other type of modems. That should be rather easy to do. Maybe a TSR is not
needed at all and driving the KISS TNC could be directly from DIGI_NED.
Sounds quite feasible to me.
Kind regards,
Henk.
----------------------------------------------------------------------
Subject: San Diego LANn-N
From: Bob Bruninga <bruninga@usna.edu>
Date: Wed, 1 Oct 2003 13:08:34 -0400 (EDT)
X-Message-Number: 32
Another great area that has an easily defined LANn-N geographic area is
San Diego. Its ALOHA circle is about 50 miles, contains about 60 users in
the area and anything beyond about 60 miles is vast quantities of
QRM from hundreds of LA stations!
THey have only 5 digis that cover the area. IDEAL for a LANn-N named SD.
All they would need to do is set UITRACE SD,28 in those digis, and
BINGO they have a LAN that can fully communicate using SD3-3. ANd they
can practice with this and try it if they like with NO IMPACT to current
users. (except they loose TRACEn-N)
Now, when they are happy with how well that works, they can take the next
BIG step to cut out all their LA QRM. Switch the UIFLOOD in all
digipeaters between THEM and the LA FLOOD to something OTHER than
WIDE. Now they wont hear any WIDEn-N QRM from LA, yet WIDEn-N still works
across the mountains and to the south and east and the vast uncharted
desert of whatever exists out there.
And for someone who knows the area well, they could probably pick ONE digi
towards LA and another one well to the EAST of LA and another one to the
NORTH of LA and give them the UIFLOOD of SOCALn-N and then anyone wanting
to talk hundreds of miles north and south bypasing LA could use the
SOCALn-N path from Sandiego to Fresno and beyond...
And for someone in SD that still wants to send WIDEn-N into LA, they
simply send their packets via KA6DAC,WIDE3-3 or whatever (though I dont
think that is fair to the folks in LA)... but that is phase 3, user
education...etc.
So many opportunities, so much planning to consider...
Kibitzing from afar...
de WB4APR, Bob
----------------------------------------------------------------------
Read previous mail | Read next mail
| |
