| |
PA2AGA > HDDIG 08.08.00 12:18l 166 Lines 7446 Bytes #999 (0) @ EU
BID : HD_2000_213A
Read: GUEST
Subj: HamDigitalDigest 2000/213A
Path: DB0AAB<DB0SL<DB0RGB<DB0MRW<DB0ERF<DB0FBB<DB0GOS<DB0PKE<DB0OVN<PI8JOP<
PI8ZAA<PI8HGL
Sent: 000808/0103Z @:PI8HGL.#ZH1.NLD.EU #:3334 [Den Haag] FBB $:HD_2000_213A
From: PA2AGA@PI8HGL.#ZH1.NLD.EU
To : HDDIG@EU
Date: Mon, 07 Aug 00 23:46:49 MET
Message-Id: <hd_2000_213A>
From: pa2aga@pe1mvx.ampr.org
To: hd_broadcast@pa2aga.ampr.org
X-BBS-Msg-Type: B
Ham-Digital Digest Mon, 7 Aug 2000 Volume 2000 : Issue 213
Today's Topics:
56Kb Packet Radio Modem (4 msgs)
Looking for Nico Palermo -YAM author
Next Saturday PSK31 relay experiment
Palm Pilot, Email, Packet? (2 msgs)
Vote Green Party (2 msgs)
Weekly Speaking 15.0- Poetry Slam Nationals
Send Replies or notes for publication to: <Ham-Digital@UCSD.Edu>
Send subscription requests to: <Ham-Digital-REQUEST@UCSD.Edu>
Problems you can't solve otherwise to brian@ucsd.edu.
Archives of past issues of the Ham-Digital Digest are available
(by FTP only) from ftp.UCSD.Edu in directory "mailarchives/ham-digital".
We trust that readers are intelligent enough to realize that all text
herein consists of personal comments and does not represent the official
policies or positions of any party. Your mileage may vary. So there.
Loop-Detect: Ham-Digital:2000/213
----------------------------------------------------------------------
Date: Sun, 06 Aug 2000 13:54:32 -0400
From: Gary Coffman <ke4zv@bellsouth.net>
Subject: 56Kb Packet Radio Modem
On Tue, 25 Jul 2000 15:45:08 +0300, Paul Keinanen <keinanen@sci.fi> wrote:
>On Mon, 24 Jul 2000 08:36:54 +1000, "David Findlay" <nedz@bigpond.com>
>wrote:
>
>>Thanks for this info, I am probably going to build a WA4DSY 56kb radio
>>modem, unless I can find a better design. Anyone know of what sort of range
>>you can get? I need around about 50km. All stations will be withing a square
>>surounding a digipeater, at a maximum of 25km away. They need to be able
>>each others transmissions as it will be using an ethernet topology.
>
>Ethernet style CSMA will not work reliably with such distances and
>high speeds. In order to work reliably, omnidirectional antennas would
>have to be used and in order to get the required ERP, quite high
>transmitter powers would be required. My guess is that the 56 kbit/s
>receiver will require about -110 dBm of receiver power. Looking at
>some old troposcatter diagrams (The Radio Amateur's VHF manual,
>1972), the path loss for 50 km and 99 % reliability is about 160 - 170
>dB. Thus, a transmitter power of +50 .. +60 dBm (100 W .. 1 kW) is
>required.
The WA4DSY RF modem requires a signal strength of about -110 dBm
to maintain a 1 in 1E6 bit error rate. But that's the signal strength at the
modem's 29 MHz input. The transverter ahead of it may have gain, as
may the antenna, so the off air field strength may be considerably lower.
(Typically with our setups about -140 dBm is the point where bit errors
exceed 1 in 1E6.)
We run a pair of 56 kb stations over a 70 mile path. The stations run
about 4 watts transmit power, as do all of our sites. It is a mountain top
to mountain top LOS path, though, and we use a long boomer gain antenna
at one end. The path is very reliable. But we have another path that is
only 18 miles, but obstructed, that has periods of high errors, correlating
with the formation of an inversion layer over the area on summer evenings.
One site is above the layer, the other is below it.
Gary
Gary Coffman KE4ZV | You make it |mail to ke4zv@bellsouth.net
534 Shannon Way | We break it |
Lawrenceville, GA | Guaranteed |
------------------------------
Date: Sun, 06 Aug 2000 14:15:45 -0400
From: Gary Coffman <ke4zv@bellsouth.net>
Subject: 56Kb Packet Radio Modem
On Wed, 26 Jul 2000 10:25:28 +0300, Paul Keinanen <keinanen@sci.fi> wrote:
>On Wed, 26 Jul 2000 08:27:46 +1000, "David Findlay" <nedz@bigpond.com>
>wrote:
>
>>Could a single frequency be used?
>
>Yes, that would be traditional digipeating. The problem is that you
>have to receive the full frame before it is retransmitted. Also the
>reply/acknowledge signal will suffer from this delay. If multiple
>digipeaters are used to relay the message and end to end acknowledge
>is used, the latency times becomes so huge, that the system is more or
>less useless.
>
>Remember also that in AX.25, the maximum data size is about 250 bytes
>and at 56 kbit/s, the frame transmission times are below 50 ms, so a
>very fast Rx/Tx turn around time is required.
We run much longer frames than that (that limit is forced by some TNCs,
but we don't use TNCs, and set frame lengths of between 1k and 4k bits),
and run a large maxframe too, but you're right that latency becomes a very
serious problem as multiple hops are traversed in a simplex digipeating
environment.
>>What about full duplex operations?
>
>Full duplex is in principle only point to point, but if multiple
>stations are required, the simplest solution is to use one master
>stations with opposite frequencies, while the slaves communicate one
>at the time with the master in full duplex.
>
>Note that full duplex operation on a single amateur band will require
>duplexers on all stations or the receive and transmit antennas must be
>kept very far away from each other. Using different amateur bands for
>the uplink and downlink simplifies the system, but some frequency
>planning is required to avoid harmonics.
We do that, using 222 MHz and 440 MHz. You need two transverters,
and usually two antennas too, so it isn't a really popular approach for
users. It is worthwhile for node to node links though.
>The use of a half duplex (phone) repeater is simpler, since only this
>station has to operate in full duplex and require a duplexer if both
>frequencies are on the same amateur band. The slave stations needs to
>be able to receive on one frequency and transmit on an other, but not
>at the same time, so no duplexer is required.
We do that too, on 440 MHz. It is more popular with our users since the
equipment required at their stations is simpler, ie only one transverter
and antenna.
>>I still don't know how to work mobile unless just having lots of
digipeaters.
>
>With normal AX.25 you are going to have a lot of "hidden transmitter"
>problem, so the CS (Carrier Sense) part of CSMA will not work
>reliably. You need some other mechanism to control when the mobiles
>are allowed to transmit.
Or just run pure Aloha and tolerate the collisions. The latter works ok,
expecially if you implement a random exponential back-off algorithm,
if the frequency is not very congested. You will eventually get congestive
collapse (or just as bad, back-off to practical infinity) if too many stations
are on, though.
>Note also that most amateur 1200, 9600 and 56 kbit/s are based of
>narrow band modulation. These suffer from multipath (selective fading)
>cancelling certain frequencies. When the vehicle is moving, the
>frequencies that are notched out varies constantly and the narrow band
>data stream fades to noise quite frequently or at least cause severe
>phase shifts, corrupting the data. For reliable communication with
>mobiles some modulation method with a large bandwidth (such as spread
>spectrum, TDMA or COFDM) and strong FEC may be required, so that the
>multipath notch will take out only part of the signal and the missing
>part can be reconstructed from the ECC bits.
>
>If you can guarantee that the mobiles are constantly moving, even
>narrow band modes may be useful at least above 1 GHz, since the
To be continued in digest: hd_2000_213B
Read previous mail | Read next mail
| |
