llh@drutx.UUCP (HaymakerLL) (09/25/85)
Volume 2, Issue 3
September 17, 1985
Published by:
ARRL
225 Main Street
Newington, CT 06111
203-666-1541
Editor:
Jeffrey W. Ward, K8KA
LOW-POWER TNC FROM GLB
GLB Electronics has released its PK-1L TNC. The PK-1L is an
enhanced, CMOS version of the successful PK-1 TNC. Running with
a CMOS Z80A microprocessor, the PK-1L draws only around 25 ma,
and it can even be run on a 9-v transistor battery. GLB engineer
Ed Jackson, WB2OIF, estimates that the PK-1L would run for about
20 hours from a 9-v alkaline battery. This TNC should be great
for solar-powered operation. As well as supporting all the
features of the PK-1, the PK-1L will have an on-board watchdog
timer, battery-backed-up RAM to store operating parameters in
case of power failure, standard DB-25 connectors for radio and
terminal I/O, a bit on the radio connector that indicates whether
the TNC is connected, and two spare I/O bits available for later
expansion. The TNC comes assembled and tested, in a 4.5 X 6 X 1-
inch, all-metal cabinet. It should make a great portable or
remote station.
From WB2OIF.
TAPR NETWORK CONTROLLER
The availability of several commercial TNCs has spurred the
growth of packet radio. In most areas, several new calls are
heard on the air each week, and Gateway is getting between 50 and
100 new subscribers per issue. How will these new users be
served by the existing network? Moreover, how will network
continue to offer more services in the face of a steadily growing
user community? The single-frequency amateur packet-radio
network is already nearly overloaded in many metropolitan areas.
When the network becomes overloaded, the packet community will
look to its experimenters for new hardware and software to keep
packet radio growing.
Several groups of experimenters, both formal and informal, are
investigating the various hardware and software choices that face
the amateur packet-radio community. The necessary software
includes protocols for network (also called ISO layer 3) and
transport (ISO layer 4) services, and standards for addressing,
routing and mail forwarding. Hardware will be needed for
sophisticated mountain-top digipeaters, remotely operated PBBSs,
satellite teleports and HF-to-VHF gateways. While Gateway is a
newsletter and not a technical journal, much of the packet-radio
news in the next few months will concern the projects and
standards mentioned above. When these topics are covered in
Gateway, we will try to define terms as we introduce them, to
stick to newsworthy items, and to list sources of more extensive
information.
One of the groups interested in what is broadly called
"networking hardware," is TAPR. Now that several hundred TAPR
TNC 2s have been shipped, TAPR is turning to the design and
debugging of a Network Node Controller (NNC). TAPR president
Lyle Johnson provides the following overview of the TAPR NNC
project.
"To keep everyone in the loop, here is the present status of the
TAPR NNC hardware project.
"The schematics are in St. Louis at Interconnections, the company
that does all the CAD (computer-aided design) layout work for
TAPR. If all goes well, we should have artwork for all three
boards by the end of the month!
"Board 1 is the NNC itself. Its present configuration is as
follows:
1) HD64180 microprocessor. This is the CMOS Z80 superset chip
with on-chip DMA (direct memory access), dual UARTs [for
asynchronous communications], 16-bit timers, MMU (memory
management unit) [to manage 512 kbytes of memory] and a
clock. This is the same microprocessor that is featured in
"Build the SB180 Single-Board Computer" in the September
issue of Byte magazine.
2) Dual SIO/2s. This allows four channels of HDLC (high-level
data link control) capability. [HDLC is necessary for packet
operation.] One SIO may be configured (via push-on jumpers)
to have either or both of the channels operate DMA.
3) One PIO (parallel I/O) chip. This provides a parallel
printer port and several lines to fiddle with (for bells,
whistles, buzzers and item 4).
4) A battery-backed-up real-time clock.
5) An SCSI interface, which allows this board to communicate at
high speed with other nearby devices. This will allow the
NNC to be a smart Level Two "front end" for a later board
that can handle all the networking and transport functions
when the network outgrows the capacity of the 64180.
6) Eight byte-wide sockets, for 64 kbytes of battery backed-up
RAM (bbRAM) with jumper selection for 256k bytes of bbRAM!
7) Eight more byte-wide sockets mapped for 32-kbyte parts...This
allows the full 1/2 Mbyte of memory to be put on the board.
8) Expansion interface for Board Three (described below).
"This will run on 5 v, DC and has RS-232 compatible ports for the
two asynchronous channels that are part of the 64180
microprocessor. The serial interface will meet the proposed
WESTNET standard.
"But, you may ask, what good is a NNC without modems? Glad you
asked that!
"Board Two, to go to St. Louis next week, consists of:
1) Multiple XR2206/XR2211 modems. [Like the modems in existing
TAPR TNCs.] Each modem will have a clock generator, a state
machine and a tuning indicator. Board size constraints will
determine whether we only get two modems or if we can squeeze
on four modems per card.
"Yes, these are only 1200-baud (or 300-baud) modems. But, the
local users need a port or two to get in (1200 baud) and long-
haul stuff is going to be HF for a while to come (300 baud now,
perhaps 1200 later).
"Both of these boards to be sized per the WESTLINK standard, so
that they can screw on the side of a 5.25-inch floppy-disk drive.
"Why a floppy drive? I'm glad you asked that, too!
"Board Three is a plug-in floppy-disk interface! The I/O is
mapped to be compatible with the SB180 to allow a simple port of
the "Z" system [disk operating system]. Thus, the NNC can become
its own software development engine, and the hard work of placing
a decent operating system on the NNC is already done and readily
available at a reasonable price.
"If we are lucky, all boards will be laid out by the end of
September. Prototype boards should be populated in October, then
debugged by hardware types while the software types (I hope) will
get cranking on some level-three [networking] software.
"Thank you each and every one for your inputs to date. Keep the
comments coming. Happy packeting!"
[If you are interested in what the 64180 microprocessor is like,
be sure to read the referenced Byte article. For further, still
introductory, discussion of some of the computer hardware
concepts introduced in the above item, read chapters 8 and 19 in
the 1985 ARRL Handbook. -- Ed.]
Via DRNET.
MAPRC MEETING REPORT
The Mid-Atlantic Packet Radio Council (MAPRC) met on September
11, in conjunction with the Gaithersburg, Maryland, hamfest. The
following item is condensed from a report filed by Mike
Chepponis, K3MC, and Bob Hoffman, N3CVL.
The meeting was moderated by Dr. Tom Clark, W3IWI. There were
handouts detailing use of the W3IWI PBBS, MAPRC coverage areas an
the KY2D UNIX system. There were also a map of EASTNET and a
copy of communications between Jeff Ward, K8KA, and Tom Clark,
W3IWI, concerning the KE3Z dual-port digipeater software.
About 40 hams were present. The first hour of discussion
concerned MAPRC organization. A copy of the proposed MAPRC
constitution was distributed and discussed. It was resolved to
collect $24 per year membership dues with $50 per year dues for
clubs. Clubs would get two votes, and individual members would
get one vote. The official purpose of MAPRC is to promote
linking in the Mid-Atlantic area and to make funds and expertise
available to install and maintain digipeaters. Projects funded
by MAPRC will include digipeaters in areas that are critical to
network growth, but without large populations of packeteers.
The group then discussed EASTNET and MidNet coverage. Bob
Hoffman, N3CVL, distributed copies of a map of MidNet and three
possible paths between MidNet and Eastnet were analyzed:
A) Through Northern Pennsylvania and the Buffalo area. This
link would enter EASTNET territory at K3RLI in Wilkes Barre,
Pennsylvania.
B) Through Harrisburg and into Philadelphia. Gary Hoffmann,
AK3P, is already working on this path.
C) Through West Virginia. This route would enter EASTNET at
Cumberland, Maryland, via N8FJB, near Martinsburg, West
Virginia.
The discussion turned to multiport digipeaters and whether it is
best to use the 440-MHz band or the 220-MHz band for 9600-bit/s
links. 440 MHz is preferred by some, especially Brian Lloyd,
WB6RQN, because high quality commercial gear is available for
that band. Groups without well-equipped RF labs have had problems
with the Hamtronics FM-5 transceivers used with the K9NG 9600-
baud modems. On the other hand, many folks already have some
kind of 220-MHz equipment, and Bob, N3CVL, pointed out that there
are other 220-MHz transceiver boards that are fairly easy to
align. It was also noted that WB4JFI-5 is already on 220 MHz.
There was some confusion as to what frequencies are allocated on
220 MHz and 440 MHz for packet. Some people are reluctant to use
the high-speed (100-kHz) channels allocated on 220 MHz for
narrow-band (20-kHz) modems. Others, pointed out the advantage
of being able to change to higher speeds in the future without
the need to change channels. It was resolved to use 220 MHz for
the backbone.
Will Xerox 820s be used as the backbone boards? [In the near
future, the answer is probably "yes," because the multiport
digipeater code runs on the Xerox 820. -- Ed.] Discussion was
inconclusive, but it was noted that the 820 [since it uses an 8-
bit microprocessor] is running out of steam. Possible
alternatives are machines based on the 8088, like the IBM PC, or
machines with large linear address spaces, like the 68000. When
choosing a second-generation networking controller, we must pay
close attention to the availability of inexpensive development
tools -- in particular, compatible computers on which software
can be written and tested.
Brian Lloyd and Phil Karn, KA9Q, discussed some technical aspects
of the network. For example, making FRACK [the time that a TNC
waits before retransmitting a packet] and DWAIT [the time that a
TNC waits after the channel is clear before transmitting] fairly
large numbers really helps out on busy or noisy channels. There
was some discussion of how present TNCs handle retries and of a
couple of bugs in the TAPR TNC-1 software (version 3.3) that add
to network congestion.
There are now three UNIX nodes known to be operating on EASTNET.
They are run by Phil Karn, Jim Kutsch, KY2D, and Brian Lloyd.
Brian gave details on how to get USENET mail onto EASTNET through
these systems.
Via HAMNET.
SOUTHNET II UPDATE
We have received further details on the SOUTHNET II conference to
be held at Georgia Tech, on November 23 and 24. The featured
presentations are:
o Lyle Johnson, WA7GXD, and Pete Eaton, WB9FLW, on the TAPR
Network Node Controller.
o Ed Jackson, WB2OIF, from GLB, will discuss the AX.25
protocol, the PK-1 command set and future GLB projects.
o Demonstrations of all available packet hardware.
o SOUTHNET organizational meeting.
o SOUTHNET awards, including the "SOUTHNET Packeteer of the
Year."
Technical sessions will include a report on the progress of
digipeaters in the SOUTHNET region, an update on the 9600-baud
modem, a status report on 220-MHz linking, discussions of various
networking proposals, sessions covering the W0RLI GateWay and
MailBox, and a roundup of Xerox 820 information. There will also
be an introductory presentation for those just getting started or
interested in packet radio. The conference will be capped by a
banquet on Saturday evening.
Members of GRAPES, the Georgia Radio Amateur Packet Enthusiasts
Society, will provide transportation to and from the airport.
Those that need this service should contact Bill Crews, WB2CPV,
at least 3 weeks prior to the conference. With super-saver air
fares available, you should prepare early to attend SOUTHNET II.
There will be a final mailing of conference information in early
November, including a complete agenda, maps of the meeting
location and information on sightseeing in Atlanta. To get this
information, send an SASE to
Bill Crews, WB2CPV
1421 Hampton Ridge Road
Norcross, GA 30093.
Via DRNET.
CENTER FOR W0RLI ROUTING INFORMATION?
With new W0RLI MailBox and GateWay stations coming on the air
frequently, it is hard for sysops (system operators) to keep
track of all of the interconnected PBBSs. In order for the
system to work, each PBBS must know which of its neighbors to
send mail to for every PBBS in the network. While it is not
practical to keep track of all PBBS users and their whereabouts,
it is a smaller task to keep track of each PBBS and the paths
through which it can be accessed. David Dodell, WB7TPY, has
suggested that someone maintain a database of MailBoxes and
GateWays, allowing sysops to get timely information about new
systems and paths. If you have any comments about this, send
them via packet to WB7TPY @ K7PYK or put them in the mail to
Gateway.
Via HAMNET.
SOUTH TEXAS PACKET
The Texas Tech University Amateur Radio Club will be putting up
several new repeaters this fall. One of the machines will be a
packet-radio digipeater on 145.01 MHz. Ronald Cole, N5HYH, is in
charge of the club's station operations, and he is planning to
put a packet-radio bulletin board system (PBBS) on the air to
stir up some local interest in packet. The other repeaters to be
installed by the Texas Tech club include a duplex repeater
dedicated to experimental modes (on 147.38/.98 MHz) and a machine
on the new 902-MHz band.
Via HAMNET.
SCANNING DIGIPEATER
In a message from W6CUS-1 PBBS that was posted on CompuServe
HAMNET, Don Simon, NI6A, reports that there is now a scanning
digipeater serving the San Francisco area. The digipeater,
N6IJP-1, is located about 2000 feet above the city of Angwin,
California. It scans all packet frequencies between 145.10 MHz
and 145.09 MHz, stopping for six seconds on each frequency. Once
it has been accessed on a given channel, the digipeater stays on
that channel until it has not been accessed for two minutes. If
your group would like to have a digipeater that serves occasional
contacts on several frequencies, this scheme may be the way to
go. For further details on how the scanning digipeater is
implemented, contact
Randy Fischer, N6IJP
455 Bay Street
Angwin, CA 94508.
Via HAMNET.
NEW CALIFORNIA DIGIPEATER
A new digipeater in northern California, WA6YNG-1, should provide
communications between southern Oregon and Sacramento. The
digipeater was installed in the Mt. Shasta area, at an elevation
of 7000 feet. Both county and state Offices of Emergency
Services cooperated to make this high-altitude site available.
Operating on 145.01 MHz, WA6YNG-1 can be reached through W6AMT-7.
When the W6AK digipeater comes on the air, WA6YNG-1 will be
moving to 145.05 MHz.
Via HAMNET.
HF GATEWAY IN COLORADO
Dave Shavey, K0HOA, now has an HF gateway operating in Colorado
Springs, Colorado. Dave is using the W0RLI GateWay software, and
provides full service, including message forwarding 24 hours a
day. Like most of the other HF gateways, K0HOA is on 14.103 MHz.
We hear that the midwest is not far behind, with the crew at
WB9FLW in St. Louis, Missouri, moving quickly to get a W0RLI
GateWay and MailBox on the air.
Via N0CCZ, DRNET.
ST. LOUIS MOVES TO 145.01 MHZ
The St. Louis Amateur Packet Radio Club (SLAPR) is moving its
packet operation from 147.555 MHZ to 145.01 MHz. This change has
been coordinated in Eastern Missouri and Southern Illinois. Our
coordination agreement states that the move will be complete by
October 15th.
Via WB9FLW, DRNET.
REPRODUCTION OF GATEWAY MATERIAL
Material may be exerpted from Gateway without prior permission,
provided that the original contributor is credited and Gateway is
identified as the source.