dna@dsd.UUCP (05/23/84)
K5JB06.TXT 15-May-84 8580 Accesses: 18
Keywords: K5JB PACKET
Column by Joe Bushwell, K5JB, describing his Packet Radio
experiences at the 1984 Dayton HamVention.
DAYTON - PACKET UPDATE
During the visit to Dayton, I got an opportunity to attend the
Packet Radio presentation and meet some of tne movers and
shakers. Two people whom I especially admire, Lyle Johnson,
WA7GXD, President of Tucson Amateur Packet Radio Corp. and Gil
Boelke, W2EUP, of GLB Electronics took turns on the platform
along with representatives from St Louis (SLAPR), Dayton (Miami
Valley FM Association), AMRAD, and the ARRL (Paul Rinaldo, W4RI)
. After the presentation I got to meet Gil and ask him if he
minded my modifying the code in the PK-1. He said heck no and
to call the company later to get specific answers to my
questions. (...which I subsequently did, and to which they
responded in writing with specific answers to my questions.
This is without any knowlege of my reputation as an agitator/
instigator). I chose my questions carefully though because
firmware is a really touchy subject. I have a real warm feeling
about the GLB folks. I think their aim is to please. The next
morning I got to visit briefly with Lyle Johnson, WA7GXD, who is
certainly one of the most energetic and creative persons I have
encountered. I can understand how he has been such a sparkplug
for the TAPR activity...).
There are currently six sources of TNC's or alternate methods of
getting into Packet Radio; AEA's, Ashby and Son's (K2TKN &
KA2OEG), GLB's, Bob Richardson's (W4UCH), TAPR's, and VADCG's.
Someone in AMRAD is working on a TNC called a PAD (Packet
Assembler - Disassembler) which is apparently slanted toward S-
100 Bus Computer systems. AEA (Advanced Electronic Applications
) announced its TNC at the Hamvention but had no representative
on the program. AEA has a handsome design that looks like a
development from the TAPR TNC, for a price of about $500. I saw
no evidence of the Ashby and Son TNC at the show but I could
have missed it. Likewise, I saw no evidence of Bob Richardson's
(W4UCH) Radio Snack model 1 and 3 program, though it was
respectfully mentioned by Gil Boehm from GLB. It seems that
since Bob and Gil are neighbors, they regularlly communicate on
program design questions but developed their programs
independently.
Gil defended the GLB's Spartan design by statingthat he Value
Engineered the problem to strike an economical compromise in
hardware v.s. software trade-offs and decide whether to add
audio shaping and filtering to the PK-1. He was challanged by a
later AMRAD speaker who felt that the five bucks the filter
would have cost would be a cheap 1.5 dB improvement if one were
working satellites. Later when Gil answered a question about
whether the TP-1 would run 300 Baud for HF operation, he
answered both points when he stated that its design goal just
didn't include those applications.
There still is no absolute answer to the question, "Which is
better, the GLB or the TAPR?" It depends for which use the TNC
is intended. For the benefit of you who have not been following
this discussion since last January, the following salient
differences exist between the two.
The TAPR TNC costs approximately $250.00 in kit form, has an
optional cabinet for about $70.00, runs Vancouver and AX.25
protocol and runs on 120 Vac.
The GLB costs approximately $150.00 assembled, has an optional
cabinet for about $25.00, runs both protocols and runs on 12
Vdc.
The TAPR TNC uses hardware (Western Digital 1933 SDLC) to make
and read packets, the GLB TNC does all in software (firmware).
The hardware v.s. software is the major difference between the
two and accounts for most of the cost difference. The only
performance disadvantage I have found in the GLB approach is the
packet reception disruption that occurs if a key is touched on
the terminal during the reception process. The keyboard has
interrupt priority and the Processor must service the keyboard
at the expense of its packet activities. Gil (and the
accompanying TP-1 literature) advises that the computer, which
is most often the case used to act as a terminal, be programmed
to accept the input off-line from the TNC and send it, when
ready, to he TNC at a high data rate, reducing the probability
of a clash with packet operation. The TAPR TNC, with its packet
handling hardware, is not placing any strain on the processor to
make the data input and output process appear transparent to the
operator. Mark, WD5DYI, is now using his TP-1 regularlly and
this disadvantage is much more apparent as he creates his
epistle sized packets. I had not noticed the problem so much
because I usually send rather terse packets, or type in short,
five character, bursts of about a thousand words per minute,
spending the rest of the time looking for the next five keys!
Packets easily slip in while I am looking. Mark is not a
programmer but he will eventually want to do the line editing on
his computer. I toyed with the idea on the Model-100 but got
bogged down in the creeping elegance of interrupt handling from
the RS-232 port.
The TAPR documention is awesome and contributes to a portion of
the price difference. Since this is a kit, construction and
adjustment instructions cause some of the book's volume but, in
fact, the TNC is one of the most thoroughly documented pieces of
hardware I have ever seen.
The GLB documentation has come a long way since I bought mine.
It is now 35 typewritten pages bound looseleaf between a couple
of nice, blue covers, with a command cross reference chart
printed on the back for easy reference. Also, it now has a
schematic! The manual is certainly adequate to support the PK-1
and its many features.
The TAPR hardware was designed for flexibility in order to
enable as many conceivable modifications as possible so that it
could be used for other applications. It has a socket for an
external modem in case one is impatient with 1200 Baud, or just
wants to try external hardware. Its audio filtering can be
easily modified for HF use by installing suitable resistors on
DIP headers. As already mentioned, with the GLB you roll your
own if you want audio filtering or shaping. If you run out of
Packet things to do with the TAPR TNC, you can get the EPROM
programmer from TAPR, plug it into the socket provided on the
TNC and do your own thing. These functions were simply not part
of the requirements in the GLB design concept. (Like a friend
of mine says, "If you wanted the Penguin to fly you should have
included it in the specs!"
The TAPR TNC can be made to operate the radio side from 50 to
4800 Baud half or full duplex while the GLB PK-1 is restricted
to 600 and 1200 Baud, half duplex. N5MS and I experimented with
600 Baud over a noisy path and it didn't seem to be any better
than the 1200 Baud speed. To gain full benefit we should have
modified the firs (Mark, WD5DYI, Joe, WA5FLT, and I) rate them
in day-to-day use. So far, I have enjoyed my particular
application, using it portable with the Model-100 computer as a
terminal. Joe, WA5FLT, also has a TAPR TNC and can provide his
opinions as soon as he gains experience. He intends to put his
into remote operation. This is a special feature incorporated
in the GLB that requires adding only a transceiver, watchdog
timer and transmitter timeout time. The TNC receives its
commands via the packet channel.
The decision on which TNC to buy is a personal one. One has to
first determine the function he wishes to perform. Lyle,
WA7GXD, put it most succinctly when he commented that the TAPR
TNC was intended for the bearded experimenter down in the
basement (He sports a handsome bush). On the other hand, Gil
claims the GLB PK-1 was not intended to be all things to all
people.
The following addresses may be useful for more information:
Advanced Electronic Applications
P.O. Box C2160
Bldg O & P - 2006 - 196th SW
Linnwood WA 98036-0918
Bill Ashby & Son
P.O. Box 332
Plukemin NJ 07978
AMRAD
1819 Anderson Rd
Falls Church VA 22043
GLB Electronics
1952 Clinton St
Buffalo NY 14206
Richcraft Engineering Ltd
#1 Wahmeda Industrial Park
Chatauqua NY 14722.
Tucson Amateur Packet Radio, Corp.
P.O. Box 22888
Tucson AZ 85734
VADCG
818 Rondeau St
Coquitlam, BC, Canada
Joe Buswell, K5JB