..
`
`April 1988
`Issue *31
`
`Published by:
`Tucson Amateur Packet Radio
`PO Box 22888
`Tucson, AZ 85734
`Phone: (602) 323-1710
`
`PACKET STATUS REGISTER
`Tucson Amateur Packet Radio Cor oration
`
`PRESIDENT'S CORNER
`
`by Andy Freeborn, Nf2ICCZ
`For the first time since February 1983 TAPR has a new President. Now if you looked
`at TAPR as a run-of-the-mill ham club you just might yawn at that statement. You
`shouldn't. Very few presidents of any type of ham club hang in there for five years or
`enjoy the stature to be repeatedly reelected.
`
`Lyle Johnson, WA7GXD, your President for the last five years, is one of a rare breed
`of persons who is a highly skilled technician, possesses an intense dedication to the
`Amateur Radio Service and has been willing to put up with the routine mundane
`chores of running TAPR which is, in essence, a business.
`
`Editor:
`Scott Loftesness W3VS
`16440 Rustling Oak Court
`Morgan Hill, CA 95037
`CompuServe: 76703,407
`He has done this, not for just a couple of years, but for FIVE years. He has done it
`-------------1 while holding a full time job. He has done it while he and his wife Heather, N7DZU,
`have been raising a family of six children, the oldest now 16. Lyle, it seems, has
`In This Issue •••
`packed a whole career into a five year period.
`• New President's Comer by
`Andy Freeborn, N0CCZ
`• Dayton HamVentlon Packet
`Activities
`• T APR to Coordinate Group
`Purchase of PS-186 Kits
`• Digital Signal Processing:
`Hardware Developments
`• Applied DSP: The Telebit
`Trailblazer Modem
`• Digital Signal Processing:
`Book Recommendation
`• PS-186 Prolect Status from
`N6NKF
`• Digital Signal Processing:
`Software Developments
`• AA4RE's "In the Mailbox·
`Column on Packet Radio
`Bulletin Board Systems
`• ROSE Switch Information
`from RATS
`• Software Available from the
`TAPR Office
`• And much more!
`
`He has gone through the ups and downs of TAPR as its leader. Black Thursday in
`1982, when it was discovered that all 174 of the TNC1 beta boards were defective,
`was a bad period. The long period just before the sale of the TNC2's, when TAPR
`was incredibly big bucks in debt certainly was a worrisome time. Whenever there was
`criticism of TAPR It was always Lyle that took the heat.
`
`There were good times, too, as when Lyle accepted the Dayton HamVention first time
`award for Outstanding Technical Achievement on behaH of TAPR. August 21 st 1985
`was a great day! At 9 AM the TAPR office, manned by volunteers, started taking
`orders for the TNC2. Within hours the telephone company called to find out what in
`blazes we were dOing. Hams were calling from all over the country to the single T APR
`telephone and had put their system into gridlock. I guess one of his greatest periods
`must have been when he was able to start signing checks to payoff that debt.
`
`Why would a man put himself through all of this? I think it was because Lyle Johnson
`wanted to contribute something to the hobby that he loves so much. If that was his
`motivation, he certainly succeeded.
`
`Lest you think that this is an eulogy. be assured that Lyle is alive and well. Wild
`horses couldn't pull him away from the hobby (or TAPR). He is now devoting his time.
`freed of administrative burdens, to the much loved technical side of new and exciting
`TAPR development programs.
`
`The entire world of amateur radio owes Lyle Johnson, WA7GXD, a huge debt of
`gratitude.
`
`Next Scheduled PSR Issue:
`June 1988 Publication D,Flte
`Editorial Deadline: 5/20/88
`
`What an act to follow! Your Board of Directors, in their infinite wisdom, felt that it was
`time to keep the techies in the tech jobs and a non-techie in charge of T APR's day to
`day operations. And that is why you see my name and call at the head of this column.
`
`Continued on page 2
`
`

`
`NON-TECH TOPICS
`by Andy Freeborn N0CCZ
`
`DSP Excitement
`
`As the annual meeting progressed in
`February. and as more and more
`information concerning TAPR's Digital
`Signal Processing work was disclosed
`you could just feel the excitement
`building among the attendees. Several
`folks commented to me that the DSP
`work could result in an impact on
`amateur radio as significant as
`TAPR's TNC1 and TNC2 develop(cid:173)
`ments.
`
`TAPR Office Change
`
`President's Corner
`Continued from Page 1
`I am a 30 year Air Force retiree. I flew
`p·47's in WWII in Europe, was shot
`down on my 39th dive bombing
`mission in early 1945 and spent the
`balance of the war as a guest of der
`Fuehrer. I subsequently spent 20 of
`those AF years in Communications·
`Electronics organizations. I got my
`ham ticket in 1980. then bought a
`computer and saw the possibilities of
`the marriage of the two. In early 1982
`I heard of a group in Tucson starting
`to DO something about it. I contacted
`them and have been associated with
`TAPR ever since. My wife of 45 years,
`Vernie, says that if she were a CB'er
`her handle would be Packet Radio
`Widow. Now you know more about me
`than you were ever going to ask.
`
`As President of TAPR I pledge to the
`membership the best that I can do.
`
`In an effort to conserve our resources
`for ongoing digital development work
`the TAPR Board of Directors voted to
`close its present officelWarehouse
`arrangement. The only difference that
`the membership will observe is a
`change in office telephone number
`73, Andy· N0CCZ
`---':-~------------1 and P.O. Box address. The office will
`DAYTON HAMVENTION:
`be located in one of the rooms in the
`Packet Radio Activities
`home of our employee, Cris. You will
`still hear her pleasant voice when the
`phone is answered. We are able to
`make this change because the
`company that has been doing kit
`assembly for us will also take on some
`of the functions that required only a
`small part of the warehouse that we
`are leaving.
`
`The Packet Forum Moderator for the
`1988 HamVention is Bob Neben,
`K9BL. He reports that the planned
`schedule of events will be as follows:
`
`Friday, 29 April, Room 3
`
`1 :00 ·Packet Radio Introduction" by
`Dave Peterson, N7BHC.
`
`2:00 "Level 1 Discussion - Modems"
`by Lyle Johnson, WA7GXD.
`
`3:00 "Pacsat and DSP" by Bob
`McGuire, N4HY.
`
`4:00 "Networking· How to get
`started" by Dave Toth. VE3GYQ and
`Chris Sullivan. VE:3NRT.
`
`TNC1 Ver 4 Firmware
`
`It is time to put to rest the matter of
`Version 4 firmware for the TNC 1. The
`Board of Directors has concluded that
`existing firmware is adequate and that
`further expenditures of funds or
`technical talent. in view of the benefits
`to be obtained, is unwarranted.
`
`What's a TAPR?
`
`stress that it came from within the
`I
`amateur community, not industry.
`I Thanks Terry
`I
`
`The Board of Directors, by unanimous
`I
`resolution, extended its heartfelt
`thanks to Terry Price, N6HBB. for the
`great job she did as Treasurer of
`TAPR. Terry set up accounting
`procedures and then prepared
`quarterly reports for the BoD that are
`second to none for this type of
`organization.
`
`Hello, SecretaryITreasurer
`Scott
`
`Scott Loftesness, W:3VS, the editor of
`the PSR which you are reading is the
`new SecretarylTreasurer for T APR.
`Scott is taking over a big job from
`Terry. In addition to his regular job he
`is also Sysop of Ham Net on CompuS(cid:173)
`erve. It seems that folks like Scott,
`who continually do more than should
`be expected of them, just keep doing
`more and more. I am just mighty
`happy for TAPR that Scott is there.
`
`Our Membership Rolls
`
`TAPR membership has dropped
`consistently since the introduction of
`the TNC2. We feel that some of this is
`due to a relatively low TAPR profile in
`the recent past (see also What's a
`TAPR? above). We'll be instituting
`several new procedures soon to
`reverse this trend. Included will be
`membership renewal reminders and a
`new membership drive program.
`
`Membership Dues and Publica(cid:173)
`tion Costs
`
`The Board of Directors has directed
`the Executive Committee to review the
`costs of publishing the PSR, the
`member dues structure and related
`costs. More on this next issue.
`
`Be sure to note TAPR's new tele(cid:173)
`phone number on the cover of this
`issue. The phone number change
`was required as part of the offic.e
`changes recently implemented In
`Tucson. You'll still find TAPR's Office
`Manager, Cris, awaiting your calli
`
`TAPR's new office telephone
`number Is 602·323-1710.
`
`Friday evening packet getogether
`dinner arrangements have not been
`made, but is assumed to be at
`McNasty's.
`
`Old time packeteers (of at least a
`couple of years) will recall when the
`terms T APR and TNC were almost
`interchangeable. Since that time the
`explosive growth of packet radio has
`brought multi thousands of new
`73 Bert NaNN @ NaNN -sk·
`.:.....-----......::.---------j packeteers onto the scene. When you
`See You at the 1988 Dayton
`ask someone if he is a member of
`HamVention! TAPR will be repre-
`TAPR and he says "What's that?",
`sented at the AMSAT booth this
`take the time to explain to him the
`year.
`origin of his TNC2 clone. Be sure to
`
`Page 2
`
`April 1988
`
`PSR
`
`Issue #31
`
`

`
`operation. perhaps using a full duplex
`HDLC port with direct-memory access
`(DMA) circuitry for performance.
`
`The DSP portion will prpbably be
`based on a Texas Instruments
`TMS320C 15 DSP chip. With proper
`sohware. it should be capable of
`providing solid performance as a
`several-hundred bits-per-second (bps)
`HF modem, as a PSK satellite and
`terrestrial modem, hopefully as a
`"K9NG-compatible" 9600 bps FSK
`modem. as an MSK modem, as a
`2400 bps QPSK modem, and as a
`normal 300 and 1200 baud FSK
`modem.
`
`I
`
`Finally. the D~~-1 project will probably
`emerge as a JOInt venture of AMSA T
`, and TAPA, combining the resources
`from both organizations and benefit-
`ting both through kit sales and,
`hopefully, OEM arrangements similar
`to that of the TNC 2. The TNC 2 OEM
`agreements have proven very
`successful for TAPA, the manufactur(cid:173)
`ers and the Amateur community.
`
`T APR to Coordinate Group
`Purchase of PS·186 kits.
`
`I
`I
`
`In addition. it should be possible to
`configure it to do a good job demodu(cid:173)
`lating WEFAX and satellite weather
`transmissions, as well as a multi-grey
`scale SSTV modem. Then, of course,
`there's CW, digital voice, and any
`number of other possibilities.
`
`TAPR is in the process of organizing a
`group purchase of bare boards and
`hard to find components for the PS-
`186 packet switch in order to make
`, this new technology available to the
`advanced experimenter for sohware
`and network development.
`
`HARDWARE HAPPENINGS
`by Lyle Johnson. WA7GXD
`
`Old readers of PSA (or, perhaps I
`should say. "readers of early issues of
`PSA ... ~) will recognize this column. It
`has been out of print in PSA for
`several years due to my authoring
`other columns.
`
`Fortunately for all of us, Andy Free(cid:173)
`born is now writing the President's
`Corner.
`
`Unfortunately, you're not completely
`rid of me as I plan to keep my finger in
`the pie by dabbling in various projects
`and generally stirring things up!
`
`I hope to use this column as a means
`of disseminating technical information
`on the hardware aspects of T APA(cid:173)
`sanctioned and -funded projects in
`which I am involved. In this respect, it
`may resemble some of the Beginner's
`Corner articles I have written in the
`past.
`
`As in earlier years, I also hope to
`break down some of the major
`projects into a sort of block diagram
`discussion, highlighting features and
`explaining some of the reasons
`behind design decisions that ulti(cid:173)
`mately affect packet operators and
`network users.
`
`The big project that will affect you first
`is the Digital Signal Processor (DSP)
`initial project, currently dubbed DSP-1.
`As has been mentioned in recent
`issues of PSA by Bob McGwier, DSP
`techniques will allow us to build
`configurable modems and other
`analog/digital devices which com(cid:173)
`binehigh performance with reasonable
`cost.
`
`A suite of three projects is currently
`envisioned.
`
`DSP-1 will likely be a standalone
`device capable of working with most
`TNCs as an external modem. It
`appears that it may include an 8088-
`class general purpose processor to
`allow it to function as a TNC on its
`own, as well as operate numerous
`other modes. If it does include a TNC
`function, the hope is that it wilrbe
`capable of really high speed
`
`Sohware development can be done
`for it with the Delanco-Spry Model 10
`DSP board for the IBM PC. Support
`for downloading code into DSP-1 will
`also be provided via a serial port, so
`anyone can write code for the unit.
`Hopefully, updated modems can be
`distributed via your local packet BBS!
`
`It is still too early to give any price or
`availability estimates, other than to
`say that design work is proceeding
`along and prototype units may be in
`the hands of a few coders and testers
`by Summer.
`
`In the next issue, I hope to break
`down the design of DSP-1 in some
`detail for you.
`
`As good as the TMS32010 is for many
`amateur applications, there are some
`applications that simply require more
`horsepower. more memory or more
`resolution than DSP-1 can provide.
`
`For this reason, it may be possible to
`configure the unit to support a pair of
`DSP processor boards. A more
`powerful processor bo'ard,perhaps
`based on the TMS320C25 or the
`Motorola DSP56001, will likely be
`designed for the DSP-1 aher the initial
`configuration is debugged, placed in
`1 service and had a reasonable demand
`I generated for it.
`
`The PS-186 is a 5-port high speed (
`> 1 MbiUsec) packet switch that was
`designed by WB6HHV, KA6IQA. &
`N6NKF. The card is described in
`detail in the AAAL 6th Networking
`Conference Proceedings. Briefly,lhe
`PS-186 is built around an 8 MHz
`80186 processor and up to a 1
`megabyte of memory. 110 support is
`provided by two 8530 SCC chips
`which are interfaced with DMA. This
`provides 4 HDLC/ASYNC/SYNC
`ports. the fifth port is ASYNC only.
`Also provided are an SCSI port (for
`disk or multi-board interface). a real
`time clock. a watchdog timer. and a
`remote reset circuit.
`
`Advanced Electronic Applications
`(AEA) has acquired the rights to
`produce the card commercially. TAPR
`is organizing a group purchase of bare
`PCBs and the necessary PALs
`directly from AEA.
`
` The kit will include an electrically
`
`I,
`
`tested 6 layer PCB. programmed
`, PALs. difficult to find LSI components
`and a documentation package. A
`complete list of the parts is included ,r"'
`the documentation. The package Will
`not contain any of the readily availab's
`components such as connectors,
`LEOs. resistors, or commonly avail-
`
`I able chips.
`
`I
`
`Continued on page 7
`
`April 1988
`
`PSR
`
`Issue #31
`
`Page 3
`
`

`
`APPLIED DIGITAL SIGNAL
`PROCESSING: The Telebit
`Trailblazer Modem
`
`By Michael Ballard, UNIX Program
`Manager, Telebit Corp.
`
`{Editor's Note: This article provides
`some interesting background on rea/(cid:173)
`world uses of digital signal processing
`technology. It was ported here from
`Usenet.}
`
`I would like to provide some back(cid:173)
`ground for Unix users considering the
`use of Telebit's TrailBlazer Plus high
`speed dialup modem. I served as
`project manager and principal
`programmer for Telebit's protocol
`support developement. The UUCP
`"g~, Kermit, Xmodem and Ymodem
`protocols are directly supported in the
`TrailBlazer modem's firmware. Peter
`Honeyman, co-developer of ArT's
`HoneyDanBer/BNU UUCP, coded
`those portions of the TrailBlazer
`firmware which support the Mg"
`protocol.
`
`The Telebit modem employs a
`patented multicarrier modulation
`scheme coined DAMQAM (Dynami(cid:173)
`cally Adaptive Multicarrier Quadrature
`Amplitude Modulation). A CRC-16
`based sliding window protocol with
`selective retransmission runs on top of
`this modulation scheme insuring data
`integrity across the phone line. This
`telephone line protocol is known as
`the Packetized Ensemble Protocol or
`PEP. PEP is the trademark by which
`all modems employing this technique
`can be recognized.
`
`This technique (DAMQAM) divides the
`voice bandwidth into 51 1 individual
`channels each capable of passing 2,
`4, or 6 bits per baud based on the
`measured characteristics of the
`individual frequencies associated with
`each channel. On a typical phone
`connection, the modem uses a subset
`of about 400 of those channels.
`
`Each time the modem connects to a
`circuit established on the dialup Public
`Switched Telephone Network (PSTN),
`the TrailBlazer measures the quality of
`the connection, and determines the
`usable subset of the 511 carriers. The
`aggregate sum of bits modulated on
`
`this subset of carriers multiplied times
`the baud rate yields a bit per second
`rate that on a local telephone connec(cid:173)
`tion (i.e. round trip through your local
`telco) is 18031 bps. This 18031 bps is
`then reduced by about 20% to allow
`for the CRC overhead, to about 14400
`bps of data throughput.
`
`Long distance line quality varies with
`location and carrier, but you can
`expect this number to be in the 10000
`to 17000 bps range under most
`conditions domestically. By chOOSing
`a high quality long distance carrier,
`you will ensure the best throughput
`overall.
`
`The modem operates at 7.35 and
`88.26 baud, transparently changing
`baud rates to accomodate the pace
`and quantity of data traffic. When in
`"interactive mode" the modem sends
`data using 11 msec packets (which
`run at 88.26 baud). Each packet
`contains 15 bytes of data. In "file
`transfer mode" the modem uses 136
`msec packets (that transfer at 7.35
`baud) that contain 256 bytes of data.
`The TrailBlazer decides which packet
`size to use on an ongoing dynamic
`basis. No intervention from the user is
`required.
`
`At lower speeds, such as 300,1200,
`and 2400 bps, the TrailBlazer pro(cid:173)
`vides emulation (performed in the
`DSP section, not by a Mchip" modem)
`to support these standards. The 300
`bps standard is called Bell 1 03C. At
`1200 bps, two standards exist, Bell
`212A and CCITT V.22. Both are
`supported. At 2400 bps, the standard
`is called CCITT V.22 bis. These
`speeds are all available with or without
`MNP Class 3 Error Correction.
`
`The TrailBlazer employs a Motorola
`68000 and a Texas Instuments
`TMS32010 digital signal processor to
`accomplish this performance. Be(cid:173)
`cause of this substantial computer
`horsepower (about 7.5 MIPS), the
`TrailBlazer is really a communications
`processor, rather than a conventional
`modem.
`
`The software defined architecture
`produces a flexible product platform
`that allows broad feature development
`capabilities while allowing the
`product's installed base to benefit
`
`; from those developments by installing
`i upgrade EPROM sets.
`
`All four protocols (Kermit, Xmodeml
`Ymodem, UUCP), V.22bis support,
`MNP at low speeds, multiple releases
`to improve the interactive performance
`(earlier TrailBlazers utilized only one
`baud rate), a multitude of RS-232
`behavior related features, leased line
`capabilities, remote command
`processor access, echo suppressor
`compensation, increased data rates,
`and a myriad of user requested
`features have found their way into
`current production modems and are
`available to earlier revisioned modems
`via the EPROM uprgrade kits.
`
`PEP modems provide a full duplex
`serial interface to an attached
`computer, however they employ a half
`duplex implementation on the tele(cid:173)
`phone line. Telebit refers to this half
`duplex technique as "Adaptive
`Duplex". As the name implies, the
`ownership of the line (i.e. the ability to
`transmit) adapts to the quantity of data
`available to send at any single
`moment. Maximum efficiency is
`achieved by sending data in a
`nonstop data stream at 19.2Kbps
`relying on serial interface flow control
`to moderate the data flow into and out
`of the modem.
`
`This allows the maximum amount of
`data to be available every time a
`transmitting modem takes ownership
`of the line. In this way the modem,
`not the DTE. controls the line turn(cid:173)
`arounds. The protocol provides a
`ceiling at about 3k of sent data before
`a transmitting modem must give up its
`turn and allow the other modem an
`opportunity to send. A continuous
`19.2Kbps data flow into the modem is
`required to ensure that there is always
`3k of data to send each time a
`transmitting modem takes its turn. The
`serial interface speed must exceed
`the telephone line speed, potentially
`18,031 bps, or the maximum efficiency
`of the modems can not be reached).
`
`UUCP's "gft protocol behavior on
`dialup lines was a clear contradiction
`of the desired behavior with the PEP
`protocol. "gft sends 3 small data
`I packets at time and then waits for the
`i remote UUCP to ACK or NAK their
`I receipt. The resulting throughput
`
`Page 4
`
`April 1988
`
`PSR
`
`Issue #31
`
`

`
`when using UUCP and "g" with the
`TrailBlazer was only a little better than
`a standard 1200 bps modem. This
`was unacceptable.
`
`What did we do about it?
`The TrailBlazer can be configured to
`"spoof" the protocol by sening a
`register (5111) to one of several
`values. The spoof can support four
`different protocols: UUCP "g",
`Xmodem, Vmodem, and Kermit.
`
`·Spoof" means to fool the various
`protocols into thinking that they are
`getting their acknowledgment packets
`from the remote computer, when in
`reality they are getting them from the
`modem.
`
`All of these protocols ale what are
`commonly referred to as "send and
`wait" protocols. This type of protocol
`builds a packet in computer A. sends
`it out through the modems. where it is
`received by computer B. Next.
`computer B looks at the packet to
`determine whether or not it arrived
`intact. If it did, it sends an ACK
`(acknowledgement) packet back to
`computer A. If it did not arrive intact. it
`sends a NAK (non-acknowledgement)
`packet. In either case. computer A
`can't send the next packet out until it
`gets the ACK from the first packet.
`This is slow I
`
`Since our modems are error-frea
`between the modems. the only place
`data could get "broken" is between the
`modems and their respective comput(cid:173)
`ers. Let me draw the connection
`diagram below:
`
`Co <-> To <-> Tb <-> Cb
`
`Co - Cocaputer R
`Cb • COllIputer B
`To • Teleblt nodelll R
`Tb • Teleblt nodeca B
`
`••• RS-232 Coble
`Phone Line
`
`When we are running our protocol
`support. we look at tha packet coming
`from Ca. Ta checks the packet fc,r
`validity and sends tha ACK or NAK.Ca
`can begin building tha next packe t
`immediatly upon receipt of Ta's ACK.
`This results in Ca building and
`sending packets as fast as it can ..
`
`Many packets are now forwarded to
`Th. Th now delivers the packets to
`Cb, observing the rules of the proto(cid:173)
`col. Tb will deliver the next packet or
`retransmit the previous packet based
`on the ACK or NAK received from Cb.
`Cb ACKs and NAKs are then thrown
`away so as not to return to Ca.
`
`DIGITAL SIGNAL
`PROCESSING:
`Book Recommendation
`
`[Editor's Note: The following message
`is from the HamNet Forum on
`CompuSefVe.j
`
`Protocol support can be configured 10
`run in parallel with data compression
`enabled. The real world result of this
`is to increase protocol transfers from
`2-3 Kbps to 10-19.2 Kbps.
`
`#: 70570 S5/Amateur Satellites
`23-Feb-88 07:53:50
`Sb: #70474-#DSP and WEFAX I
`Fm: Bill Coleman AA4LR 76067.2327
`To: Bill Bard 75366.2557 (X)
`
`This covers most of the commonly
`asked questions about the TrailBlazer.
`If any of the above information is
`unclear. or you have questions
`regarding other aspects of modem
`technology or performance, send mail
`to:
`
`Richard Siegel
`Senior Systems Engineer
`Telebit Corporation
`ARPA: telebitlmodems@ames.ARPA
`UUCP:
`{uunet,ames,hoptoad} Itelebitlmod e ms
`
`TAPR AT DAYTON
`
`TAPR will once again be officially
`represented at the Dayton
`Hamvention. We will be sharing a
`booth through the courtesy of AMSAT.
`There will be a limited number of PSK
`modems available for sale. and orders
`may be placed for additional items.
`such as the K9NG 9600 baud
`modems.
`
`We hope that this booth, as well as a
`booth in the courtesy area, manned by
`the OHIO PACKET RADIO COUNCIL.
`will serve as rallying points for packet
`get-togethers.
`
`Of course, no TAPRlPacket gathering
`would be complete without a trip to a
`Malibu raceway, and tentative plans
`include a trip to the Miamisburg track
`on Saturday nite. There will ba a
`packet "dinner"/gel-logathar. likaly at
`tha now-famous McNasty's (famous
`only because of our 2 previous
`bashesl ...
`
`So, be there, or be square! Packet
`Pete Eaton and Dr. Death (AKA
`VE3GVQ) will be leading an all-star
`cast in manning the booth!!!
`
`I just found a great book on DSP.
`"Designing Digital Filters" by Charlas
`S. Williams. Prentice-Hall 1986. ISBN
`0-13-201856-X. This book is for 'he
`rest of us· who are mathematical
`wizzos. All that is required is a working
`knowledge of trigonomitry and
`calculus. The book even contains a
`review of complex numbers and an
`introduction to analog filters. All
`notational conventions are thoroughly
`explained before they are used. and
`the book introduces each new concept
`with a easy-to-understand example.
`
`The book stans with an overview of
`the impetus around designing digital
`filters, discusses the issues associ(cid:173)
`ated with digital filtering (aliasing.
`quantization). The next chapter
`reviews the mathematical concepts of
`frequency response. The next two
`chapters discuss the design and
`implementation of nonrecursive filters.
`then two chapters on recursive filters.
`The final two chapters talk of polyno(cid:173)
`mial modeling of digital signals and
`the OFT and FFT.
`
`An excellent book for those of us who
`are not EEs. nor Math majors. In all. a
`very readable book!
`
`73. Bill
`
`About your membership ...
`
`PSR is one of the major benefits of
`membership in TAPR. If you find PSR
`of value. please remember to check
`the mailing label on this issue for your
`membership expiration date. TAPR
`does not send you any other notice of
`membership expiration.
`
`If you want to continue to receive PSR
`uninterrupted. please check your
`membership expiration NOWI
`
`April 1988
`
`PSR
`
`Issue #31
`
`Page 5
`
`

`
`DSP UPDATE
`by Bob McGwier, N4HY
`
`The DSP project is moving along
`nicely with basically good news all
`round. All the hardware for the initial
`team has been distributed and the
`software and manuals for that
`software written by team members
`was distributed last week. We hope to
`begin to see some return from the
`folks in the project soon. Our first
`tasks are still emphasizing those
`areas that will have the largest and
`most immediate impact on amateur
`radio in general and packet radio in
`particular. Tom spent his entire week
`away for Christmas "Vacation" putting
`the initial offering of software together
`and sending out the final DSP boards.
`Diskcopying 120 disks, photocopying
`20 sets of manuals of one variety or
`another, is not my way to spend
`Christmas but Tom apparently
`survived the worst part of managing a
`project like this. Courtney Duncan,
`N5BF, has agreed to take on the
`documentation and record keeping
`task for this project. Before we go on,
`let me tell you who all the team
`members are that are working on the
`software.
`
`Tom Clark, W31WI
`Bob McGwier, N4HY
`Richard Allen, WSSXD
`Mike Parker, KT7D
`Dan Morrison KT7B
`John Shew, N400
`Courtney Duncan, N5BF
`Randy Cole, KN6W
`John Conner, WDOFHG
`Mike Chepponis, K3MC
`Phil Karn, KASO
`Hugh Aller, WD80BL
`Andy Demartini, KC2FF
`Junior De Castro, PY2BJO
`Alberto Zagni, 12KBD
`John Molnar, WA3ETD
`Dave Borden, K8MMO
`Ned Johnson, K1 NJ
`Fred Williams, W1 FWJ
`Ken McGuire, WA2VFN
`David Backus, KA8ZLA
`Mike Lamb, N7ML
`AI Chandler, K6RFK
`Barry McLarnon, VE3JF
`Rick Hambly, WB2TNL
`James Miller, G3RUH
`Steve Sagerian, KAOYRE
`Dave Truly, NN2Z
`
`These are listed in the order that they
`joined the project and received the
`TMS32010 Delanco Spry boards with
`the exception of Sagerian and Truly;
`whom I will return to in a moment. We
`believe with this very strong team, we
`can produce some very interesting
`applications for the amateur radio
`community. We have already told you
`about the OSP modem software
`developed for demodulating JAS-1
`using an unmodified tnc and about the
`software that does FITs to do
`spectrum analysis (and weak signal
`work) by myse", Tom Clark, and
`Richard Allen W5SXD. Modems are
`still a hot topic with the group and
`recently I worked out a scheme to use
`some mathematical tricks that allows
`you to do modems of several different
`varieties of phase continuous
`schemes with minor changes in the
`carrier tracking algorithms and
`different clock recovery and bit
`decision algorithms. The modems
`that I have tested In the signal
`processing lab at work are for sse
`bandwidth (telephone also). The ones
`that are working well enough to shout
`about are
`
`BPSK 1200, 2400 bps
`OPSK 1200,2400,4800 bps
`GMSK 1200,2400 bps
`V.29 4800,9600 (not full duplex)
`V.32 9600 (not full duplex)
`
`All of these are done with a new
`adaptive equalization scheme that is
`blind and self training and can
`maintain equalization with channel
`changes that occur over several bit
`times (rapidly changing channels such
`as are guaranteed on HF are not
`treated YET). We should begin to
`look at Noncoherent schemes such as
`M·ary FSK and others with an "eye"
`towards doing some adaptive Ea. The
`reader is referred to a recent article in
`aEX by team member VE3JF, Barry
`McClarnon. Also upcoming in OEX
`are a series of articles by team
`members. The first two are by Tom
`Clark and myself and are a general
`introduction to the OSP project and a
`detailed look at DSP modems. The
`second (or third depending on the
`editor) will be about spectrum analysis
`and weak Signal work. The reader is
`referred to the papers in the 6-th
`annual Networking Conference by
`Tom, Barry, and myself. Also you
`
`might wish to read the recent Above
`50 Mhz column by Bill Tynan on the
`weak si9nal applications end of the
`spectrum analysis code. He outlines
`how the work that Tom, Richard Allen.
`and I have been doing could impact
`the weak signal arena.
`
`Tom Clark and I gave a talk at the
`AMRAO meeting on February 1. The
`AMRAD group has started a group
`DSP effort and devoted two talks in a
`row to DSP and half of their most
`recent newletter.
`
`Phil Kam has become active with the
`DSP software effort and is working
`with myself on a 400 bps BPSK
`modem for Phase III-C. Phil is also
`working on a program to run in the PC
`that allow the OSP board and the PC
`to become a very versatile command
`station. We are also beginning to
`work on 2400 BPSK autocorrelation
`routines that should greatly outperform
`the technique used for ranging to
`determine the orbital elements of
`Phase III-B. Phil also did a very clever
`spectrum folding trick in TMS3201 0
`software that allowed inverted
`sideband signal (Donald Duck) to be
`fed to the DSP board and out comes
`normal sounding audio.
`
`WBOMPO, AI Shjarback, recently
`loaned me a receiver that receives in
`the range 137·138 Mhz. The sweet
`sound of WEFAX-APT has been
`flowing into my DSP hardware. This
`demodulator is now working. The
`picture is displayed as false color on
`an EGA monitor. Both the infrared
`and visible pictures are shown side by
`side. This was demonstrated at the
`TAPR meeting in late February.
`Meteor (Soviet) WX satellites were
`were added on 2110/88.
`
`At the TAPR meeting, a great deal of
`interest in the OSP project was shown
`and a great deal of encouragement
`was given the members of the DSP
`team.
`
`On the hardware:
`
`Lyle WA7GXO is beginning the layout
`of the first DSP project to be produced
`by this group. It is basically the same
`box described to the audience at the
`T APR annual meeting. There are
`
`Page 6
`
`April 1988
`
`PSR
`
`Issue #31
`
`

`
`several design questions being
`addressed at this time, so the details
`are still in a state of flux. We believe
`that the project will be announcing a
`DSP kit sometime in the near future,
`
`Steve Sagerian, KAOYRE, an engi(cid:173)
`neer with Motorola, Inc. of Chicago
`has acquired some Motorola
`DSP56000 development system
`boards and is working up an AID DIA
`board for them. We have also gotten
`a complete set of software tools for
`this device at this point in the game
`including an early release of
`Motorola's ·C· compHer, Assembler
`and Linker for the DSP56000 chip. We
`will be developing code for this chip
`and the development will initially
`parallel the TMS320 development
`effort so that we might compare the
`two hardware types.
`
`Not to be outdone, AT&T has an(cid:173)
`nounced the DSP-32C. Another
`blistering DSP engine in the same
`class as the TMS320C25 and the
`DSP56000. Dave Truly NN2Z and I
`are writing a proposal to Dave's boss
`who is charge of marketing the DSP- '
`32C for AT&Tto help the DSP project
`with