Text content (OCR)
ISSUE 22 OCTOBER 1975
ws + Build a
GRAPHICS DISPLAY
* TELEVISION INTERFACE
« NOVAL ASSEMBLER
*« KLUGE HARP: Music ?
* LIFE Line: Game
Computers -
The World’s Greatest
Toy!
ie
CT-1024
TERMINAL
SYSTEM
When we designed the CT-1024 we
knew that there were many applica-
tions for an inexpensive TV display
terminal system. Even so, we have
been surprised at the many additional
uses that have been suggested by our
customer in the last four months since
we introduced this kit.
The basic kit, consisting of the charac-
ter generator, sync and timing circuits,
cursor and 1024 byte memory gives
you everything you need to put a six-
teen line message on the screen of any
TV monitor, or standard set with a
video input jack added to it. Input
information to the CT-1024 may be
any ASCII coded source having TTL
logic levels. Two pages of memory for
a total of up to one thousand and
twenty four characters may be stored
at a time. The CT-1024 automatically
switches from page one to page two
and back when you reach the bottom
of the screen. A manual page selector
switch is also provided. The main board
is 9% x 12 inches. It has space pro-
vided to allow up to four accessory
circuits to be plugged in. If you want a
display for advertising, a teaching aid,
or a communication system then our
basis kit and a suitable power supply
is all you will need.
CT-1 TERMINAL SYSTEM with
MEMORY KIT... $175.00 ppd
Power supply kit to provide + 5 Volts @
2.0 Amps and - 5 Volts, -12 Volts @ 100
Ma. required by the CT-1 basic display
system.
CT-P POWER SUPPLY KIT.
$15.50 ppd
A very nice convenience feature at a
very reasonable cost is our manual cur-
sor control plug-in circuit. The basic
kit allows you to erase a frame and to
bring the cursor to the upper left cor-
ner (home up). By adding this plug-in,
you can get Up, Down, Left, Right,
Erase to End of Line and Erase to End
of Frame functions. These may be
operated by pushbutton switches, or
uncommitted keyswitches on your key-
board. Although not essential to ter-
minal operation, these features can be
very helpful in some applications.
CT-M MANUAL CURSOR CONTROL,
wresesaeneer 11.50 ppd
If you plan to use your terminal with a
telephone line modem, or any other
system that requries a serial data out-
put; you will need our serial interface
(UART) plug-in circuit. This circuit
converts the ASCII code from a par-
allel to a serial form and adds “Start”
and “Stop” bits to each character. The
standard transmission rate for this
circuit is 110 Baud, but optional rates
of 150, 300, 600 and 1200 Baud may
be obtained by adding additional parts
to the board. The output of this cir-
cuit is an RS-232 type interface and
may be used to drive any type modem,
or coupler system using this standard
interface.
CT-S SERIAL INTERFACE (UART)
KIT... .. $39.95 ppd
If you are using the CT-1024 as an 10
{input - output) device on your own
computer system, you will probably
C1 Enclosed is $.
(0 or BankAmericard #
O ct-1024 kit
O cts Serial interface Kit
want to connect it to the computer
with a parallel interface system. A di-
rect parallel interface allows for much
faster data transmission and reception
and is basically a simpler device than a
serial interface system. Our parallel
interface circuit contains the necessary
tristate buffers to drive either a separ-
ate transmitt and receive bus system, or
a bidirectional data bus system. TTL
logic levels are standard on this inter-
face. Switch selection of either full, or
half duplex operation is provided. The
terminal may write directly to the
screen, or the computer may “echo”
the message and write to the screen.
CT-L PARALLEL INTERFACE
KIT... see 22.95 ppd
We would be happy to send you a com-
plete data package describing the CT-
1024 and a achematic. If you want this
additional information, circle our num-
ber shown below on your reader infor-
mation service card. The CT-1024 kit
has complete assembly instructions
with parts location diagrams and step-
by-step wiring instructions. If you
would like to check the instruction
manual before you purchase the kit,
please return the coupon with $1.00
and we will rush you the manual and
the additional data mentioned above.
MAIL THIS COUPON TODAY
or Master Charge #
Card Expiration Date
1 CT-M Cursor Control Kit
C1 ct-t Parallel Interface Kit
ADDRESS.
CITY.
STATE ZIP.
|
H
I
!
I
I
| Name
|
|
|
|
|
CO $1.00 Enclosed send manual and data package
Southwest Technical Products Corp., Box 32040, San Antonio, Texas 78284
In the Queue BUTE #2
OCTOBER 1975
Foreground
ADD A KLUGE HARP TO YOUR COMPUTER ...... 14
Fun and Games — Carl Helmers
LIFE Line 2 00... cece cece cece teeter eee 34
Software — Carl Helmers
A QUICK TEST OF KEYBOARDS ..........-..065 43
Hardware — Walters
ADD A STACK TO YOUR 8008 ........ 0. serene 52
Hardware — Chamberlin
A NOVAL ASSEMBLER FOR THE 8008 ........... 64 Here
Software — Peter Helmers . ve in
ASYNCHRONITISs 0s isc ie ve seis ee oe ws oe ee 68 ‘Maer
p. 34
Hardware — Bancroft
BUILD A GRAPHICS INTERFACE..........+00005 70
Hardware — Hogenson
Background
TELEVISION INTERFACE ...... 0. sce e cece eee 20
Hardware — Lancaster
MODULAR CONSTRUCTION ......... cece evens 46
Hardware — Walters
BUCK ROGERS AND THE HOME COMPUTER......58
Speculation — Gardner
Nucleus
Speaking of Computers ....... 0.000. ee eee eel 5
BYTE magazine is published
monthly by Green Publishing, LOHERS: scans Sees i ere corey aa yum HeTaS 4 8
Ine., Peterborough, New
Hampshire 03458.
Sibsctigtion slasere.b18 for Clubs and Newsletters ..........600.0s sees AZ:
y ldwide. Ti F
scutes $29. ‘Three: years $80. A Word From the Publisher 61
Second class postage .
application pending at Book Reviews. .......- ee eeee eee eee eee
Peterborough, New Hampshire
03458 and at additional Bits’atid Pisces)... 0. scene cisiersccacs uninioe anew ar
mailing offices. Phone:
603-924-3873. Entire contents 4 i
@eponaht 1976 by Gren Reader's Service ... 0... .. 00 cee cee ee cerns
Publishing, Inc., Peterborough, F .
NH 03458. Address editorial The BYTE Questionnaire
correspondence to Editor,
BYTE, Box 378, Belmont MA
02178.
The Sphere System...
Tomorrow’s Computer Today!!!
IT WILL Dra ve cheng long ater our kit aso poet tend
IT WILL to expand to the largest set of peripherals and
le in today’s lowcost computer market.
IT WON'T BREAK YOUR BANK !!!
MORE —
Are you? send info to
[an educator
O@ hobbyist
tan OEM... QTY/YR —__
O/ooking for
product to
sell.
STAMP
HERE
What is your application
SPHERE CORP.
791 South 500 West
Bountiful, Utah 84010
19H 1908 PIOF
cut here cut here
Clip out and mail coupon or call (801) 292-8466
Before After
KITS Sept. 30 Sept. 30
SYS1 4K Byle computer with TV terminal, $ 650 $ 860
keyboard, and ROM monitor
SYS2 9 4K Byle computer with TV terminal. 750 999
keyboard. ROM monitor. and audio
cassetle interface.
SYS3 20K — Byle computer with TV terminal, 1345 1765
keyboard, ROM monitor, audio
cassette interface, and extended
basic programming language.
NEW FEATURES
PEN Light pen kit for CRT board 80 100
BAS 1K micro-basic prom set (a modified
basic language that makes your computer
talk basic without loading from pLte) 90 140
RAC2 Altractive plastic chassis with CRT display
and acoustic coupler. 270 300
SPH
791 South 500 West, Bountiful, Ulah 84010 (801) 292-8466
Carl Helmers
Editorial :
A whole line of home
computer experimentation
can be started by the person
who designs — and writes up
as an article for BYTE — a
simple and practical circuit to
gencrate speech output from
phoneme codes in a program.
The output problem in digital
speech representation
techniques is well within the
range of a microcomputer
system and _ inexpensive
digital to analog conversion
methods (such as the
Motorola 6-bit el cheapo
MC1406 DACs).
Consider, for instance, a
brute force technique in
which the voice info is
encoded as 3-bit (eight level)
quantities which are sent out
ata rate of once every
millisecond for conversion.
(Rate: 3000 bits per second.)
A typical voice “phoneme
string’ in the micro’s
memory would specify data
for maybe 100 milliseconds.
This would require a total of
100 data values or 300 bits,
stored in a packed bit string
format in 38 bytes plus one
overhead byte for string
length. If one considers a
reasonable vocabulary of
speech elements, say 100
basic sounds, the data
requirement is thus 3900
bytes — well within the
memory budgets of many
BYTE readers’ systems. The
thruput required to output
the voice bytes is also well
within a micro’s capacity —
1000 microseconds is a lot of
time to fool around with. It’s
even long enough for a
tortoise of a computer like
the 8008 to do enough bit
diddling to prepare a
command code for a 3-bit
DAC port.
Now, what’s all the
purpose to the home voder?
Well, if you can’t think of a
use for it, ’ll supply a couple
of suggestions. Wouldn’t it be
neat to put up a home
security system using your
micro in which the burglar
gets scared out of his wits by
a threat issued in a
computereeze voice while
lights are flashing and other
ominous things are
happening? Or in the area of
gamesmanship — when the
game program gets erroneous
input, have your program
issue the text of the message
“foull” Or, getting less
exotic, simply make yourself
a calculator which will
literally read out the answers
to your problems as well as
show them on an output
screen. Getting exotic again,
suppose you make a Star
Trek oriented space war
game. You really should have
Functional Specifications
“The Home Brew Voder”
the “ship's computer’ give
status reports aurally to make
the game more exciting —
perhaps coupling in a few
bells and whistles (literally) in
the form of special purpose
synthesizers of photon
torpedo, phaser, transporter
and other sound effects
Trekkies know and love.
Then there is the ultimate
application — making a higher
grade voice synthesizer which
can sing, literally, so your
machine can play vocal
music.
How can such audio
peripherals be brought into
existence? One way is
through the encouragement f
can give by publishing articles
on various approaches — the
design articles of readers
which make BYTE an
essential publication for the
home computer
experimenter. A second way
is for the various
entrepreneurial readers in the
audience to take a cue from
this little essay and get
working on packaged
products in the audio output
line which the majority of
readers will be able to put
together and program.
The essential elements of
the inexpensive voice
synthesizer are the familiar
hardware and software
combination: The hardware
part is the n-bit (3 will do)
DAC output port (and latch)
which converts internal codes
into one of 8 voltage levels.
The DAC in turn will drive an
active low pass filter (op
amp) and a power amplifier
for the speaker. If you get
fancy, a second DAC could
be used as a gain control
output (with an appropriate
8-bit latch for storage) — and
the 5 unused bits of the
typical speech value word
could be used to control 5
additional bells and whistles.
The software of the
application is in two
segments. There is the
machine-independent data
table which specifies the basic
phoneme information — and
the rules for combination
into understandable words.
Then there is the machine-
dependent programmed
“talker’? routine which
accepts an “‘n-byte” character
string with phoneme codes
and synthesizes the phonemes
one by one from the data
table by outputting the
selected series of 3bit codes
in real time. In a design
article, “talker” would be
specified functionally in a
high level language along with
a global flow chart, and the
phoneme data information
would be specified as a table
in hex and/or octal codes. So
here you have the idea — let's
see what BYTE readers can
come up with in the way of
articles on the subject.
(Oh yes, one parenthetical
note — the proof of the
putting is in the speaking.
Prospective voder designers
should send along a tape
recording of their design in
operation — preferably
speaking the first two lines of
Lewis Carroll’s poem
“Jabberwocky,” the universal
test string:)
A Need in Search of a Product
Have you ever run a
business, newsletter or club
which requires periodic
mailing of information? Have
you ever tried to maintain a
mailing list? The purpose of
this short essay is to identify
an opportunity which exists
for the entrepreneurial
persons in BYTE’s readership
— the opportunity to create
and market a specialized
mailing information system
using contemporary
technology in the form of
microcomputers and
inexpensive peripherals. Here
is what you have to compete
against in the general market
— the free market of all the
possible solutions to the
problems of mailing labels.
Multipart Carbon Forms
My old standby in the
mailing list area is the
multipart carbon form which
is manufactured by Dennison
among others. These forms
hold 33 names per sheet and
come with four parts to
reduce typing and retyping of
names. They come with water
base glue backing and are
perforated — but the main
problem is typing. The 33
name per sheet figure only
holds if you don’t make any
typing errors! The degree of
automation of this system is
usually zero, unless you have
a power typewriter and are
using the carbon forms to
save output. Cost is also low
— paid for in typing time of
course — at about $2 to $3
for a package of ten. No
competition — labor
intensive.
Spirit Duplicator Methods
The next step up in the
world of office automation is
a system involving spirit
6
duplication stencils — the
blue ink smeary reproduction
that was in vogue for high
schools and grade schools
before Xerox and its
imitators became so
widespread. A Sears version
of this system, per their
office products catalog, costs
from about $70 upwards.
Similar units run up to the
range of about $200 to $300
and address master blanks for
the “Elliott” system cost
about a dime each. This type
of system has a higher degree
of automation (retyping of
names is much less frequent)
— but still involves a costly
“use-only-once” part, the
stencil.
Addressograph-Multigraph
The next step up in cost —
a much more permanent
system for large usage — is
the Addressograph-
Multigraph type of system
which uses metal plates
prepared in advance. These
systems have a much longer
lasting and more expensive
label master blank — and
minimal systems can be had
in the $800 range and up.
The system involved is
essentially the same as the
duplicator style systems —
but more costly, due to
fancier equipment, some
automation of envelope
handling and longer lasting
media. There is still no
automation of the typing and
related information handling.
Here is where the new
computer systems will begin
to effectively compete.
The Hypothetical Small
Computer Addressing System
A “trivial” application of
the microprocessor
technology which can be
assembled by any of BYTE’s
more experienced readers is a
hardware/software system
consisting of the following
items:
1. Microprocessor (8-bit)
with 1024 bytes RAM, 2048
to 4096 bytes ROM program,
three serial ports, one 8-bit
parallel output and one 8-bit
parallel input.
2. Adding machine tape
ASCII printer — accepts
parallel output and prints it
as characters on rolls of tape.
3. ASCII input keyboard
— parallel input of text and
commands.
4. Triple audio cassette
interface capable of two
input operations and one
output operation
simultaneously.
5. Edit/Merge program —
software for editing of
“current label” records held
in RAM, using the keyboard
for commands and text input,
using the printer for output.
Allow optional input from
one tape port, output to a
second tape port, with
automatic sequence checking
to maintain a sorted sequence
on the files.
6. Update/Merge program
~— software for “batch”
changes to the file, in which a
previously prepared (and
sequenced) update tape is
merged with an old tape to
Create a new output tape.
7. Print program —
software to print the file —
either as an unconditional
dump of all labels, or a
selective dump such as “every
nth name” or “names with
zips 07932 to 07860.”
8. A Percy Wing Machine.
This is a hand-operated gizmo
which costs about $100 and
is used to automatically apply
labels from the roll of
printout — slicing each label
off the roll and automatically
applying it with glue to the
mailing piece.
Put together a packaged
product based on these ideas,
and it could most likely be
sold in the $1500 to $2000
range, supplying a nice profit
margin and a product which
competes effectively with the
Addressograph-Multigraph
“‘systems,’’ yet provides
automatic features and a
much more compact storage
method (cassette tape) for
lists of moderate size.
[RIGS ELECTRONICS)
DISCOUNTS: 10% OFF ORDERS OVER $25.00; 20% OFF ORDERS OVER $250.00.
SPECIAL
1-8008
8-2102
$50
ANOTHER POWER SUPPLY...
PS 25-1 0 to 25v 1a lab type power supply with adjustable current
limiting; remote sensing & remote programming for voltage & current.
Insturctions included. All parts except chassis, meter(s), p.c, board.
Kit of parts with schematics. $14.95
P.C. boards available, No. 007 $3.00 ea.
2K RAM BOARD KIT. ALL
PARTS INCL. SOCKETS
$84.50
ICs
8008 MICROCOMP. CHIP$30.95
2102 1K STATIC RAM 3.00
5203 256x8 PROM 15.00
5204 512x8 PROM 25.00
INFO ON ABOVE CHIPS IF
ASKED FOR.
ORDERS OF $50 OR
MORE GET FREE BYTE
SUBSCRIPTION IF ASKED
FOR (CONTINENTAL U.S.
ONLY).
008A MICROCOMPUTER KIT
8008 CPU, 1024 x 8 memory; memory is expand-
able. Kit includes manual with schematic, program-
ming instructions and suggestions; all |Cs and parts
supplied except cabinet, fuses & hardware. Includes
p. c. boards. $375.00
MANUAL ONLY, $25.00
(no discount on manual)
008A-K ASCII keyboard input kit. $135.00
O08A-C Audio cassette adapter kit. $100.00
Details on computer, peripheral kits in our flyer.
RGS ELECTRONICS
3650 Charles St., Suite K # Santa Clara, CA 95050 = (408) 247-0158
We sell many ICs and components not listed in this ad. Send a stamp for our free flyer. TERMS OF SALE:
All orders prepaid; we pay postage. $1.00 handling charge on orders under $10.00. California residents please
include sales tax. Please include name, address and zip code on all orders and flyer requests. Prices subject to
change without notice.
_I
7
LETTERS
“IT'S ALL IN THE
EPHEMERIS...”
Dear Mr. Helmers,
| have sent in a
subscription order for BYTE.
IL seems to be what I’m
looking for. Already | have
been looking at surplus ads
for CPUs with an eye to
making my own digital
computer though “simple
minded” it would probably
have to be.
Among other things, | am
a ham and a_ consulting
engineer for EBASCO. What |
have in mind is to have
available a programmable
computer to solve filter
design problems (and
antenna/feedline impedance
problems parametrically).
However — as a first
question: Do you think it
may be in the realm of a
“non-computer expert” to
make a home built computer
solve weather satellite orbits
and read out azimuth and
elevation information for
tracking purposes? | can feed
in very accurate time
information. ! don’t know
yet how to get from “here”
to “there.”
W. J. Byron
New York NY
You have an interesting
question... For the first
portion, how to solve the
analytical equations of
engineering, there is a range
of solutions depending upon
your purposes, budget and
8
other factors. The simplest
approach, with the least
amount of money and the
least amount of computer
technology ‘‘learning”’
experience is to simply go out
and buy one of the new
HP-25 calculators. You'll find
it quite capable of solutions
to a large number of
numerical analysis problems
for engineering — although
it’s adequate for extensive
linear algebra and matrix data
calculation. My associate
Chris Bancroft has been using
the slightly more powerful
HP-65 for more than a year
now to arrive at exact
analytical (and very
predictive) solutions to
engineering problems in
applied electronics — and the
HP-25 should be able to
perform similarly on many
problems, But a pre-packaged
calculator may be “‘no fun” —
and certainly is not useful for
anything other than
calculation. Further, the
programming of a_ really
complicated satellite position
algorithm (eg. adding in
second and third order
perturbations of the earth’s
gravitational potential) may
be beyond the range of a
simple hand calculator
without multiple mental
overlays.
The problem of simple
celestial mechanics
computation is well within
the range of a small home
microcomputer system which
contains perhaps 4k bytes (or
equivalent) and the facility to
do overlay programming
using at least one audio
cassette drive. In order to do
this kind of calculation,
however, | am making the
following assumption: you
already have the analytical
solution to the problem in
the form of an algorithm
specification. (In case you
don't, maybe a reader with a
bent for applied celestial
mechanics, coordinate
transformation problems and
numerical integration might
be so kind as to supply a
solution to this problem
either to you privately or for
publication.)
In order to program the
typical three-space navigation
problem, assuming well
determined static orbital
elements, the following items
are required:
7. A computer with
software for SIN, COS, TAN
and ARCTAN trig functions
(vou can get this set from an
interface calculator chip).
2, Vector-matrix algebra
subroutines for three-space.
Since you are talking about
static orbits (no active
control altering parameters),
you don't need much in the
nature of the more advanced
error correcting algorithms.
3. A floating point
arithmetic package to go
along with the computational
requirements of #1 and #2.
This should probably include
data conversions to integer
and character string forms for
convenience of programming.
4, An orbital model which
uses a given set of ephemeris
information at some time “t"’
and calculates new satellite
position at time “t + At”
later. (The “now” position.)
5. An output model which
uses various coordinate
transformations to turn the
“‘now’’ position in
“geocentric coordinates” into
an apparent position on the
unit sphere for a known
latitude, longitude and time.
This is the right ascension and
declination of an optical
telescope or equivalent terms
for your radio antenna.
The place to look for
detailed information is an
advanced undergraduate or
graduate level book on
classical mechanics,
spacecraft navigation or
astronomy if you want to
find info on the calculations.
One possible source of
information, mentioned to
me by Bob Baker of
Littleton, Mass., is a radio
amateur organization called
the Amateur Satellite
Corporation — AMSAT, Box
27, Washington DC 20044,
Bob tells me that their
newsletter frequently
mentions computational
algorithms for the OSCAR
amateur radio satellite —
algorithms which could be
adapted to any sutellite
orbital elements. Another
possible source of
information is the HP-65
Users Club run by Richard
Nelson — in scanning through
his back issues recently | saw
several listings of names of
programs sounding
suspiciously like what you
want, e.g., “Orbital Element
Determination” and the like.
(HP-65 Users Club, 2541
West Camden Pl., Santa Ana
CA 92704.) While by no
means a complete design, |
hope this information proves
useful to you.
.»- CARL
EVOLUTION
NOT EROSION
Dear Mr. Helmers,
1! am a subscriber to ECS
magazine and was a little
uncasy to receive the
announcement of its
transformation into “BYTE”.
| hope this is an evolutionary
development, and not an
erosion into the murkiness of
total commercialism. | have
very much enjoyed and
benefited from ECS.
Thanks for the help, and
good luck in your new
venture.
Duane L. Gustavus
Denton TX
1 think 1 can sympathize
with your uneasiness about
ECS Magazine's
transformation into BYTE.
Yes, it is indeed an
evolutionary transition, in
several respects. First, if you
look in the first issue you will
find a much larger and more
varied editorial content than |
was ever able to achieve in
the course of preparing my
self-published 24page photo
offset magazine. Second, by
providing a place for
commerce — a free market —
it serves as a unifying element
in the whole sphere of the
computer hobby endeavor.
Throughout history, it is the
market place which has
sustained the progress in
ideas, and technology which
has brought the human race
from crude wheels to $20
bus-oriented microcomputers.
Third, there is the element of
professionalism in format and
execution. BYTE is being put
together by a fine
organization of craftspeople
who take pride in the work
which results. Sure, the
magazine is commercial and
has advertisements — but
then, don't you work for
yourself, some employer, or
other agency? This pride in
the quality and value of work
extends throughout the
publishing operation | have
joined — and it will be the
element which makes the
reputation for the magazine
as a source of ideas and fun,
-.. CARL
“BILL ME”
Dear Sir:
Yes, I'll byte, but | think
I've been taken, | was a
subscriber of ECS and was
told that subscription was
okay for the charter BYTE
publication; | guess that
promise ... and ten bucks
can also buy a year’s
subscription. I've been bit
(for more than the ten bucks)
by others feeding off the
crumbs of micro machine
data, so can't begrudge you
yours.
1 am eternally hopeful for
BYTE, but not optimistic. |
have been a ham for a long
time (23 years), and Mr.
Green is not my favorite; also
I didn’t think Mr. Helmer’s
ECS info was very good. |
agree, however, that there
“sure is a need for a good
magazine,” so hope BYTE
can serve that need. Let’s
hope it’s better than 73.
Please bill me for the lousy
sawbuck.
C. Southard WAGIOT
Cedar Rapids IA
Jam printing your letter in
BYTE for a reason, a matter
of principle if you will: | am
personally responsible for the
editorial content of BYTE
magazine, and have my
reputation on the line as a
result. | believe that after you
have read your first issue of
BYTE you will find it well
worth a $10 which you had
no reason to send in at this
time other than (perhaps) a
negative attitude and a
promotional circular which
came your way as a result of
one of our mailing lists. 1
have a number of items of
unfinished business in
connection with the M. P.
Publishing Co. operation |
was running in my spare time
until BYTE started — one of
the first such items was a
condition of the arrangement
| made with Green Publishing
that all ECS subscribers
Should be picked up by
BYTE ona two issues for one
basis, Thus since your ECS
subscription of $21 was
fulfilled by mailings of 10
issues, the remaining two ECS
issues become four BYTE
issues. Adding to that your
inadvertent resubscription
gives a total of 16 BYTES...
OK, you say, “he promises
16 BYTES — will 1 get
them?" You can best evaluate
that after you've gotten
BYTE #16. The aim of BYTE
magazine is — as is the aim of
the large corporation for
which you work — to turna
reasonable profit in the long
run and provide enduring and
satisfying work for those
connected with the
enterprise. It can only do so
by providing good service to
its customers — the readers
and advertisers who patronize
the magazine. BYTE is very
much a market phenomenon,
and cannot exist unless it
maintains a readership. of
intelligent and active persons
such as yourself. If BYTE
were to consistently turn you
off — as well as others on a
large scale — we'd be heading
into bankruptcy faster thana
Penn Central express train
should be running.
/ invite you or any reader
who thinks he or she is
getting a bad deal on the
magazine to write me
personally at any time. 1
won't promise to publish all
such letiers — yours, Mr.
Southard, is published by
distinction of being the first
— but I will endeavor to
answer each one personally.
As to technical content,
you can peruse the first issue,
then make a judgment. | will
endeavor to produce the best
possible magazine by
selecting the best possible
articles. | am not about to
ignore the _ biggest
multiprocessor system of
them all — the human race
... all the people in the
readership of BYTE who will
be coming up with ideas for
articles and submitting them
will set the level of much of
the material available for
publication, If you don’t like
my personal work, kindly
give me the courtesy of
stating why ~ 1 know that
the ECS Series articles 1
previously put out had many
flaws, I have heard some good
and some bad evaluations of
the ECS Series and ECS
Magazine items.
1 will not however disown
anything | have done — and
that series of self-published
articles is my product with all
its flaws and imperfections. If
you wish to exercise the
moneyback guarantee, let me
know... and send back your
copies to fulfill your part of
the exchange.
.. CARL
10
WHAT SINGLE ELECTRONIC
MACHINE CAN BE USED TO
PERFORM/CONTROL ALL
THE FOLLOWING TYPES
OF SERVICES?
Send morse code
Control repeater stations
Operate as a calculator
Receive/send/buffer data
between a wide variety
of communication devices
Monitor instruments
Control machines
Sort/compile data
Test other devices
ee the SCELBI-8B MINI-comPuTER CAN !
SCELBI COMPUTER CONSULTING, INC.- The company that pioneered in producing the small computer for the
individual user with the popular SCELBI—8H, now brings you the new SCELBI—8B with increased capability!
Like the former SCELBI—8H, the SCELBI—8B is built around the amazing ‘8 0 0 8 “CPU-on-a-Chip” which has been
revolutionizing the electronics world.
However, the NEW SCELBI—8B offers extended memory capability at reduced cost! It is directly expandable to
16,384 words of RAM/ROM/PROM memory. This increased memory capability now means the user has the potential in
a small and compact computer to support compiler type languages, manipulate sizable data bases for business and
scientific applications, and support a wide variety of programs including those that take advantage of external mass
memory storage devices.
The NEW SCELBI-—8B still retains the outstanding features of its predecessor. Decoding logic for 8 Output and 6
Input Ports is built into the basic computer. Plug-in capability for 1/O devices is provided on the chassis. A unique,
simple to operate console that utilizes just 11 switches on the front panel makes the SCELBI—8B a pleasure to use.
The NEW SCELBI—8B is backed by a line of low cost SCELBI interfaces which currently include: an interface that
turns an oscilloscope into an alphanumeric display system, low cost keyboard and TTY interfaces, and an interface that
turns a low cost audio tape cassette into a “Mag—Tape”’ storage and retrieval unit.
Last, but certainly not least, SCELBI has a wide selection of software ready to run on the NEW SCELBI—8B
including: Editors, Assemblers, calculating programs, I/O and general utility routines. Additionally, SCELBI produces
publications that can show you how to develop your own custom tailored programs.
The NEW SCELBI—8B is available NOW. (We have been delivering since June!) It is available in three forms. Ultra-low
cost “Unpopulated” card sets with chassis kits starting at $259.00*. Complete parts kits for a 1,024 word
mini-computer as low as $499.00". An assembled and tested 4,096 word computer is just $849.00*. Interfaces,
accessories, and software sold separately.
(*Domestic prices.) (Prices, specifications and availability subject to change without notice)
Literature available for S.A.S.E.
SCELBE COMPUTER
CONSULTING INC
1322 REAR BOSTON POST ROAD
MILFORD, CONNECTICUT 06460
MACHINE LANGUAGE
PROGRAMMING
i i
for the F=Lete}=4 (AND SIMILAR MICROCOMPUTERS)
Written to provide you with the detailed knowledge you need to know in order to successfully
develop your own MACHINE LANGUAGE PROGRAMS! This information packed publication
discusses and provides numerous examples of algorithms and routines that can be immediately
applied to practical problems. Coverage includes:
DETAILED PRESENTATION OF THE “8008” INSTRUCTION SET MATHEMATICAL OPERATIONS
FLOW CHARTING — MAPPING MULTIPLE-PRECISION ARITHMETIC
EDITING AND ASSEMBLING DEBUGGING TIPS FLOATING-POINT PACKAGE
FUNDAMENTAL PROGRAMMING TECHNIQUES MAXIMIZING MEMORY UTILIZATION
LOOPS, COUNTERS, POINTERS, MASKS 1/0 PROGRAMMING REAL-TIME PROGRAMMING
ORGANIZING TABLES SEARCH AND SORT ROUTINES PROGRAMMING FOR “PROMS”
CREATIVE PROGRAMMING CONCEPTS
Virtually all techniques and routines illustrated also applicable to ‘8080’ and similar types of
micro/minicomputers, with appropriate machine code substitution. Orders now being accepted
for immediate delivery at the LOW price of just $19.95.* Add $3.00 if PRIORITY mailing
service desired. (*Domestic prices.) Pricing, specifications, and availability subject to change
Order direct from — without notice.
SCELBE COMPUTER 5.0 REAR BOSTON POST ROAD
CONSULTING ENG. witForD CONNECTICUT 06460
—-Ged. | enclose $19.95. Send me a postpaid copy of:
MACHINE LANGUAGE PROGRAMMING for the ‘8008’ (and similar microcomputers)
Please send my copy by Priority Mail. | enclose $3.00 extra.
Charge it to my Mastercharge Card #____. 9
Bank #__________.____Exp. Date.
Card Holders Signature.
NAME:
ADDRESS:
Clubs
Newsletters
LA Update
Thanks to the efforts of
Derek McColl, I’ve received a
further update on the Los
Angeles club activities ... a
very active bunch of people.
Derek sent me a copy of the
Southern California
Computer Society Interface
— Volume 1, #0 (do | detect
an algolmaniac at work?) for
August. The motto in the
heading line reads “an
announcement for computer
hobbyists designed to
connect people and ideas
...'’ — and it lives up to that
billing in its six pages of
typewritten copy.
The person who is
handling the administrative
details of SCCS for the time
being is Hal Lashlee, who can
be reached by phone at
1-213-682-3108, or by mail
at PO Box 987, South
Pasadena CA.
The following topics and
concerns were drawn from
the suggestion boxes of a
brief survey form handed out
to members at the first
meeting of SCCS (and printed
in Volume 1 #0 of /nterface):
— Members would like to
see a computerized
clearinghouse for computer
hobby information.
— One purpose of the
organization should be
mutual assistance with
specific problems.
— Hardware procurement
by group purchases.
— Standardization.
— The club might act as a
brain pool for small business
needs.
— The usual club type
activities of social meetings,
lectures, seminars,
workshops, public service,
etc.
12
creative corpatirg
Creative Computing is the
name of a magazine which is
edited and published by
David H. Ahi. The motto of
the operation is “a non-profit
magazine of educational and
recreational computing.”
Creative Computing is
published bi-monthly, printed
in a saddle-stitched 8% by 11
format (similar to BYTE)
with 60 pages in a typical
issue. The editorial content is
heavily oriented toward
information useful in an
educational context. David
Ahl was formerly the
Educational Marketing
Manager at Digital Equipment
Corp., where he was
responsible for the creation
of DEC’s EDU publication.
To quote from his editorial in
the March-April 1975 issue of
Creative Computing:
“Over the years EDU
flourished and grew into a
48-plus page magazine.
However there were certain
aspects of educational
computing which EDU could
=
INTERFACE
not satisfactorily address. In
particular, school users, both
college and elementary/
secondary, need more
classroom activities, exercises,
problems and ideas than are
available in textbooks and
other magazines. Also there
ought to be a discussion of
the social aspect of the
computer, its effect on jobs,
medical care, privacy and the
like. Furthermore, what
about the user of non-DEC
computers? Clearly to be
responsive to these needs
another vehicle was needed.
Thus Creative Computing was
born...”
The various issues | have
seen to date include
numerous puzzles, BASIC
games, and articles on
computers, computer
education and computer
careers. To order a
subscription, send $8 for one
year or $21 for three years to
Creative Computing, Box
789-M, Morristown NJ
07960. ++ CARL
Notes from the Garden State
via the Goethels Bridge
ACSN]’s second meeting
was held at the Union County
Technical Institute on July
18, 1975. The meeting was
presided over by Bohdan in
the absence of Sol Libes.
Thirty hobbyists showed up
including 8 new members. In
deference to Stephen Gray,
founder of the original ACS,
we are considering changing
our name to the New Jersey
Amateur Computer Group
(NJACG).
Roger Amidon gave a
presentation on the UART.
Marty Nichols held a
discussion on the differences
between 8008 and 8080
microprocessors. Andy Vics
talked about his experiences
with the construction and
operation of TVT 1 and TVT
2. Wayne Ahlers showed us
his octal keyboard built
around PEs 12-74 low cost
computer terminal. Literature
and other information was
disseminated before and after
the formal meeting.
Later that night, a small
group visited Roger Amidon’s
site. In addition to his 16k
Altair, TTY and magnetic and
Paper tape peripherals, Roger
K2SMN also has a home brew
RTTY controller which is
affectionately named Spider.
(If 1 can ever get a picture of
it, you'll understand why.)
The third meeting of the
NJACG will be held in
September, not August, on
the third Friday, September
19, 1975, at the Middlesex
County College. For more
information contact George
Fischer, 1-212-351-1751.
Oklahoma City Club
In Oklahoma City OK, Bill
Cowden reports that he is
organizing a computer
enthusiasts’ club. Contact Bill
at his home address, 2412
SW 45th, Oklahoma City OK
73119,
STOP fussin’ and cussin’ at soldering,
. ma
P: a rticy | igh-P J heat sinks, short circuits and ruined compo-
nents. That’s a DRAG... when circuit build-
ts ing should be FUN and components should
be REUSABLE. Now you can plug in, power
wy, u L h up and test your experimental circuits FAST,
iv SAFELY and CONVENIENTLY (and reuse
CG e those components) with all the assurance
and satisfaction of A P HIGH PERFOR-
MANCE in every circuit-building device.
TRY THEM AND SEE!
Super-Versatile™ TERMINAL and DISTRIBUTION STRIPS to build your own breadboards
SQUARE HOLES for ROUND LEADS? A/GHT! For easy com-
ponent plug in, no soldering, better gripping and solid electri-
calcontact. These beauties have matrices of universal plug in
terminals on .10" centers for accepting all DIP’s and discrete
components with leads to .032" dia.
Model 264L (shown) holds up to nine
14-pin DIP’s. Interconnect with any
solid wire up to No. 20 A.W.G, NEW,
integral, non-shorting, instant-mount
backing permits quick build-up of
custom breadboards using any mix of
A P terminal and distribution strips.
Superior, non-corrosive, nickel-silver
terminals. Other models available.
Model 212R
Distribution Strip...$2.50
Contains 2 continuous buses of 12 connected
4-tie-point terminals. Size: 6.5” by 35”.
Model 264L Terminal Strip...$
Contains 128 5-tie point terminals. Size: 65" by 1.36".
AC:E 200-K ALL-CIRCUIT EVALUATOR
MORE SQUARE HOLES...728...count ‘em.
Same high-performance features as abovel ~
This handy breadboard kit offers excellent cir-
cuit-building versatility. Holds up to eight 16- SAVES
pin DIP’s. Universal matrix of solderless, plug- OWKITS
in tie points includes 136 separate 5-tie-point
terminals and 2 distribution buses, each consist- Other models
ing of 6 connected 4-tie-point terminals... typi- _ ae autocrine
cally for voltage and ground. Includes two 5- ACE Model
way binding posts, 4 rubber feet. Aluminum 200-K only $18.95 =—
base, serving as ground, is gold-anodized for goard size: 4-9/16" x 5-9/16"
] surface protection. Assembles in 12 min. Complete instructions included
for FAR-OUT DIP TESTING . .. Super-Grip™ IC TEST CLIPS
No more shorting across DIP leads... . just clip on an IC TEST CLIP to
bring DIP leads out for SAFE attachment of probes and other leads.
Idea! for signal inputs, tracing, trouble- terminals in card racks (photo shows
shooting, etc. Patented, precision, this unique feature). Engineered mech-
“contact comb’ design guarantees no _anical clamping plus gold-plated phos-
shorting between DIP leads. Probes phor bronze terminals provide superior
can hang “‘no-hands” free on Test Clip electrical contact. Unequaled as a safe
DIP-pulling tool, too!
Models to fit all DIP’s:
ORDER BY MODEL NUMBER
TC-16 fits 16-pin DIP’s etc. We honor 4 Add proper fees from ths chart—aer_SHIPPING/HANDLINGIC.O.,
laster Charge ‘(Add sales tax on OH and CA order Upto $10.00 $1.00 | $ .70
1C38.. 47.38 TC-20. ..11.55 BankAmericard —(¢.,8. Painesville on company ?.0.'s.) S200): 8%500 150 | Bo
TC-22... 11.55 charge orders. Dealer inquiries invited. 50011010000 280} 1.00
78 TC-24, .. 13.85
TC-16LSI8.95 TC-28...15.25
TC-18...10.00 TC-36...19.95
TC-40, ..21.00
All products guaranteed to meet or exceed published specifications
AP PRODUCTS INCORPORATED
Box 110-G e 72 Corwin Drive e Painesville, OH 44077
or phone 216/354-2101 . or Twx 810-425-2250
by
Carl Helmers
Editor, BYTE
Add a Kluge Harp
to Your Computer
One of the most
interesting computer
applications is that of
electronic music. This is the
use of software/hardware
systems to produce sequences
of notes heard in a loud
speaker or recorded on
magnetic tape. The idea of
generating music — if well
done ~— is of necessity
complex. If | want to put my
favorite Mozart piano sonata
into an electronic form, I'd
have to record a very large
number of bits in order to
completely specify the piece
with all the artistic effects of
expression, dynamics, etc...
The magnitude of the
problem can be intimidating.
But, never let a hard problem
get in the way of fun!
Fig, 1. The Kluge Harp Circuit .., minus computer.
14
Ald OF COMPUTER
al
CLR SETLOC
AI3 OF COMPUTER |
a
CLR RESETLOC
7437 jo
fresno
Simplify the music problem
to one channel of melody,
and you can use a virtually
bare CPU with a very simple
peripheral to play music.*
The combination of the CPU
with this simple peripheral is
what I call the “Kluge Harp”
—a quick and dirty electronic
music kluge.
| invented this electronic
music kluge to answer a
specific problem: | had just
gotten a new Motorola 6800
system’s CPU, memory and
control panel up and running.
(*ALTAIR owners: Write an
8080 version of this program and
your machine can do more than
blink its lights.)
+5V
[en
SPEAKER
i i
i Hl
Law
OPTIONAL
RESISTOR
(1009)
The next problem (since |
wasn’t using the Motorola
ROM software) was to make
a test program which could
be loaded by hand. By
combining a little
imagination, my predilections
for computer music systems
and an evening getting the
whole mess straightened out,
the Kluge Harp resulted.
While the program and
schematic are specific to the
system | was using, the fdea
can be applied to your own
system just as well.
The Kluge Harp Hardware
The hardware of a Kluge
Harp is simplicity at its
essence. The peripheral is
driven off two “un-used”
high order address lines (I
used Al4 and A13), and
consists of a set-reset flip
flop. A program running in
the computer alternately will
set and reset the flip flop by
referencing one or the other
of two addresses. These
addresses are chosen so that
the address lines in question
will change state, actuating
the set or reset side of the flip
flop. A “note” at some pitch
consists of a delay loop in the
program followed by
instructions to change the
state of the flip flop. Since
the same count is used for the
two halves of a complete
cycle of the note, this will
produce a_ perfect square
wave. The actual music
program organization is a bit
Michael Gipe
~e oh
beeeeece ~
e -%-=
Control Panel
Speaker
Backplane Interconnections
Ne
KLUGEHARP Peripheral
Control Panel Interface
CPU and Buffers (M6800)
The Kluge Harp peripheral and the KLUGEHARP program were concocted in order to test out a Motorola 6800 system's
operation. This photo shows a test bench mounting of the three main cards and control panel. The Kluge Harp peripheral, such
as it is, is the single isolated wire wrap socket in the foreground, with wires dangling from connections on the CPU card.
more complex and is
described in detail below.
Fig. 1 illustrates the
hardware as implemented in
my system. The 7437 circuit
is used to form the NAND
gate flip flop. This flip flop in
turn drives a_ parallel
combination of the two
remaining 7437 gates, acting
as a buffer. The output of
this buffer is used to drive the
speaker; an 8 Ohm 5” speaker
produced more than adequate
volume. (A 100 Ohm resistor
in series will limit the volume
level to spare the ear drums.)
Generating Music With
Program Loops
Fig. 2 illustrates the basic
concept of the one-channel
music generator, expressed in
a procedure-oriented language
for compactness. The main
program loop begins at line 2
of the listing — “DO
FOREVER’ means repeat
over and over again all the
lines of code down through
the “END” at the same
margin, found at line 17. This
is the main loop used to cycle
through the SCORE stored at
some point in memory as
pairs of note selection/length
data bytes.
Lines 3 to 4 compute the
“next’’ pointer to the SCORE
— incrementing NOTER by 2.
Then LNGTH is set equal to
the second byte of the
current pair, SCORE
(NOTER+1). The length
codes are taken from Table |
along with note codes when
you set up a SCORE, and
represent a fixed interval of
time for the note in question,
measured as the number of
cycles.
Line 6 begins a note length
loop which extends to line
14. This “note length” loop
repeats the generation of the
note a number of times
indicated by the length code
just retrieved. The note
generation is accomplished by
delaying a number of time
units (CPU states) set by the
pitch code found at
SCORE(NOTER), then
changing the state of the
output flip flop and repeating
the process. The loop at lines
8-10 counts down the pitch
code and has a fixed delay
multiplied by the pitch code
to give the time for one half
cycle of the desired
frequency. Lines 11 to 15
change the state of the Kluge
Harp output device (0 to 1,1
to 0) — remembering in the
software location IT what the
previous state was.
Generating Codes
Table | is a reference table
of 21 notes “roughly” spaced
at equal intervals on the well
tempered scale. The integer
numbers in the “divide ratio”
column were determined
using the prime number 137
as an arbitrary starting point
and calculating the integer
closest to the result of the
following formula:
(In(137}+ n tn(2)/12)
tae
Where e is the usual
mathematical number 2.717
++. and the natural logarithm
of x (base e) is indicated by
{n(x). This is the standard
mathematical calculation of
the musical “well tempered”
scale — the 8-bit
approximation used by the
Kluge Harp is not perfect by
any means, but comes close
enough for the purposes of
this project.
The length count columns
are determined based upon
the assembly language
generated code for this
15
Fig. 2, The KLUGEHARP program specified in a procedure-oriented
computer language.
CDNAMAGHAWN=
17
18
that for each
corresponding
length count column will
measure a nearly identical
interval of time. The formula
is:
Len = time / (oh + di# pcp)
where:
routine, so
pitch, the
KLUGEHARP:
DO FOREVER;
PROGRAM;
NOTER = NOTER + 2;
IF NOTER = NOTEND THEN NOTER = NOTESTART;
LNGTH = SCORE(NOTER+1); /* SECOND OF TWO BYTES */
DO FOR |
LNGTH TO 1 BY —1;
PITCH = SCORE(NOTER); /* FIRST OF TWO BYTES */
DO FOR J = PITCH TO 1 BY -1;
/* COUNT DOWN THE PITCH DELAY */
END;
IT = IT + (—127); /* SWITCH SIGN BIT OF IT */
IF IT O THEN
SETLOC = 0;/* SET FLIP FLOP WITH MEMORY REF */
ELSE
RESETLOC = 0; /* RESET FLIP FLOP WITH REF */
END;
END;
CLOSE KLUGEHARP;
Len = nth length count.
time is the total number of
states for one “beat” of the
music (e.g., the shortest
note).
oh is the overhead of the
length counting loop.
dt is the number of states in
Table I. Kluge Harp Synthesizer pitch/length specification codes (HEX).
n divide hi
ratio
SOMIYDTAWNHAOSNHOKHDHUHOS
a
81
86
91
97
102
108
115
122
129
137
145
154
163
173
183
194
205
217
230
244
16
the pitch count innermost
loop.
Pcp is the pitch count for the
nth frequency.
Table 1 shows the divide
ratio in decimal, a
hexadecimal equivalent note
pitch code, and seven
ex note Note Length Codes (second byte of pair)
code = 4 6 8 16 32
4D 19 32 «64 96 C8 - =
51 18 30 60 90 CO ~ -
56 17 20 5A 87 B4 al sd
5B 16 28 56 81 AC - =
61 14 29 61 TA AQ F3 =
66 13 27. 4D 74 9A E7 a
6c 12 25 «49 6—E 92 DB al
73 W 230°«43 68 8A CF ad
7A 10 21 a 62 82 C3 -
81 10 iF 3E 65D 7€ BA F8
89 OF 1D 3A 57° 74 AE EB
91 OE 1¢ «37 53 6E AS DC
9A oD 1A 34 4€ 68 9C DO
AS oc 19 31 4A 62 93 C4
AD oc 18 QF 47 5E 8D BC
B7 OB 16 2c 42 58 84 BO
c2 OB 15 2A 3F 54 7E A8
cD OA 14 28 3c 60 78 AO
D9 09 130 25 «438 «4A GF 94
E6 o9 12 23 35 #46 #69) «8C
F4 o8 11 21 32 42 63 84
Data assumed by KLUGEHARP:
NOTER: 16-bit (two-byte)
address value. Initialize to
point to the address of the
first byte of SCORE.
SCORE: An array of data in
memory containing the code
sequence of the music (see
Table II). Initialize with the
music of your heart's desire
or use the example of Table
iT
NOTEND: 16-bit address
value, the address of the last
byte of SCORE (must be an
even number).
NOTESTART: 16-bit address
value, the address of the first
byte of SCORE (must be an
even number).
SETLOC: An unimplemented
address location which if
referenced turns off one bit
among the high order address
lines, bit 14 in the author's
case.
RESETLOC: An
unimplemented address
location which if referenced
turns off one bit among the
high order address lines, bit
13 in the author's case.
Data used but not initialized:
LNGTH
PITCH
IT
iJ
Fig. 3. Motorola 6800 Code for KLUGEHARP program.
Address Data
F800 FE
F801 FA
F802 00
F803 08
F804 08
F805 FF
F806 FA
F807 ile}
F808 8c
F809 FCO
F80A 80
F80B 26
F80c 03
F80D cE
F80E FC
F80F 00
F810 FF
F811 FA
F812 00
F813 FE
F814 FA
F815 00
F816 6
F817 01
F818 5A
F819 26
F81A 03
F81B JE
F8IC FS
F81D 00
FS1E AG
F81F 00
F820 4A
F821 26
F822 FD
F823 86
F824 80
F825 BB
F826 FA
F827 02
F828 28
F829 05
F82A 7F
F82B BO
F82C 00
F82D 20
F82E 03
F82F 7F
F830 bo
F831 00
F832 B7
F833 FA
F834 02
F835 cis
F836 FS
F837 18
columns of hexadecimal
length codes weighted to 1, 2,
4, 6, 8, 16 and 32 unit
intervals of time. A note is
placed in the score by picking
a note code, putting it in an
even numbered byte, then
placing a length code from
the same line of the table in
the odd numbered byte
which follows it. The actual
Add 2 to location in score
by incrementing and then
saving 16-bit new address
compare against immediate
Skip if not at end...
otherwise recycle
save in either case .. .
This is superfluous!
Skip if length remains .. .
Data allocations for KLUGE-
HARP:
FAOO-FAO1 = Current
pointer to SCORE, NOTER,
which should be initialized to
FCOO before starting the
program.
FAO2 = IT — an arbitrary
initialization will do.
FAO3-FFF7 = memory area
available for SCORE — the
example uses FCOO to FC7F
and puts the relevant
initializations into locations
F809-F80A (NOTEND) and
(NOTESTART).
NOTE: In the label column,
the numbers followed by colons
(e.g, “)are used to cate
F80E-F80F
Label Opcode Operand
KLUGEHARP: Lox NOTER
3:
INX
INx
STX NOTER
4: cPx #NOTEND
NOTEND: (last address of
SCORE plus 2)
BNE
4342
LDx #NOTESTART
NOTESTART: (first address of
score...)
STX NOTER
Lx NOTER
LDAB 1,X
LENGTH: DECB
6: BNE
74243
Jmp KLUGEHARP Restart piece
LDAA 0.x
FLOOP: DECA
8: BNE FLOOP
42:3
WW: LDAA #80
(127)
ADDA IT
12: BMI
4245
13: CLR SETLOC
(address with bit 14 off...)
BRA
4243
15: CLR RESETLOC
(address with bit 13 off...)
STAA IT
16: Jue LENGTH
pitches you'll get from these
codes depend upon the
details of the algorithm in
your own particular
computer and the clock rate
of the computer. For the
6800 system on which Kluge
Harp was first implemented,
the lowest note (code F4) is
approximately 170 Hz with a
500 kHz clock - and the unit
interval of time is
approximately 2000 CPU
states or about 4
milliseconds.
The hand assembled
M6800 code for the
KLUGEHARP program is
listed in Fig. 3. The
mnemonics and notations
have been taken from the
Motorola M6800
corresponding places in the high
level language version of the
program of Fig. 2.
In the system for which this
program was written, all active
memory is found at addresses
F800 to FFFF. Thus for all
normal program activity, bits A14
and A13 at the back plane of the
system are logical ‘1. When the
location SETLOC (B000) is
cleared, the high order address
Portion changes and bit A14 goes
to negative for a short time,
setting the Kluge Harp flip flop.
When the location A13 is cleared
(D000) on an alternate cycle,
address bit A13 goes to logical 0
for a short timer resetting the
Kluge Harp flip flop . ..
7
Table Il. WOLFGANG: Set the content of SCORE in memory to the
codes in this table — given for the addresses of the M6800 program
version — and KLUGEHARP will play four bars from the classical
period.
6800 Address Value 6800 Address Value
FCOO 90345 Fcao 5856 7
Fco2 9034 FC42 5856
FCo4 9034 Fc44 5856
FCO6 9034 FCa6 5B56
FCO8 9A34 | Note 1 Fc4B 5856 | Note8
FCOA 9034 FCaA 5B56
FCoc 9434 Feac 5B56
FCOE 9A344 FC4E 5B56 J
FC10 7Aaa1y FC50 6640 7
FCI2 7041 | Note 2 FC52 664D | Note
FC14 7Aa1 FC54 664D
FC16 7aai J FC56 6640 J
FC18 66409 FC58 4064 J
FCIA 664D | Note 3 FCSA 4D64 | Note 10
FCIC 6640 FC5C 4064
FCIE 6640 FCSE 4064 J
FC20 3319 FC6O 6640 7
FC22 A331 FC62 6640
FC24 A331 | Note 4 FC64 664D | Note 11
FC26 A331 FC66 eeap J
FC28 A331
FC2A A331 J FC68 7343 — Note 12
FC6A 664D — Note 13 NOTE: This program is
Fc2c 9A34 — Note 5 FC6C 7343 - Note14 — gimpleminded and not at all
FC6E 7A41 — Note 1S optimized. As a challenge to
FC2E 893A — Note 6 FC70 7343 — Note16 — yeaders, figure out a way to make
the notation more compact yet
FC30 9A34 FC72 7A4a1 preserving the total length of each
FC32 9034 FC74 7Aat pate:
FC34 9A34 FC76 7Aat
FC36 9934 | Note 7 FC78 7A41 | Note 17
FC38 9034 FC7A 7A41
FC3A 9034 FC7C 7A41
FC3C 9A34 FCTE TAAt
FC3E 9A34] FC80 —_(end pointer points here)
7 . Fig. 4. Timing of the Kluge Harp Output Waveform, At is the amount
Microprocessor Programming of time spent in the inner loop, and is set by choice of pitch codes. AT
Manual available from the is the length of the note, measured as a count of half-cycles at its
manufacturer. frequency. See Table I for a consistent set of length codes.
While not the greatest « AT =
musical instrument in the
world, the Kluge Harp
represents an interesting and
challenging diversion. The
program presented here is by
no means the ultimate in 2A Wy
music systems — and can FIG! |_|
serve as a basis for further
experimentation and
elaboration. Some challenges
for readers: modify the Al4 LINE
program to change the
es At oe At —ohe—At — +}
1
I
i
frequency of the notes 7 |
without changing the SCORE clr settoc |!
data; write another (longer) INSTRUCTION |
music program which only i i
specifies the pitch 1
code/length information once AS"TINE. J}
— and represents the score as
a series of one-byte indices 41
i i CLR RESETLOC
into the table of pitch Rie Reset
code/length information. EXECUTED
18
ALTAIR 8800 USERS!
Did you know...
That ali our modules are 100% compatible with the Altair 8800
computer, NO modifications necessary!
© That our 4KRA Static Read/Write Memory module doesn’t have
to lose it’s data when you pull the plug!
* That our 3P+S Input/Output module will fully interface two TV
Typewriters with keyboards and a modem or teletype at the
same time!
© That we make the most powerful alphanumeric Video Display
module anywhere!
© That our software is FREE, or close to i
© That all our modules are truly high quality, computer grade, but
that our prices are the lowest in the industry!
* That we have already shipped hundreds of modules on time, and
we will continue to deliver what we promise, FAST!
CHECK THE SPECS:
4KRA Static Read/Write Memory
This 4096 word STATIC memory provides faster, more reliable
and less expensive operation than any currently available dynamic
memory system. The 4KRA permits Altair 8800 operation at
absolute top speed continuously. All RAM's (Rancom Access
Memories) used in the 4KRA are 91L02A’s by Advanced Micro
Devices, the best commercial memory IC on the market today.
91L02A‘s require typically 1/3 the power of standard 2102 or
8101 type RAM'’s and each one is manufactured to military
specification MIL STD-883 for extremely high reliability. These
memories can be operated from a battery backup supply in case
of power failure with very low standby power consumption. (Ask
for our technical bulletin TB-101 on power down operation.) In
short we have done everything we could to make the best 4K
memory module in the computer field, and because we buy in
large quantity, we can make it for a very reasonable price.
Available now.
2KRO Erasable Reprogrammable Read Only Memory Module
With this module the Altair 8800 can use 1702A or 5203 type
Erasable Reprogrammable ROM’s. The 2KRO accepts up to eight
of these IC's for a capacity of 2048 eight bit words. Once
programmed this module will hold its data indefinitely whether
or not power is on, This feature is extremely useful when
developing software. All necessary bus interfacing logic and
regulated supplies are provided but NOT the EPROM IC's. Both
1702A and 5203 PROM's are available from other advertisers in
this magazine for well under $25. Available now.
3P+S Input/Output Module
Just one 3P4S card will fulfill the Input/Output needs of most
8800 users, There are two B-bit parallel input and output ports
with full handshaking logic. There is also a serial 1/O using a
UART with both teletype current loop and EIA RS-232 standard
interfaces provided, The serial data rate can be set under software
control between 36 and 9600 Baud. You can use your old model
19 TTY! This module gives you all the electronics you need to
interface most peripheral devices with the Altair 8800, it’s really
the most useful and versatile I/O we've seen for any computer.
Available now.
MB-1 Mother Board
Don't worry any more about wiring hundreds of wires in your
Altair to expand the mainframe. Our single piece 1/8-inch thick,
rugged mother board can be installed as one single replacement
for either three or four 88EC Expander cards, so you don't have
to replace your already installed 88EC card if you don’t want to.
The MB-1 has very heavy power and ground busses and comes
with a piece of fiat ribbon cable for connection to the front panel
board of the 8800, Available now.
VDM-1 Video Display Module
This module is the first real computer terminal display in kit
form. Under software control the VDM-1 displays sixteen 64
character lines to any standard video monitor. Characters are
produced in a 7x9 dot matrix, with a full 128 character set, upper
and lower case plus control characters. Data is accessed by the
VDM as a block from any 1K segment within the 65K address
range of the 8800 computer. Multiple cursors are completely
controlled by software and the display can begin anywhere on the
screen (this is great for many video games). When the last line is
filled the display scrolls up a line, Powerful editing capabilities are
provided with the FREE software package included in every
VDM-1 kit. Available in September ‘75.
SOFTWARE
Our Assembler, Text Editor and System Executive is being
shipped now. This software package gives you very powerful
Assembly Language capability in the Altair 8800. The Executive
and Editor allow you to call programs by name (including
BASIC) and then add, delete, change, or list programs by line
number. The Assembler provides a formatted symbolic mnemonic
listing as well as octal or binary object code from Assembly
Language programs written using the Editor. The Assembler also
gives valuable error messages to help in debugging those inevitable
errors. The Assembler, Editor, Executive Package No. 1 will be
available in read only memory along with an expanded Executive
and a powerful Interpretive Simulator by October or November
of 1975.
We are working on two BASIC Language packages which should
be ready by October. One will be a basic BASI€ needing about
BK of memory as a minimum and the other will be an Extended
version with additional string manipulation, matrix operations
and double precision arithmetic capabilities requiring about 12K
Both these packages will be available in Read Only Memory for 2
reasonable price.
PRICE LIST
Item Kit Assembled Delivery
2KRO EPROM module $50. $75, 2weeks ARO
3P+S 1/0 module 125. 165. 3 weeks ARO
4KRA-2 RAM module
w/2048 8-bit words 135. 185. 2 weeks ARO
4KRA-4 w/4096 B-bit
words of RAM 215, 280. 2 weeks ARO
RAM only, AMD 91LO2A
500n sec low power 8/$40 0 — 2 weeks ARO
MB-1 Mother Board 3.0 2weeks ARO
VDM-1 Video Display module 160. 225. Sept. 29, 75
then 3weeks ARC
Send for our FREE flyer for more complete specifications and
for pricing on additional items,
TERMS: All items postpaid if full payment accompanies order.
COD orders must include 25% deposit. MasterCharge gladly
accepted, but please send us an order with your signature on it.
DISCOUNTS: Orders over $375 may subtract 5%; orders over
$600 may subtract 10%.
Berkeley, Ca.94710 (415) 549.0857
by
Don Lancaster
We can get between a TV
typewriter and a television
style display system either by
an rf modulator or a direct
video method.
In the rf modulator
method, we build a
miniature, low power, direct
wired TV transmitter that
clips onto the antenna
terminals of the TV set. This
has the big advantage of
letting you use any old TV
set and ending up with an
essentially free display that
can be used just about
anywhere. No set
modifications are needed, and
you have the additional
advantage of automatic safety
isolation and freedom from
hot chassis shock problems.
There are two major
restrictions to the rf
modulator method. The first
of these is that transmitters
of this type must meet
Television
Anyone with a bunch of memory circuits, control logic and
a wire wrap gun can whip up a digital video generator with
TTL output levels. The problem as | see it is to get that digital
video signal into a form that the TV set can digest. The care
and feeding of digital inputs to the TV set is the subject of
Don Lancaster’s contribution to BYTE 2 — an excerpt from
his forthcoming book, 7V Typewriter Cookbook, to be
published by Howard W. Sams, Indianapolis, Indiana.
. +. CARL
certain exactly spelled out
FCC regulations and that
system type approval is
required. The second
limitation is one of
bandwidth. The best you can
possibly hope for is 3.5 MHz
for black and white and only
3 MHz for color, and many
economy sets will provide far
less. Thus, long character line
lengths, sharp characters, and
premium (lots of dots)
character generators simply
aren't compatible with
clip-on rf entry.
In the direct video
method, we enter the TV set
immediately following its
video detector but before
sync is picked off. A few
premium TV sets and all
monitors already have a video
input directly available, but
these are still expensive and
rare. Thus, you usually have
to modify your TV set, either
Fig. 1. Standard video interface levels. (Source impedance = 72 or 100 Ohms.)
WHITE
)
a
BLACK LEVEL
SYNC LEVEL
(OPTIONAL,
GRAY
Lov
LEVEL
2VOLTS
0.5 VOLTS
== OvoLTs
adding a video input and a
selector switch or else
dedicating the set to exclusive
TV typewriter use. Direct
video eliminates the
bandwidth restrictions
provided by the tuner, i-f
strip, and video detector
filter. Response can be
further extended by removing
or shorting the 4.5 MHz
sound trap and by other
modifications to provide us
with longer line lengths and
premium characters. No FCC
approval is needed, and
several sets or monitors are
easily driven at once without
complicated distribution
problems.
There are two limitations
to the direct video technique.
One is that the set has to be
modified to provide direct
video entry. A second, and
far more severe, restriction, is
that many television sets are
“hot chassis” or ac-de sets
with one side of their chassis
connected to the power line.
These sets introduce a severe
shock hazard and cannot be
used as TV typewriter video
entry displays unless some
isolation technique is used
with them. If the TV set has a
power transformer, there is
usually no hot chassis
problem. Transistor television
sets and IC sets using no
vacuum tubes tend to have
power transformers, as do
older premium tube type sets.
All others (around half the
sets around today) do not.
Direct Video Methods
With either interface
approach, we usually start by
getting the dot matrix data,
blanking, cursor, and sync
signals together into one
composite video signal whose
Interface
form is useful to monitors
and TV sets. A good set of
standards is shown in Fig. 1.
The signal is dc coupled and
always positive going. Sync
tips are grounded and blacker
than black. The normal open
circuit black level is positive
by one-half a volt, and the
white level is two volts
positive. In most TV camera
systems, intermediate levels
between the half volt black
level and the two volt white
level will be some shade of
gray, proportionately brighter
with increasing positive
voltage. With most TV
typewriter systems, only the
three states of zero volts
(sync), half a volt (black),
and two volts (white dot)
would be used. One possible
exception would be an
additional one volt dot level
for a dim but still visible
portion of a message or a
single word.
The usual video source
impedance is either 72 or 100
Ohms. Regardless of how far
we travel with a composite
video output, some sort of
shielding is absolutely
essential.
For short runs from board
to board or inside equipment,
tightly twisted conductors
should be OK, as should
properly guarded PC runs.
Fully shielded cables should
be used for interconnections
between the TVT and the
monitor or TV set, along with
other long runs. As long as
the total cable capacitance is
less than 500 pF or so (this is
around 18 feet of RG178-U
miniature coax), the receiving
end of the cable need not be
terminated in a 72 or 100
Ohm resistor. When
terminated cable systems are
in use for long line runs or
multiple outputs, they should
be arranged to deliver the
signal levels of Fig. 1 at their
output under termination.
Generally, terminated cable
systems should be avoided as
they need extra in the way of
drivers and supply power.
The exact width of the
horizontal and vertical sync
pulses isn’t usually too
important, so long as the
shape and risetime of these
pulses are independent of
position control settings and
power supply variations. One
exception to this is when
you’re using a color receiver
and a color display. Here, the
horizontal sync pulse should
be held closely to 5,1
microseconds, so the
receiver's color burst
sampling does in fact
intercept a valid color burst.
More on this later.
Intentional Smear
Fig. 2 shows us a typical
composite video driver using
a 4066 quad analog switch. It
gives us a 100 Ohm output
impedance and the proper
signal levels. Capacitor C1 is
used to purposely reduce the
video rise and fall times. It is
called a smearing capacitor.
Why would we want to
further reduce the bandwidth
and response of a TV system
that’s already hurting to
begin with? In the case of a
quality video monitor, we
wouldn’t. But if we’re using
an ordinary run-of-the-mill
TV set, particularly one using
rf entry, this capacitor can
Fig. 2, Analog switch combiner gen
+5
H SYNC 9-8
+8
CURSOR 8
+5
vere ots
+5) (WHITE)
VIDEO DOTS 8
+5
erates composite video.
cl
SMEARING
CAPACITOR
(SEE TEXT)
GOOCOOCOO
4066 (CMOS)
ANALOG SWITCH
680
100
very much improve the
display legibility and
contrast. Why?
Because we are interested
in getting the most legible
character of the highest
contrast we can. This is not
necessarily the one having the
sharpest dot rise and fall
times. Many things interact to
determine the upper video
response of a TV display.
These include the tuner
settings and the i-f response
and alignment, the video
detector response, video
peaking, the sound trap
setting, rf cable reflections,
and a host of other responses.
Many of these stages are
underdamped and will ring if
fed too sharp a risetime
input, giving us a ghosted,
$0 veo out
- AMI,
SYNC + OV
BLACK#0.4V
WHITE *1.5V
shabby, or washed out
character. By reducing the
video bandwidth going into
the system, we can move the
dot matrix energy lower in
frequency, resulting in
cleaner characters of higher
contrast.
For most TV displays,
intentional smearing will help
the contrast, legibility, and
overall appearance. The
ultimate limit to this occurs
when the dots overlap and
become illegible. The
21
Fig. 3. Block diagram of typical B and W television.
ANTENNA
SOUND nf
PROCESSOR SPEAKER
DIRECT VIDEO
ENTERS HERE
VIDEO I-F
L,| vioeo VIDEO
TER ae DETECTOR [Jamptirier [| CRT
SYNC
EXTRACTOR
POWER
SUPPLY
SWEEP.
CIRCUITS
optimum amount of true of recent small screen,
intentional smear is usually
the value of capacitance that
is needed to just close the
inside of a ‘“W" presented to
the display.
Adding a Video Input
Video inputs are easy to
add to the average television
set, provided you follow
some reasonable cautions.
First and foremost, you must
have an accurate and
complete schematic of the set
to be modified, preferably a
Sams Photofact or something
similar. The first thing to
check is the power supply on
the set. If it has a power
transformer and has the
chassis properly safety
isolated from the power line,
it's a good choice for a TVT
monitor. This is particularly
solid state portable TV sets.
On the other hand, if you
have a hot chassis type with
one side of the power line
connected to the chassis, you
should avoid its use if at all
possible. If you must use this
type of set, be absolutely
certain to use one of the
safety techniques outlined
later in Fig. 8.
A block diagram of a
typical TV set appears in Fig.
3. UHF or VHF signals
picked up by the tuner are
downconverted in frequency
to a video i-f frequency of 44
MHz and then filtered and
amplified. The output of the
video i-f is transformer
coupled to a video detector,
most often a small. signal
germanium diode. The video
detector output is filtered to
remove the carrier and then
routed to a video amplifier
made up of one or more
tubes or transistors.
AL some point in the video
amplification, the black and
white signal is split three
ways. First, a reduced
bandwidth output routes
syne pulses to the syne
separator stage to lock the
set's horizontal and vertical
scanning to the video. A
second bandpass output
sharply filtered to 4.5 MHz
extracts the FM sound
subcarrier and routes this to a
sound i-f amplifier for further
processing. The third output
is video, which is strongly
amplified and then
capacitively coupled to the
cathode of the picture tube.
The gain of the video
amplifier sets the contrast of
the display, while the bias
setting on the cathode of the
picture tube (with respect to
its grounded control grid) sets
the display brightness.
Somewhere in the video
amplifier, further rejection of
the 4.5 MHz sound subcarrier
is usually picked up to
minimize picture
interference. This is called a
sound trap. Sound traps can
be a series resonant circuit to
ground, a parallel resonant
circuit in the video signal
path, or simply part of the
transformer that is picking
off the sound for more
processing.
The video detector output
is usually around 2 volts peak
to peak and usually subtracts
from a white level bias
setting. The stronger the
signal, the more negative the
swing, and the blacker the
picture. Sync tips are blacker
than black, helping to blank
the display during retrace
times.
Fig. 4 shows us the typical
video circuitry of a transistor
black and white television.
Our basic circuit consists of a
diode detector, a unity gain
emitter follower, and a
variable gain video output
stage that is capacitively
coupled to the picture tube.
The cathode bias sets the
brightness, while the video
gain sets the contrast.
Amplified signals for sync
and sound are removed from
the collector of the video
driver by way of a 4.5 MHz
resonant transformer for the
sound and a low pass filter
for the sync. A_ parallel
resonant trap set to 4.5 MHz
eliminates sound interference.
Peaking coils on cach stage
extend the bandwidth by
providing higher impedances
VIDEO
LAST VIDEO DETECTOR
FF OXEMR
L
t
a
S|
> /
and thus higher gain to high
frequency video signals.
Note particularly the
biasing of the video driver. A
bias network provides us with
a stable source of 3 volts. In
the absence of input video,
this 3 volts sets the white
level of the display, as well as
establishing proper bias for
both stages. As an increasing
signal appears at the last
video output transformer, it
is negatively rectified by the
video detector, thus lowering
the 3 volts proportionately.
The stronger the signal, the
blacker the picture. Sync will
be the strongest of all, giving
us a blacker than black bias
level of only one volt.
The base of our video
driver has the right sensitivity
we need for video entry,
accepting a maximum of a 2
volt peak to peak signal. It
also has the right polarity, for
a positive going bias level
means a whiter picture. But,
an unmodified set is already
biased to the white level, and
if we want to enter our own
video, this bias must be
shifted to the black level.
We have a choice in any
TV of direct or ac coupling of
our input video. Direct
coupling is almost always
better as it eliminates any
Fig. 4. Typical video circuitry of transistor B and W TV set.
+12V
« [a
BRIGHTNESS
*800———w
HOOK
agus +150V
UN!
PICKOFF BBUND 6800
3V (WHITE)
Iv (SYNC)
v
VIDEO
DRIVER 620nH
(PEAKING)
22
i ov yt
r rhs
) sti8.,
UT PU
250KH (L8v)
(PEAKING) 3 479 nee
.SMHz
SOUND ac
TRAP
BIAS
SOURCE 100pF
200 $+
th
CONTRAST
PICTURE TUBE
+12KV
shading effects or any change
of background level as
additional characters are
added to the screen. Fig. 5
shows how we can direct
couple our video into a
transistor black and white set.
We provide a video input,
usually a BNC or a phono
jack, and route this to a PNP
Darlington transistor or
transistor pair, borrowing
around 5 mils from the set’s
+12 volt supply. This output
is routed to the existing video
driver stage through a SPDT
switch that either picks the
video input or the existing
video detector and _ bias
network.
The two base-emitter
diode drops in our Darlington
transistor add up to a 1.2 volt
positive going offset; so, in
the absence of a video input
or at the base of a sync tip,
the video driver is biased to a
blacker than black sync level
of 1.2 volts. With a white
video input of 2 volts, the
video driver gets biased to its
usual 3.2 volts of white level.
Thus, our input transistor
provides just the amount of
offset we need to match the
white and black bias levels of
our video driver. Note that
the old bias network is on the
other side of the switch and
does nothing in the video
position.
Two other ways to offset
our video input are to use
two ordinary transistors
connected in the Darlington
configuration, or to use one
transistor and a series diode
24
Fig. 5. Direct coupled video uses 1.2 volt offset of Darlington
transistor as bias.
OV (SYNC)
2V (WHITE)
VIDEO
INPUT
to pick up the same amount
of offset, as shown in Fig. 5.
If more or less offset is
needed, diodes or transistors
can be stacked up further to
pick up the right amount of
offset.
The important thing is
that the video driver ends up
with the same level for white
bias and for black bias in
either position of the switch.
L2v (SYNC) [RF
S.2VWHITE)
DARLINGTON
EXISTING
DETECTOR &
BIAS
VIDEO
TRANSISTOR
PAIR |
Ac or capacitively coupled
video inputs should be
avoided. Fig. 6 shows a
typical circuit. The TV’s
existing bias network is
lowered in voltage by adding
a new parallel resistor to
ground to give us a voltage
that is 0.6 volts more positive
than the blacker than black
sync tip voltage. For instance,
with a 3 volt white level, and
Fig, 6. Ac coupled video needs shift of bias to black level plus a
clamping diode.
*New components.
EXISTING
VIOEO
DRIVER
VIDEO* a
2N5139
EXISTING +12
VIDEO
ORIVER
TRANSISTORS
onaaoz ¥ INII4
ALTERNATE
USING TRANSIS-
TOR AND DIODE
2 volt peak to peak video, the
sync tip voltage would be 1
volt; the optimum bias is then
1.6 volts. Input video is
capacitively coupled by a
fairly large electrolytic
capacitor in parallel with a
good high frequency
capacitor. This provides for a
minimum of screen shading
and still couples high
frequency signals properly. A
clamping diode constantly
clamps the sync tips to their
bias value, with the 0.6 volt
drop of this diode being
taken out by the extra 0.6
volts provided for in the bias
network. This clamping diode
automatically holds the sync
tips to their proper value,
regardless of the number of
white dots in the picture.
Additional bypassing of the
bias network by a_ large
electrolytic may be needed
for proper operation of the
clamping diode, as shown in
Fig. 6. Note that our bias
network is used in both
switch positions — its level is
shifted as needed for the
direct video input.
Tube type sets present
about the same interface
problems as the solid state
versions do. Fig. 7 shows a
typical direct coupled tube
interface. In the unmodified
Fig. 7. Direct coupled video added to tube type B and W television.
+140
100x8 4 +
FROM VIDEO NIDEOSAME. crt
LAST
Vibe! es RE, 130v |-——+ catHove
ue VIDEO. CONTRAST
\ (2V- VIDEO)
T T (OV- RF)
\
‘ + *
FRIOOpF
* icon
VIDEO ft NEW CATHODE SELF-BIAS
INPUT SHIFTS TO SYNC LEVEL IN
VK VIDEO POSITION
mm
*New components,
circuit, the white level is zero
volts and the sync tip black
level is minus two volts. If we
can find a negative supply
(scarce in tube type circuits),
we could offset our video in
the negative direction by two
volts to meet these bias levels.
Instead of this, it is usually
possible to self bias the video
amplifier to a cathode voltage
of +2 volts. This is done by
breaking the cathode to
ground connection and
adding a small resistor (50 to
100 Ohms) between cathode
and ground to get a cathode
voltage of +2 volts. Once this
value is found, a heavy
electrolytic bypass of 100
microfarads or more is placed
in parallel with the resistor.
Switching then grounds the
cathode in the normal rf
mode and makes it +2 volts in
the video entry mode.
In the direct video mode, a
sync tip grounded input
presents zero volts to the
grid, which is self biased
minus two volts with respect
to the cathode. A white level
presents +2 volts to the grid,
which equals zero volts grid
to cathode.
Should there already be a
self bias network on the
cathode, it is increased in
value as needed to get the
black rather than white level
bias in the direct video mode.
Hot Chassis Problems
There is usually no shock
hazard when we use clip-on rf
entry or when we use a direct
video jack on a transformer-
powered TV. A very severe
shock hazard can exist if we
use direct video entry with a
TV set having one side of the
power line connected to the
chassis. Depending on which
way the line cord is plugged
in, there is a 50-50 chance of
the hot side of the power line
being connected directly to
the chassis.
Hot chassis sets,
particularly older, power
hungry tube versions, should
be avoided entirely for direct
video entry. If one absolutely
must be used, some of the
suggestions of Fig. 8 may ease
the hazard. These include
using an isolation
transformer, husky
back-to-back filament
transformers, three wire
power systems, optical
coupling of the video input,
25
and total package isolation.
Far and away the best route
is simply never to attempt
direct video entry onto a hot
chassis TV.
Making the Conversion
Fig. 9 sums up how we
modify a TV for direct video
entry. Always have a
complete schematic on hand,
and use a transformer style
TV set if at all possible. Late
models, small screen, medium
to high quality solid state sets
are often the best display
choice. Avoid using junk sets,
particularly very old ones.
Direct coupling of video is far
preferable to ac capacitor
coupling. Either method has
to maintain the black and
white bias levels on the first
video amplifier stage. A shift
of the first stage quiescent
bias from normally white to
normally black is also a must.
Use short, shielded leads
between the video input jack
and the rest of the circuit. If
a changeover switch is used,
keep it as close to the rest of
the video circuitry as you
possibly can.
Extending Video and Display
Bandwidth
By using the direct video
input route, we eliminate any
bandwidth and response
restrictions of an rf
26
modulator, the tuner, video
i-f strip, and the video
detector filter. Direct video
entry should bring us to a 3
MHz bandwidth for a color
set and perhaps 3.5 MHz for a
black and white model, unless
we are using an extremely
bad set. The resultant 6 to 7
million dot per second rate is
adequate for short character
lines of 32, 40, and possibly
48 characters per line. But
the characters will smear and
be illegible if we try to use
longer line lengths and
premium (lots of dots)
character generators on an
ordinary TV. Is there
anything we can do to the set
to. extend the video
bandwidth and display
response for these longer line
lengths?
In the case of a color TV,
the answer is probably no.
The video response of a color
set is limited by an essential
delay line and an essential
3.58 MHz trap. Even if we
were willing to totally
separate the chrominance and
luminance channels, we'd still
be faced with an absolute
limit set by the number of
holes per horizontal line in
the shadow mask of the tube.
This explains why video color
displays are so expensive and
so rare. Later on, we'll look
at what's involved in adding
color to the shorter line
lengths.
With a black and white
TV, there is often quite a bit
Fig. 8. Getting Around a Hot Chassis Problem.
Hot chassis problems can be avoided entirely by
using only transformer-powered TV circuits or
by using clip-on rf entry. If a hot chassis set
must be used, here are some possible ways
around the problem:
1, Add an isolation transformer.
A 110 volt to 110 volt isolation trans-
former whose wattage exceeds that of the
set may be used. These are usually expen-
sive, but a workable substitute can be made
by placing two large surplus filament trans-
formers back to back. For instance, a pair
of 24 volt, 4 Amp transformers can handle
around 100 Watts of set.
2. Use a three wire system with a solid ground.
Three prong plug wiring, properly
polarized, will force the hot chassis connec-
tion to the cold side of the power line. This
protection is useful only when three wire
plugs are used in properly wired outlets. A
severe shock hazard is reintroduced if a
user elects to use an adaptor or plugs the
system into an unknown or improperly
wired outlet. The three wire system should
NOT be used if anyone but yourself is ever
to use the system.
3. Optically couple the input video.
Light emitting diode-photocell pairs are
low in cost and can be used to optically
couple direct video, completely isolating
the video input from the hot chassis. Most
of these optoelectronic couplers do not
have enough bandwidth for direct video
use; the Litronix 1L-100 is one exception.
Probably the simplest route is to use two
separate opto-isolators, one for video and
one for sync, and theri recombine the
signals inside the TV on the hot side of the
circuit.
4. Use a totally packaged and sealed system.
If you are only interested in displaying
messages and have no other input/output
devices, you can run the entire circuit hot
chassis, provided everything is sealed inside
one case and has no chassis-to-people
access. Interface to teletypes, cassettes,
etc., cannot be doné without additional
isolation, and servicing the circuit presents
the same shock hazards that servicing a hot
chassis TV does.
we can do to present long
lines of characters, depending
on what set you start out
with and how much you are
willing to modify the set.
The best test signal you
can use for bandwidth
extension is the dot matrix
data you actually want to
display, for the frequency
response, time delay, ringing,
and overshoot all get into the
act. What we want to end up
with is a combination that
gives us reasonably legible
characters.
A good oscilloscope (15
MHz or better bandwidth) is
very useful during bandwidth
extension to show where the
signal loses its response in the
circuit. At any time during
the modification process,
there is usually one response
bottleneck. This, of course, is
what should be attacked first.
Obviously the better a TV
you start with, the easier will
be the task. Tube type gutless
wonders, particularly older
ones, will be much more
difficult to work with than
with a modern, small screen,
quality solid state portable.
Several of the things we
can do are watching the
control settings, getting rid of
the sound trap, minimizing
circuit strays, optimizing spot
size, controlling peaking, and
shifting to higher current
operation. Let’s take a look
at these in turn.
Control Settings
Always run a data display
at the lowest possible
contrast and using only as
much brightness as you really
need. In many circuits, low
contrast means a lower video
amplifier gain, and thus less
of a gain-bandwidth
restriction.
Eliminate the Sound Trap.
The sound trap adds a
notch at 4.5 MHz to the
video response. If it is
eliminated or switched out of
the circuit, a wider video
bandwidth automatically
Fig. 9. How to Add a Direct Video Input to a TV Set.
Get an accurate and complete schematic of
the set — either from the manufacturer's
service data or a Photofact set. Do not try
adding an input without this schematic!
Check the power supply to see if a power
transformer is used. If it is, there will be no
shock hazard, and the set is probably a
good choice for direct video use. If the set
has one side of the power line connected to
the chassis, a severe shock hazard exists,
and one of the techniques of Fig. 8 should
be used. Avoid the use of hot chassis sets.
Find the input to the first video amplifier
stage. Find out what the white level and
sync level bias voltages are. The marked or
quiescent voltage is usually the white level;
sync is usually 2 volts less. A transistor TV
will typically have a +3 volt white level and
a +1 volt sync level. A tube type TV will
typically have a zero volt white level and a
-2 volt sync level.
Add a changeover switch using minimum
possible lead lengths. Add an input con-
nector, either a phono jack or the premium
BNC type connector. Use shielded lead for
interconnections exceeding three inches in
length.
Select a circuit that couples the video and
biases the first video amplifier stage so that
the white and sync levels are preserved. For
transistor sets, the direct coupled circuits
of Fig. 5 may be used. For tube sets, the
circuit of Fig. 7 is recommended. Avoid the
use of ac coupled video inputs as they may
introduce shading problems and changes of
background as the screen is filled.
Check the operation. If problems with
contrast or sync tearing crop up, recheck
and adjust the white and sync input levels
to match what the set uses during normal rf
operation. Note that the first video stage
must be biased to the white level during rf
operation and to the sync level for direct
video use. The white level is normally two
volts more positive than the sync level.
Fig. 10, Removing the sound trap can extend video bandwidth.
(a) Response (b) Parallel resonant trap —
short or bypass.
4.5 MHz
VIDEO
WITH SOUND
TRAP DRIVER
FREQUENCY
WITHOUT
FREQUENCY
CONTRAST
(4) Combined trap and
pickoff — open or SOUND I-F
(c) Series resonant trap —
open or remove.
remove (series resonant);
short or bypass (parallel
i
results. Fig. 10 shows us the
response changes and the
several positions for this trap.
Generally, series resonant
traps are opened and parallel
resonant traps are shorted or
bypassed through suitable
switching or outright
elimination. The trap has to
go back into the circuit if the
set is ever again used for
ordinary program reception.
Sometimes simply backing
the slug on the trap all the
way out will improve things
enough to be useful.
Minimizing Strays
One of the limits of the
video bandwidth is the stray
capacitance both inside the
video output stage and in the
external circuitry. If the
contrast control is directly in
the signal path and if it has
long leads going to it, it may
be hurting the response. If
you are using the TV set
exclusively for data display,
can you rearrange the control
location and simplify and
shorten the video output to
picture tube
interconnections?
28
Additional Peaking
Most TV sets have two
peaking networks. The first
of these is at the video
detector output and
compensates for the vestigial
sideband transmission signal
that makes sync and other
resonant).
VIDEO VIDEO
OUTPUT OUTPUT
CRT
SlSecal
VIDEO
OUTPUT
SOUND.
IF
45
MHz
low frequency signals double
the amplitude of the higher
frequency ones. The second
of these goes to the collector
or plate of the video output
stage and raises the circuit
impedance and thus the
effective gain for very high
Fig, 11. Adjusting the peaking coil can extend video response.
(a) Circuit. +
VIDEO
OUTPUT
(b) Response.
AMPLITUDE
COLLECTOR
LOAD
PEAKING
COIL
CRT
OUTPUT,
STRAY, &
INPUT
CAPACITY
TOO MUCH
L TOO LARGE
OPTIMUM
TOO LITTLE
L TOO SMALL
FREQUENCY
frequencies. Sometimes you
can alter this second network
to favor dot presentations.
Fig. 11 shows a_ typical
peaking network and the
effects of too little or too
much peaking. Note that the
stray capacitance also enters
into the peaking, along with
the video amplifier output
capacitance and the picture
tube’s input capacitance.
Generally, too little peaking
will give you low contrast
dots, while too much will give
you sharp dots, but will run
dots together and shift the
more continuous portions of
the characters objectionably.
Peaking is changed by in-
creasing or decreasing the
series inductor from its design
value.
Running Hot
Sometimes increasing the
operating current of the video
output stage can increase the
system bandwidth — IF this
stage is in fact the limiting
response, IF the power
supply can handle the extra
current, IF the stage isn’t
already parked at its
gain-bandwidth peak, and IF
the extra heat can be gotten
rid of without burning
anything up. Usually, you can
try adding a resistor three
times the plate or collector
load resistor in parallel, and
see if it increases bandwidth
by 1/3. Generally, the higher
the current, the wider the
bandwidth, but watch
carefully
limits. Be sure
extra ventilation and
additional heatsinking, and
check the power supply for
unhappiness as well. For
major changes in operating
current, the emitter resistors
and other biasing components
any dissipation
to provide
should also be
proportionately reduced in
value.
Spot Size
Even with excellent video
bandwidth, if you have an
out-of-focus, blooming, or
changing spot size, it can
completely mask character
sharpness. Spot size ends up
the ultimate limit to
resolution, regardless of video
bandwidth.
Once again, brightness and
contrast settings will have a
profound effect, with too
much of either blooming the
spot. Most sets have a focus
jumper in which ground or a
positive voltage is selected.
You can try intermediate
values of voltage for
maximum sharpness. Extra
power supply filtering can
sometimes minimize hum and
noise modulation of the spot.
Anything that externally
raises display contrast will let
you run with a smaller beam
current and a sharper spot.
Using circularly polarized
filters, graticule masks, or
simple colored filters can
Fig. 12, Contrast Enhancing
Filter Materials.
Circularly polarized filters:
Polaroid Corp.
Cambridge MA 02139
Anti-reflection filters:
Panelgraphic Corp.
10 Henderson Dr.
West Caldwell NJ 07006
Light control film:
3M Visual Products Div.
3M Center
St. Paul MN 55101
Acrylic plexiglas filter sheets:
Rohm and Haas
Philadelphia PA 19105,
Fig. 13. Standard rf interface levels. Impedance =
frequency per Fig. 14.
{HH
minimize display washout
from ambient lighting. Fig.
12 lists several sources of
material for contrast
improvement. Much of this is
rather expensive, with pricing
from $10 to $25 per square
foot being typical. Simply
adding a hood and
positioning the display away
from room lighting will also
help and is obviously much
cheaper.
Direct Rf Entry
If we want the con-
venience of a ‘‘free’’
display, the freedom from
hot chassis problems, and
“use it anywhere” ability,
direct rf entry is the obvious
choice. Its two big limitations
are the need for FCC type
approval, and a limited video
bandwidth that in turn limits
the number of characters per
line and the number of dots
per character.
An rf interface standard is
shown in Fig. 13. It consists
of an amplitude modulated
carrier of one of the standard
television channel video
frequencies of Fig. 14
Channel 2 is most often used
300Q. Carrier
SYNC TIPS=
100 % AMPLITUDE
4mvV RMS TYPICAL
BLACK =
75 % AMPLITUDE
| | 3mV RMS TYPICAL
°
| ~WHITE= 10% OR
LESS AMPLITUDE
.3mV RMS TYPICAL
with a 55.250 MHz carrier
frequency, except in areas
where a local commercial
Channel 2 broadcast is
intolerably strong. Circuit
cost, filtering problems, and
stability problems tend to
increase with increasing
channel number.
The sync tips are the
strongest part of the signal,
representing 100%
modulation, often something
around 4 millivolts rms across
a 300 Ohm line. The black
level is 75% of the sync level,
or about 3 millivolts for 4
millivolt sync tips. White level
is less than 10% of maximum.
Note that the signal is
weakest when white and
strongest when sync. This is
the exact opposite of the
video interface of Fig. 1.
Rf modulators suitable for
clip-on rf entry TV
typewriter use are called Class
1 TV Devices by the FCC. A
Class 1 TV device is supposed
to meet the rules and
regulations summarized in
Fig. 15.
Fig. 16 shows us a block
diagram of the essential parts
of a TV modulator. We start
Fig. 14, Television Picture
Carrier Frequencies.
Channel 2 . 55.25 MHz
Channel 3 . 61.25 MHz
Channel 4 . 67.25 MHz
Channel 5 . -77.25 MHz
Channel 6 ...... .83.25 MHz
Fig. 15, FCC Regulations on Class
1 TV Devices. More complete
information appears in subpart H
of Part 15 and subpart F of Part 2
of the Federal Communications
Commission Rules and Regula-
tions. It is available at many large
technical libraries.
A Class 1 TV device generates
a video modulated rf carrier of a
standard television channel
frequency. It is directly
connected to the antenna
terminals of the TV set.
The maximum rms rf voltage
must be less than 6 millivolts
using a 300 Ohm output fine.
The maximum rf voltage on
any frequency more than 3 MHz
away from the operating channel
must be more than 30 dB below
the peak in-channel output
voltage.
An antenna disconnect switch
of at least 60 dB attenuation must
be provided.
No user adjustments are
permitted that would exceed any
of the above specifications.
Residual rf radiation from
case, leads and cabinet must be
less than 15 microvolts per meter.
A Class 1 TV device must not
interfere with TV reception.
Type approval of the circuit is
required. A filing fee of $50 and
an acceptance fee of $250 is
involved.
with a stable oscillator tuned
to one of the Fig. 14
frequencies. A crystal
oscillator is a good choice,
and low cost modules are
widely available. The output
of this oscillator is then
amplitude modulated. This
can be done by changing the
bias current through a silicon
small signal diode. One
milliampere of bias current
makes the diode show an ac
and rf impedance of 26
Ohms. Half a mil will look
like 52 Ohms, and so on. The
diode acts as a variable
resistance attenuator in the rf
circuit, whose bias is set and
changed by the video circuit.
Since diode modulators
are non-linear, we can’t
simply apply a standard video
signal to them and get a
standard rf signal out. A
differential amplifier circuit
called a video slicer may be
used to compensate for this
non-linearity. The video slicer
provides three distinct
currents to the diode
modulator. One of these is
almost zero for the white
level, while the other two
provide the black and sync
levels. A contrast control that
sets the slicing level lets you
adjust the sync tip height
with respect to the black
level. The video slicer also
minimizes rf getting back into
the video. An attenuator to
reduce the size of the
modulated signal usually
follows the diode modulator.
An upper side band filter
removes most of the lower
sideband from the AM
modulated output, giving us a
30
Fig. 16. Block diagram of rf modulator.
crow
anoehna
-—---------—---—~-~~-|-~,
| ye e0 war ioe Oren antenna] |
|| cavsrar™ LI Nooucaroa bs) Sioesano fe! siscownecr [150
| [ B8eiEbron [P| BSR ENCRron [BCE Swen +
| ; er
| viogo |
o— +] S08,
viveo | |
INPUT RFI SHIELO J
vestigial sideband signal that
stays inside the channel band
limits. This same filter
eliminates second harmonic
effects and other spurious
noise. The filter’s output is
usually routed to an antenna
disconnect switch and the
TV’s antenna terminals. A
special switch is needed to
provide enough isolation.
Some of the actual
circuitry involved is shown in
Fig. 17. The video slicer
consists of a pair of high gain,
small signal NPN transistors,
while the oscillator is a
commercially available
module.
Rf entry systems always
must be direct coupled to the
antenna terminals of the set
and should never provide any
more rf than is needed fora
minimum snow-free picture.
They should be permanently
tuned to a single TV channel.
Under no circumstances
should an antenna or cable
service hookup remain
connected to the set during
TVT use, nor should
radiation rather than a direct
rf cable connection ever be
used.
Color Techniques
We can add a full color
capability to a TV typewriter
system fairly easily and
cheaply — provided its usual
black and white video dot
rate is low enough in
frequency to be attractively
displayed on an ordinary
color TV. Color may be used
to emphasize portions of a
message, to attract attention,
as part of an electronic game,
or as obvious added value toa
graphics display. Color
techniques work best on TV
typewriter systems having a
horizontal frequency very
near 15,735 Hertz.
All we basically have to do
is generate a subcarrier sine
wave to add to the video
output. The phase of this
subcarrier (or its time delay)
is shifted with respect to
what the phase was
immediately after each
horizontal sync pulse to
generate the various colors.
Fig. 18 shows us the
differences between normal
color and black and white
operation. Black and white
baseband video is some 4
MHz wide and has a narrow
4.5 MHz sound subcarrier.
The video is amplitude
modulated, while the sound is
narrow band frequency
Fig. 17. Channel two oscillator, modulator, video slicer and attenuator.
R sets output level.
OX OSCILLATOR MODULE &
EX CRYSTAL (INTERNATIONAL.
CRYSTAL) OR EQUIV
2N3643
+5V
ETC
VIDEO
INPUT
2N5129
2N4400
|-7T-1/4in. DIA (NO SLUG)
DIRECT COUPLE OUTPUT
iN914 or
igiaa 47°F op
CONTRAST
+5
IK 470
modulated. This translates up
to a 6 MHz rf channel witha
vestigial lower sideband as
shown in Fig. 18(b).
To generate color, we add
a new pilot or subcarrier at a
magic frequency of 3.579545
MHz — see Fig. 18(c). What
was the video is now called
the luminance, and is the
same as the brightness in a
black and white system. The
new subcarrier and_ its
modulation is called the
chrominance signal and
determines what color gets
displayed and how saturated
the color is to be.
Since the black and white
information is a sampled data
system that is scanned at the
vertical and horizontal rates,
there are lots of discrete holes
in the video spectrum that
aren't used. The color
subcarrier is designed to stuff
itself into these holes (exactly
in a NSTC color system, and
pretty much in a TVT
display). Bo
Showing the first 100000 of 299743 characters.