Please chime in and show us a picture
of your vintage cw keyboard
commercial or homebrew
Here’s a picture of my HAL MKB-1 keyboard. I built it from a kit back in 1975 when it sold for around $175. It was my favorite keyboard for about 25 years until I got a K1EL K20. As you can see there is no spacebar due to no buffering. If you held a key down long enough it would repeat the character. That could be an advantage for sending certain words, but getting off of the E key quickly enough could be a challenge at high speeds. The keyboard still works, but it would take me a few days of practice to be able to send on it at a good speed. Back in the good old days, I could send over 60 wpm on it.
Below here is a picture of my SkipJack - sold by the ECM Corporation
SkipJack 48B keyboard keyer that was sold as a kit in the late 1970
KF7CX, SkipJack internals
Well, in 67 I built a code typewriter using a typewriter with a bobby pin switch mounted beneath each key lever, a diode matrix, and a relay counter to control a transistor speed key. Then in 68 made a code typewriter using integrated circuits with a core matrix. Later adding a one character magnetic core memory and in 69 added a four character magnetic core memory to it. All of these machines and my 75 one were built using my plans although I , no doubt, made use of circuits from books and the many magazine articles. The May, 61 QST "Codemite" article used a semiconductor and magnetic core shift register.The Sep., 64 CQ article used a semiconductor and magnetic core shift register. The March, 52 Radio & Television News magazine told about a "Codetyper" using 40 miniature tubes.
Here is a picture of my 1975 Morse keyboard with a 64 character buffer memory and N-key rollover. The speed is set with the binary decade switches next to the weight control. The toggle switches allow you to pre-load a message into the memory and then send it once or over and over, or you can clear the memory with the clear button. The other switches select speaker or phone, transmit or local, or continuous dots. The memory counter display, tone, and volume controls are to the right. I used a ferrite core matrix to enter the characters into the First-in First-out memory chips.
OK, I'll chime in with the attached photo. This shows the internal view of a keyboard that used a set of Cherry ASCII encoded keys, probably built in the 80's but I do not recall now. The enclosure went through a number of iterations of insides and users as it got shipped around. I think in the mid-90s it went to K6KX until I sent him a more modern one. Not sure but I think W4TY also used it during one of it's iterations.
The circuit board shown probably also had my logic for accepting BAUDOT input and converting to ASCII and then to CW from the period where I made a few "RTTY" contacts, copying the incoming stuff as CW. I also converted the ASCII output to BAUDOT to drive a audio FSK circuit (fuzzy details in my head now). I made about 10 contacts total and told each one that I was not using a TTY machine or computer (I think some people were using early computers then). My recollection is that most of them had no idea what I was talking about and maybe two clearly understood and commented it was the dumbest thing they ever heard of.
I have no idea how many kbds I had built between the first three W4UX Touchcoder II (?) toroid encoded guys from the early 70's and this one, but would guess ten or more as I went through newer designs. W4TY ( né W4ZAU) supplied three sets of keypunch keyboards around 1973 after I had breadboarded the W4UX design and I built them up for Art, myself, and my elmer from high school. Running all those tiny wires through the toroid matrix was great fun of course.
I don't remember how long I used the unbuffered design, but probably within a year or so I started using FIFOs.
The kbd in the photo lived into my retirement move to Las Cruces, but a few years ago I removed the circuit board to save and tossed the poor old enclosure.
I find it interesting that a single chip (and a lot of software code) now implements all the features I used to do with many TTL and CMOS chips.
Chuck B. (W5UXH)
I built my keyboard keyer in college in the late 60's. I graduated in May 1970 from the Univ of Texas at Arlington so it was a couple of years before then that I designed and built the keyer. It's one of a kind. I had been looking at building a keybord keyer using ferrite cores. I visited a ham in Dallas with one with ferrite cores. But then I thought I might try transistor flip flops in a shift register. I built one with transistor flip flops but they proved to be sensitive to noise, and I abandoned that effort. While I was working at TI in Dallas one summer, one of the engineers told me to pick up some of those new integrated circuits (in buckets on the floor) and see what I could do with them. There were flip flops and nand and nor gates and timers, etc. I was in heaven. Here was what I needed to build a keyboard keyer at last! I laid out a circuit with a set of IC flip flops in a register. Seems like the flip flops were 7474 or something like that. I reasoned that a clock would shift them out the end. The last flip flop would be the one sending the cw. I used diodes to set the flip flop shift register for each different letter key. The diodes were also from TI. I designed the circuit, laid out a pc board, photo etched it, put on the components, made the case out of fiberglas wrapped around a balsa mold, used old typewriter keys, and you know what? The darn thing worked right off the bat! I had to learn how to sync my fingers with the sending of the letters since it held only one letter at a time in the shift register. Basically you have to press and hold a key until it starts sending that letter and then you go to the next letter before the sending letter is finished. You have to be quick when sending e's hi. A few years ago one of the electrolytic caps failed and I had to replace that. My Icom 756pro3 has about a 10 ms shortening of each dot and dash when in qsk mode, so I put in a relay and put a capacitor across the relay coil so the relay would hold in for an additional 10 ms and compensate for the qsk error in the Icom rig. Without this delay the dits become much lighter than 50% duty cycle using qsk on the IC756P3. With the compensation, the dits hold at 50% duty cycle even through 50 wpm. The keyer is now about 45 years old and it looks almost exactly the same as it did when it was brand new. I won an IEEE student paper contest the year after I finished it. I was light years ahead of my competition showing off that shiny new electonic keyer. IC's were new then so even the use of IC's was novel at that time. I use this keyboard keyer every morning on 7047 kHz on the waterway cw net that starts at 6 am central time (both daylight and non daylight savings time).
The keys are caps from an old typewriter. I made flat springs from a sheet of metal and screwed the end of a spring to the key top using a short sheet metal screw. The top cover the springs stick though is actually a single sided pc board made of fiberglass. Holes were put in the pc board so the keys appear on the other side. I then soldered the other end of the spring to the pc board. All the keys are grounded and the pc board provides a shield from my hand capacitance. A set of diodes is on another pc board beneath the pc board holding the keys. The diode pc board has brass screws sticking up that can be adjusted to provide a small clearance to each key. A set of diodes for an individual letter is connected in the pc diode board to the brass screw. Pressing a key down grounds the brass screw causing certain flip flops to be set to the other state thus loading the shift register with a specific letter. I listen to the cw being sent and move to the next letter accordingly. I have to manually leave the space bar space by ear. Since the clock is running it doesn't send the next letter until the correct time unless I hit the next letter too soon and it leaves no space at all. So it either leaves the correct space or none at all. I can tell if it leaves no space at all by listening to my cw and make an adjustment on my sending to correct my mistake the next time a space is needed. 73 de k5gp, gene www.egpreston.com
Bio—Wally – K5TO
First licensed in 1959 as KN5VCP. Upgraded to general and K5VCP in 1960. Upgraded to Extra in 70’s applied for and received K5TO call in 1977. In amateur radio 49-years.
My interest in CW continued over the years. In 1960 I had a TO-keyer (Hallicrafters HA-1 by Jim Ricks W9TO) used in conjunction with HT37 and SX-111. However, at speeds approaching 45-WPM the HA-1 left much to be desired in terms of accuracy which was very operator dependent.
I followed keyboard development since ‘60s when QST featured a ferrite-core keyboard. However, it did not have a space bar or any buffer memory. This design resulted in mechanical aberrations created by ops with less than perfect typing skills. A keyboard design in the 70’s by K4KN contained a 1-character buffer but no space bar. The operator created both letter and word spacing. My preference was a buffered design to eliminate letter and word spacing problematic with a 1-character buffer. A Curtis design was available in the ‘70s but was expensive. The solution was to design my own keyboard. Since I worked as a design engineer of geophysical displays at Shell Oil Company I decided to design a Morse keyboard from the ground up.
The design was a digital sequencer with a multi-phase clock. It contained TTL logic, FIFO’s, and PROM arrays. The keyboard encoder design outputted a Hexadecimal character directly into a First-In First-Out (FIFO) memory. In the 70’s PC keyboards were not available so it was necessary to also design a keyboard encoder to get keyboard character output in digital form. The FIFO output directly addressed a Programmable Read-Only Memory (PROM) which was clocked parallel into a serial register that serially outputted the appropriate Morse character representing all defined Morse characters. The multi-phase clock provided 8-phases of decoding and serially extracted the Morse character retrieved from the PROM. The logic circuitry was extremely fast in comparison to actual keyboard speeds encountered in amateur radio. In order to troubleshoot the design on an oscilloscope without flicker a speed of 1500-WPM was established totally beyond human capacity to copy or send. Thus the keyboard’s accuracy at normal operating speeds was unparalleled. Gene Farthing K4KHT one of the best QRQ ops in the world at that time told me that the keyboard produced the best sounding CW he’d ever heard. Of course this necessitated making keying circuit changes to the Kenwood T-599X to accommodate the keyboard’s fast speeds and to preserve the typical 3:1 dash-to-dot ratio without additional weighting although the board did contain weighting circuitry.
Several prototype boards were built using wire-wrap techniques requiring a relatively bulky enclosure. Production PCB’s and injection molded cases were evaluated but became cost prohibitive yet much less expensive than the Curtis. Soon other buffered keyboards came onto the scene, such as the “Skipjack” and Heathkit among others. Consequently the K5TO keyboard was never mass produced. The board was in use well into the 90’s when PC software evolved to the point of meeting the accuracy demands of QRQ CW. I now use a software design by VE6YP called YPLog, a logging program which has one of the best CW generators.
The ACCU-MILL project was developed by a young brilliant MIT graduate and I can recall purchasing the PROMs and PC board from him.It used a Standard ASCII style keyboard ( Provided by the user ) to the input to of his circuit board. His circuit was a ROM look table ( TWO Proms ) that provided an output to the famous ACCU-Keyer. It was amazing what some of the hams used as keyboards back then. I recall folks using the old Radio Shack TRS 80 computers for CW. The Commodore VIC-20 was also used to blast out CW. HeathKit made a CW keyboard - Marty (KW1C ) has one of these boards. There was a CW keyboard made by HAL. K0JVX - John, still uses his old SkipJack keyboard. Back in 1999 WA2TDL - Rick, was still using his Woody (K4KN) Keyboard.
My first keyboard was one called a Skipjack. After that I got an Info-Tech M-300 and then an update of that one, the M-300C by Info-Tech. The first board my dad had was an Info-Tech M-300 and the M-300C later.
These were originally designed and built here in St Louis (Westport) just a few miles from my house. Info Tech later moved to Florida just north of where my Dad, W4BI lived. Info Tech supplied Universal Radio in Ohio with not only keyboards but many digital mode decoding boxes for SWLs and some of the first RTTY TUs with CRTs. Info Tech is no longer in business. Dave Kelce was the founder of the company and a ham, K0DGF. The 300 and 300C were popular with the CFO because they sent very good CW, had a space bar and recallable memories. The 300 had individual key switches which were not as reliable as one piece keyboard assembly on the 300C but switches could be easily replaced. Both boards used an 8 bit Fairchild 1802 processor and I think had around 4kbyte of memory. Since it was "all digital," it had no speed knob but controlled using Control M keys followed by the numbers for the speed and finally, the return key. Its default speed/power up speed was very slow maybe around 10-12 wpm which was frustrating for the QRQ operator! When my dad picked up his second 300C, Info Tech used his call sign., W4BI for the serial number. :-)
The K4KN KEYBOARD
When i was trying to improve my cw speed in 1975, i managed to obtain a k4kn keyboard from a w5 fella - sure wish i could remember his call now - and so needless to say i was thrilled. imagine finding out that such great cw ops like k4kht, gene, and w9lrv(k5hib)ken were using the same morse keyboard to pound out the music as i now had. sure has been fun using the "woody", K4KN keyboard !
FROM K1LKP about the k4kn, "woody", keyboard creator: