Morse Code Practice Trainer

Almost everyone meets Morse code as a chart: a letter, then a row of dots and dashes beside it. The chart is accurate and it is the wrong place to start. The code was never a set of pictures. It is a set of durations, and the people who use it well do not see anything at all when they copy. They hear a short rhythm and a letter arrives.

That difference between looking up and recognizing is the whole subject. It explains how the code is defined, why its speed is measured with the name of a city, why the best known training method begins at full speed with only two characters, and why a learner who memorizes the chart first so often stalls.

From an American patent fight to a European standard

Samuel Morse was a painter before he was an inventor. By the account of the United States Senate Historical Office, he gave his first demonstration of a working telegraph in September 1837 with Leonard Gale, a chemistry professor, and then turned to a former student and toolmaker, Alfred Vail, to refine and produce the instruments. In demonstrations that followed he introduced the code of dashes and dots that carries his name.

The famous moment came later. On May 24, 1844, Morse sent a message over wire from the Capitol in Washington to Baltimore, where Vail was waiting. The International Telecommunication Union dates the code’s working life from the same year: its current Recommendation notes that versions of the Morse code have been in use since 1844.

Who shaped the alphabet itself is less tidy. Wikipedia’s account, which is the only source used here for this part of the story, says Morse’s first proposal was replaced by an alphabet code developed by Vail, and that this version served commercial telegraphy in North America. It was an awkward code by modern ears, and Europe did not keep it. In 1848 Friedrich Clemens Gerke, working in Hamburg, simplified it so that only two elements remained, the dot and the dash. The same account says his version was adopted in Germany and Austria in 1851 and led to the International Morse code in 1865, which kept most of his letters.

The year 1865 matters for a second reason. Telegraph lines by then crossed borders, and messages had to be stopped and re-sent in the system of the next country. Representatives of twenty states met in Paris, and on 17 May 1865 they signed the first International Telegraph Convention and founded the body that is now the ITU.

Four rules and a word

The modern definition fits on a few pages. Recommendation ITU-R M.1677-1 lists the letters, one accented letter, ten figures, punctuation and a handful of procedural signals such as wait and end of work. Then it gives the timing in four sentences. A dash is equal to three dots. The space between the signals forming the same letter is equal to one dot. The space between two letters is equal to three dots. The space between two words is equal to seven dots.

Everything is therefore a multiple of one length. Set the dot and the whole transmission follows. That is tidy, but it leaves a practical problem: letters have different lengths, so a speed in letters per minute means little. The letter E is one dot long and the figure zero is nineteen.

The convention is to pick a standard word, and the usual choice is PARIS. A 1957 manual of the United States Army and Air Force on teaching the code calls it the PARIS standard and used it to fix the speed of practice and test tapes. Count it against the four rules: P is 11 units, A is 5, R is 7, I is 3 and S is 5, which is 31. Four gaps between letters add 12 and the gap after the word adds 7. The total is 50 dots.

So a speed of s words per minute means 50 times s dots every minute, and one dot lasts 1.2 divided by s seconds. At 20 words per minute a dot is 60 ms, a dash 180 ms, and the word PARIS takes exactly three seconds. Wikipedia notes that another standard word, CODEX, is sometimes used for random groups, so a quoted speed is ambiguous unless the word is named.

Why slow code teaches the wrong skill

Here is the trap. At five words per minute a dot lasts nearly a quarter of a second, and a beginner has time to count: short, long, short, short, that must be L. It works, and it builds exactly the habit that later has to be torn out.

The German psychologist Ludwig Koch studied that habit in the 1930s. A summary of his 1936 report prepared by the Long Island CW Club says he was sharply critical of teaching from a chart of symbols, which he called the analytical method. Copying that way means counting the elements of each signal and reassembling them for comparison with a symbol held in memory. The summary calls this inefficient but possible at a low tempo and unachievable at faster ones, and it quotes Koch placing the changeover in mental behavior at about 10 words per minute.

The Army manual states the cost in plain terms. A student who stops to count a character he did not recognize loses the several characters that follow, so he is told to leave a blank and keep copying, and to hear each signal as a complete unit of sound. It warns that the temptation to count dits and dahs may take some time to resist.

His own experiments pointed the other way. The same summary reports that he found about 12 words per minute best for first learning, and that below 10 the student could not hear a character as a single acoustic shape. Koch called that wholeness the Gestalt of the character.

His method follows directly. Send at full speed from the first minute, but with only two characters that sound clearly different. When they are copied with about 90 percent accuracy, add a third, then a fourth, and go on one character at a time until the whole set is known. Nothing is ever learned slowly and relearned fast, and the summary’s list of his rules includes a flat one: visual charts should never be used.

Which two characters come first is less settled than the method. The same club’s analysis gives the sequence Koch himself used as L F C K R D P X Y Q B G A N Z U V W J H S I E O M T. It observes that the order leans on mirror pairs such as L and F, which look alike on paper but sound nothing alike. Popular trainers use other orders. The principle survives any of them.

Fast characters, slow words

Full speed is hard on a beginner who still needs a moment to write each letter down. The Farnsworth technique gives that moment back without slowing the sound. Characters are sent at one speed, and the silences between characters and words are stretched until the overall rate is lower. Wikipedia’s example sends each character at 13 words per minute and lengthens the gaps until the word rate is only 5.

The Army manual built all its receiving practice the same way, on characters sent at 20 groups per minute: slow enough to be heard as one unit of sound, fast enough that a character does not fall apart. The station at the headquarters of the American Radio Relay League, W1AW, sends its slow practice text at 5 to 15 words per minute and its Morse bulletins at 18.

The arithmetic is short, and the Morse Code World site works it through in full, noting that its result is the same as the equations Jon Bloom published in an article for the league. Wikipedia knows that standard as ARRL Farnsworth spacing. A 50-unit word at overall speed s must take 60 divided by s seconds. Its 31 units of marks and inner gaps, sent at character speed c, take 37.2 divided by c seconds. What is left over is shared between four character gaps and one word gap in the proportion three to seven, nineteen parts in all.

Try 20 words per minute spaced out to 12. The word takes 5 seconds and its marks take 1.86, leaving 3.14 seconds of silence. Three nineteenths of that is 496 ms between characters, where standard timing would give 180. Seven nineteenths is 1,157 ms between words. The dots and dashes are untouched at 60 and 180 ms, so the ear still learns the real shapes.

Sending, and what the code is still for

Receiving is half the skill. A straight key is a switch: the operator makes every dot, dash and gap by hand, and the proportions are as good as the wrist. Paddles hand that job to an electronic keyer, which produces perfectly timed elements. Hold one paddle for a string of dots, the other for dashes, and squeeze both for an alternating stream that starts with whichever was closed first.

Keyers come in two dialects. As Chuck Olson explains, a Mode A keyer completes the element being sent when the paddles are released, while a Mode B keyer adds one more element opposite to it. His test is the letter C. Squeezed dash first, Mode A needs the paddles held until the last dot sounds. Mode B wants them released during the second dash, and the keyer supplies the final dot. Neither is better; an operator keeps whichever one their hands learned first.

Operators also save time with agreed shorthand. A companion ITU Recommendation lists the Q code, three-letter groups that work as a question or an answer: QRS asks the other station to send more slowly, and QRQ to send faster.

The code has lost its legal standing in amateur radio without losing its users. A 2003 world conference let each country decide whether to require it, and in December 2006 the Federal Communications Commission removed the five words per minute test from American licences. The ITU still records Morse in use in the amateur services, and to a diminishing extent elsewhere.

It also keeps a quiet job in aviation. The FAA’s manual for pilots says radio beacons transmit a continuous three-letter identification in code, and that the only positive way to identify a VOR station is by its Morse identification or a recorded voice. Morse has been used, too, as an assistive technology that lets people with a range of disabilities communicate.

None of this makes learning by sound a guarantee. Koch’s optimum of 12 words per minute came from one study, a 1943 review quoted in the club’s summary judged the gain from one of his refinements probably not significant, and the club says his findings still have to be reconciled with modern Farnsworth practice. A lamp, for what it is worth, is no substitute for a tone: to stay under three flashes a second, the limit web accessibility guidance sets for flashing content, a run of dots cannot be shown faster than about 7 WPM. The chart still has a use. It is where you check a character you have forgotten, and then you close it and listen again.