Learning Morse At 64 Years Old -- My Experience With The Koch Method

Discussion in 'Straight Keys - CW Enthusiasts' started by G3EDM, Aug 31, 2021.

ad: L-HROutlet
ad: l-rl
ad: MessiPaoloni-1
ad: HRDLLC-2
ad: L-MFJ
ad: Left-3
ad: abrind-2
ad: Left-2
ad: L-Geochron
  1. G3EDM

    G3EDM Ham Member QRZ Page

    Warning: this is long, as you can see.

    I have just completed the Koch method for learning Morse code so here are some of my thoughts. Perhaps the most salient personal fact in this case is that I am 64 years old and only really started learning Morse seriously this year. (In various other times in my life, I tried to learn Morse but never got all that far.)

    Advantages of the Koch method:
    • A measurable performance target that enables each learner to progress at their own speed.
    • Like all the best Morse learning methods, strong incentives to avoid "counting the dits and dahs". In the case of Koch, it's partly because in most versions of his method, you jump right in at 15wpm to 20wpm. My favourite setting was to have the G4FON Koch-method software send the characters at 20wpm but slow down the inter-character and inter-word spacing ("Farnsworth spacing") so that the effective transmission speed was 15wpm.
    • A proven method whose successful performance is backed by data carefully compiled by Koch, rather than "when I learned Morse, I did it like this, so why don't you do it that way too" which is what I so often hear from people who usually learned Morse a long time ago. Some of them know very little about the Koch method. For a long time, the Koch method was almost unknown outside Germany, and the commenters in question were much younger and therefore would have found learning almost anything easier than if they tried to learn it now, as older adults, from scratch.
    The drawbacks of the Koch method, from my own experience:
    • As you progress through the 40 characters, each new character that is introduced gets more and more "diluted" and less frequent in each drill (because it is drowned in the middle of so many characters that you know already). For example, given that in a common version of the Koch method the first characters you learn are "K" and "M" this means you end up being *really* good at copying those letters, but less and less good at the letters that are introduced later. While this is also an issue with other Morse-learning methods, I do think that the Koch method is an extreme case of it. To give an example: it is not that unusual, when you already know, say, 25 characters, for a typical five-minute Koch drill not to send the "new" character (the 26th character in this case) until maybe one minute into the five-minute drill. Note: It is perfectly possible to partially correct this problem, for instance in G4FON's software, by setting up special drills that include only a few characters, thus lessening the "dilution" of the new character.
    • Even with software, "grading" your progress is a time-consuming task unless you are typing your copy into a keyboard; I used pencil and paper. You are supposed to stick with each set of drills that introduce a new character, over and over again, until you reach 90 percent accuracy, before moving on to the next new character. In the end, I stopped grading myself but simply moved ahead to the next character when I felt comfortable, instead of laboriously counting up my mistakes. This removed a serious "block" caused by reaching various plateaux in the learning sequence. Sometimes, 70 percent or 80 percent is "good enough" for learning purposes and you just have to push through regardless.
    • Finally, and perhaps most seriously, the rhythm of a typical Koch drill (consisting of a series of random characters grouped into random word lengths) is nothing like the rhythm in Morse of the English language (or any other language for that matter). A Koch drill does not use the standard distribution frequency of English, so for example you will have far more consonants and numerals, and far fewer vowels, than in a typical English sentence. For me the hardest transition was going from the "random character" drills to actual words and sentences in English, which typically feel more "rapid fire" (because of the greater number of short characters such as "E" or "T") than the Koch drills, even at the same WPM. Morse and Vail chose the shortest Morse sequences for the most common characters, after all. Sometimes I wish Morse code had fixed character lengths, like modern-day digital codes, but then it would be quite a lot less efficient!
    In the end though, the most important thing about learning Morse is practice, practice, practice. It may not matter so much what method you use, as long as you perservere. Also, by all means, if a particular method seems to be getting nowhere, try another one.

    Having said that, I am not convinced by the common claim that "everyone can learn Morse". I think it may be true that most people, and almost all young people, can master Morse code if they have the time and motivation. But claiming that everyone can do it, especially people like me who came to it in late middle age, sounds like an exaggeration, a kind of motivational rallying cry that may possibly not be true. But I admit that I have no evidence to back my statement, just as (in my experience) those claiming the opposite also seldom provide evidence.

    To be honest, it has been quite a struggle for me, even with extremely strong motivation. I was only interested in getting into ham radio if it was using Morse, so without Morse proficiency, getting on the air was basically impossible. It is still a big struggle, but most of the time it is fun, too. I do now rate the experience a "success" even though I still have an enormous way to go before I am truly fluent in Morse.

    73 de Martin, G3EDM
    Last edited: Aug 31, 2021
    US7IGN, N0TZU, KY4GD and 5 others like this.
  2. W1BR

    W1BR Premium Subscriber QRZ Page

    Oddly, the older I get the easier it seems to copy CW in my head. Not that my speed is increasing (not at all!!) but I am able to recognize some words without resorting to scribbling on a piece of paper and then reading what I wrote.
    DM2TT and G3EDM like this.
  3. KD2RDG

    KD2RDG Ham Member QRZ Page

    Respect for anyone with the self-discipline to get through Koch.

    Your station looks great on 40m, getting out all over Europe (RBN - Reverse Beacon Network) You're about to break into the really fun part, making contacts while your speed will increase naturally. Keep at it!
    G3EDM likes this.
  4. G3EDM

    G3EDM Ham Member QRZ Page

    The G4FON Morse software was improved in recent years with drills for the most common English words. The most common 100 words and the most common 1,000 words.

    There are two competing things at play. One thing that is good about the Koch method in general is that because the usual drills consist of random characters, there is no way you can guess what is coming next. This improves copy accuracy. OTOH according to what I have read in various books, when you get to really high speed (I guess that's 30wpm and higher) it is vital to copy entire words and even entire phrases in one go. For that, you have to do substantial "copying behind" by keeping whole phrases of code in that "memory buffer" in your head before decoding it.

    I find that if I can force myself to "copy behind" my copying skills improve dramatically. It is a bit of a knife-edge though, and hard to get exactly right!!!!

    Thanks. It is already a lot of fun. RBN is a bit misleading, as it relies on excellent receivers in very quiet locations. Having said that, it is cool to know that my 40m signal is truly reaching all of Western Europe, with just 12 watts and a simple dipole up only 12 feet! Now I just need to improve my Morse skills, and my operating skills (the two are intimately linked) to "convert" that long-reaching signal into actual contacts.

    73 de Martin, G3EDM
  5. G3EDM

    G3EDM Ham Member QRZ Page

    W1BR, W4KYR and KI4ZUQ like this.
  6. W9RAC

    W9RAC Platinum Subscriber Platinum Subscriber QRZ Page

    Martin after reading your post I have some thoughts that might help other aspiring Ops. wanting to become a Morse practitioner. I believe that listening to code on-air is essential while using the program, or any computer based code program. Once you listen on air, you will find all the conditions that are not present otherwise. All the crummy code both mechanical and paddle, QRM, QRN, QSB play a big part almost everyday. Decoding computer code and decoding mechanical key sent code is completely different most times. I believe exposure to both is necessary to become fluent with code at any speed. I like to see an Op. also mess with sending, trying their best to duplicate the code character they just listen to, off air of course for the student. Lots of guys say learning code is just like ridding a bicycle, you never forget it. I have not found that to be the case since skill will deteriorate quickly when left dormant. I'm on most every day and often run across not only new Ops but those who have mostly forgotten it from non use. Try to work some time in every day for some mental exercise. Nice job with that process sir, 73 Rich
    K9VA, DM2TT, KJ7VAB and 2 others like this.
  7. G3EDM

    G3EDM Ham Member QRZ Page

    That's fascinating. So, for those not familiar with the Koch method or with its leading advocates, their advice is: while you are learning Morse code, don't listen to on-air code, or any code at all except perfect code from code drills (in the old days those were on paper tapes, nowadays it is computer software).

    Two reasons are cited for this:
    • Avoid being contaminated by "bad code" from real-life QSOs.
    • The Koch method teaches the Morse characters in a fairly random order, and even when you are halfway through the course, you learned too few characters to copy anything off the air.
    To give you an example of this philosophy: the G4FON software actually has some really useful "canned QSOs" but you cannot even listen to them until you have learned all 40 characters (or alternatively, you can lie to the software and tell it you have learned all 40 characters, at which point the magic closet it opened....).

    I don't have strong opinions on this issue. I did monitor the bands even while doing the Koch course, and I tried to copy QSOs even though I had too few characters learned. What I did learn from personal experience is that Koch is "all or nothing": you really have to complete the entire course before you have any chance of copying real QSOs.

    73 de Martin, G3EDM
  8. W9RAC

    W9RAC Platinum Subscriber Platinum Subscriber QRZ Page

    Martin if you learn code by computer code only when you get on the air you will be lost. If you can copy mechanical key code you can copy computer code with ease. If you only intend to listen to computer sent code, or very good code by paddle you will get by. The first time you run across a straight key, bug or poor paddle you will be done. If you intend to get on the air bad code is a fact of life. 73 Rich
    DM2TT, WA1GXC and G3EDM like this.
  9. G3EDM

    G3EDM Ham Member QRZ Page

    I'm on the air, of course. As I said before, I am not taking sides on this issue. I agree with you that bad code is almost impossible to copy if you have been schooled in "perfect code". I suppose the argument is between those who think that before you are unleashed on the great heathen bad-coders, you yourself should have an immaculate education; and those who think that's bunkum and ignores the "real world". (This paragraph has been edited later to make more sense!)

    My goodness though, there is some really awful code out there. Including mine, some of the time, but let's not get into that!

    Rich: I am editing this to add that I am using a straight key and plan to go on doing so for some time, until I am reasonably satisfied that my straight-key sending is quite good. At that point I will probably transition to the Vibroplex bug that you can see in the photo posted earlier in this thread.

    My current key for QSOs is an Ameco K-4, a clone of the J-38.

    73 de Martin, G3EDM
    Last edited: Aug 31, 2021
    W9RAC likes this.
  10. AG6QR

    AG6QR Premium Subscriber QRZ Page

    I started learning code well into my 50s, and I remember transitioning from computer drills to listening on the air. I could immediately copy well-sent code from the airwaves, unless it was simply too fast for me. And I quickly increased my speed so that I could copy much of what I could hear. I had no difficulty copying code sent with a poorly handled bug (dits much faster than dahs), though sometimes the tendency of some operators to send too many dits, combined with my pre-existing difficulty in distinguishing S, H, 5 or D, B, 6, would confuse me a bit. Most straight key operators were pretty easy to copy. too.

    The only thing that presented a real problem was when operators refused to send spaces between letters. Absenceofspacesbetweenwordswasntabigproblem. Though code is certainly a lot easier to copy if there are spaces between words. But when the letters are run way too close together, nothing makes sense. That problem may have been made a little worse by learning using Farnsworth spacing, which uses longer than normal space between letters.

    Be that as it may, I'm not a big fan of intentionally starting out by listening to code where the letters are smushed together too close. One of my tutors told me to "Send the spaces. They're just as important as the dits and dahs."
    US7IGN and KD2RDG like this.

Share This Page

ad: MonitorSens-1