Getting from 5 to 20 WPM: what to do about plateaus, head copy, and whether to trust a decoder
- Replies 2
- Views 94
- Created
- Last Reply
Top Posters In This Topic
-
Admin_Terry 1 post
-
Admin_Lee 1 post
A better way to browse. Learn more.
A full-screen app on your home screen with push notifications, badges and more.
This thread is for comparing practice methods for the 5 to 20 WPM stretch, which is where a lot of people stall. The early gains come fast, then progress flattens somewhere in the low teens and practice starts to feel like work. Here is a rough map of the problem, with some questions to argue about.
Character speed versus overall speed. Most structured methods, whether Koch-style or Farnsworth-spaced, send characters at 18 to 20 WPM or faster and stretch the gaps between characters and words. The goal is to learn each letter as a sound pattern ("dah-di-dit-dit" is wrong, it should simply be B) rather than a dot-and-dash count. If you learn characters at 5 WPM with normal spacing, the individual letters are so slow that you tend to count elements, and that habit is hard to unlearn later.
The usual plateau. The wall around 10 to 13 WPM is commonly attributed to a switch from letter-by-letter copy to word-level recognition. At that speed a letter is over so quickly that you can't consciously identify it and still keep up, but you haven't yet learned to hear whole words and common chunks like THE, ING, AND, or CQ. Typical ways through it:
Head copy. Writing every letter is a fine training stage, but it caps speed because your pencil becomes the bottleneck, and many people write more slowly than they can hear. Moving to head copy is usually gradual: start with callsigns and signal reports in your head, then short words, then whole phrases, then full exchanges. Letting go of the pencil is uncomfortable, so it helps to do it in low-stakes practice before attempting it on the air.
Drills. Word lists and random callsigns train different skills. Callsign trainers (RufzXP is a well-known one) build quick, accurate recognition of mixed letter-number strings, which is exactly what you need in contests and pileups. Plain-text and word drills build the chunking that makes ragchew-speed copy possible. Ideally you do both, but which should come first?
Paddles and keyer settings. If you send with a keyer, a few settings matter more than people expect:
Decoding by ear versus software. Software decoders can be handy for checking what you sent or confirming a callsign. As a learning tool for copy, though, they have a trap: they're weakest in exactly the conditions you need to learn, such as noise, QSB, hand-sent irregular timing, and overlapping signals. Relying on a screen also keeps your ear from developing the tolerance it needs. Generators and trainers (LCWO, Morse Elmer, and similar) are different, since they produce clean practice rather than decode live signals.
Schedules. Short daily sessions generally beat occasional long ones. Around 15 to 30 minutes a day, perhaps split into two blocks, is a common recommendation, with fatigue as the signal to stop. Within that you might rotate: new characters or speed push, callsign drills, then some real or simulated on-air material.
On-air practice. Slow-speed CW nets and QSO partners with a similar speed are useful for practicing the full exchange structure (calls, RST, name, QTH, 73) under realistic conditions. Contests are a good source of fast, repetitive exchanges where you hear the same format over and over. If you try one, check the current schedule and rules of the specific event yourself, since they change. Many operators suggest listening first, then answering only stations you can copy comfortably.
Questions to get things going: What changed things for you when you hit the low-teens wall: speed, spacing, or the kind of material? Do you move to head copy early or late? And is there a point where software helps instead of hurts?
Link to comment
https://www.hamradiobase.com/forums/topic/10965-getting-from-5-to-20-wpm-what-to-do-about-plateaus-head-copy-and-whether-to-trust-a-decoder/Share on other sites