What are some characteristics, when sending CW, of a "bad fist"? How bad can it get, typically, and still manage to communicate the message text sent to a human operator on the receiving end?

Contra-wise, what (slight?) variations from machine generated Morse Code are still considered as having a "good fist" (for a human operator)?

Or does machine generated Morse code have some characteristics of a bad fist as well?

  • $\begingroup$ As this question stands, it runs the risk of being closed as primarily opinion-based because the answers are likely to be more about opinion than fact. Can you rephrase the question in such a way that it is more inviting to authoritative answers than opinion-based ones? $\endgroup$
    – user
    Apr 23, 2014 at 14:00
  • $\begingroup$ The term seems to be in common use. It must have some definition including some characteristics that can be perceived during a QSO. $\endgroup$
    – hotpaw2
    Apr 23, 2014 at 14:21
  • $\begingroup$ Unfortunately, I don't think it follows from "the term seems to be in common use" that "it must have some definition including some characteristics". I'd liken Morse code with photography in this case; it's possible to objectively determine whether an object in a photo is in focus (think standard deviation in key-down lengths and spacing), but it's a lot harder to objectively define what makes a photo good (a "good fist"). $\endgroup$
    – user
    Apr 23, 2014 at 14:25
  • $\begingroup$ Focus can be measured in camera lens testing. A "good photo" can have bad focus, and a bad photo can be optimally focused. Is focus really the issue or not? Same with your hypothesized std.dev.? Is that your answer? $\endgroup$
    – hotpaw2
    Apr 23, 2014 at 14:44
  • 1
    $\begingroup$ @hotpaw2 the answer is that characteristics of a "bad fist" are that "someone doesn't like it". Isn't it clear that this is a subjective measure, just like "bad music", "bad photos", "bad food", or "bad sense of humor"? What do you expect to happen here, except for a bunch of primarily opinion-based, arbitrary, subjective descriptions of people's preferences? $\endgroup$ Apr 23, 2014 at 17:59

3 Answers 3


Some characteristics that make code harder to copy:

  • The dahs are are too short compared to the dits.
  • Not enough space between words.
  • Not enough space between characters.
  • An out-of-place pause when sending a word (as when thinking of how to spell a word).
  • Incorrect number of dits (for example, sending b or n instead of d)
  • Sending a correction with insufficient pauses between the error, the error indicator (eight dots, a question mark, whatever) and the correction.

It's been my experience that most hard-to-copy code has timing errors, and the most damaging of those errors are of spacing.

  • $\begingroup$ I would add 'unintentionally varying the sending speed' to that list, too. Sometimes during a contest you will hear software-generated CW that has a hugely speeded-up '599' followed by a slow serial number, and that's just a time-saving measure. But when someone is sending on a straight key by hand and the speed varies, it can make it difficult for the receiver - especially if the reader is done in software, rather than being manually read. $\endgroup$
    – Scott Earle
    Sep 5, 2014 at 3:54

The worst offenders earning "bad fist" have one or more of three major problems (in my opinion and I admit some of this is repeat of above):

  1. Poor spacing of dahs, that is too short mostly or changing the length which used to be a popular "swing" to a fist back in the 50s/60s. Such as a CQ with a very long dah in the first part of the letter C with a shorter dah later as in daaaah-di-dah-dit.

  2. run-on words, that is, no spacing between.

  3. Chirp. I can't stand chirp. Not a fist problem but I say get that rig fixed. Not as common today with solid-state radios as this used to often be caused by poor voltage regulation in the older tube radios but I still hate it and won't answer a call with chirp.

  • $\begingroup$ "[I] won't answer a call with chirp" Somewhat unrelated, but how does that help the operator know what's wrong? You say "get that rig fixed"; what if the operator is working on it, doesn't have every bit of test equipment available and wants a signal report to know whether they are headed in the right direction? $\endgroup$
    – user
    Sep 11, 2014 at 7:36

A bad fist in CW usually refers to improper keying, such as slurring the definition between dots and dashes, so it is harder to copy. Also, varying speed of CW dramatically and accidentally adding in characters that shouldn't be there are examples of a bad fist. Usually machine generated CW is good quality, since it is all programmed, but errors in software or components can result in sloppy machine generated CW.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .