Talk:Continuous wave
Untitled
[edit]This creates an interesting situation: CW is a redirect for Morse code. And yet, there is certainly a distinction between Morse code, which can be transmitted by means other than CW, and "Continuous wave". But, in the real world, Morse code has only been used, to any significant extent, by way of CW for many years. So the two have really come to be somewhat synonymous. I'd be inclined to merge the two under Morse code with an explanation of the problem. Most of what I have to say about CW, I'd be inclined to say in the Morse code article, probably because that's where most of it already is. Bill 11:56, 8 Aug 2003 (UTC)
Well, CW was a redirect for Morse until a few minutes ago. Bill 11:58, 8 Aug 2003 (UTC)
It would be better instead to link other uses of continuous wave to support the distinction. Waveguy 01:40, 10 Aug 2003 (UTC)
The only use of CW I know of is to send Morse code by way of radio. Historically, Morse code has been used in media other than CW, but this is addressed in the Morse code article. Unlike most forms of radio transmission, CW in unmodulated. That is, it is a pure carrier in which no intelligence has been encoded. -- Bill 12:36, 10 Aug 2003 (UTC)
- This has been here for years but it's not correct; on-off keying is a form of modulation by definition! And, pedantically, you might say ICW for "Interrupted continuous wave" but anyone who uses the term "CW" understands that; a continous wave is a mathematical fiction, anyway. --Wtshymanski 02:09, 9 May 2005 (UTC)
Feld-Hellschreiber is also a CW mode. It uses (in its basic form) on off keying to represent characters, though anti-aliasing can be done by changing the level of the carrier... Jdos2 18:58, 15 August 2006 (UTC)
This article should point out that the phrase "continuous wave" and its abbreviation "CW", are mainly used in the context of amateur radio as a misnomer for Morse code. Originally the phrase "continuous wave" was used to mean "not spark", i.e. not periodically damped. Now the original meaning of the phrase has been lost, and "continuous wave" has come to mean Morse code. In my opinion, any modern amateur radio mode could be called "continuous wave", because the wave is not periodically damped. Again in my opinion, the leading paragraph of the article, the one that defines continuous wave as an electromagnetic wave of continuous amplitude and frequency that is assumed to be switched, is merely attempting to cover up the misnomer. Unfortunately I have no sources to back me up at this time. Apparently neither does the article's original author. Rclocher3 22:51, 13 April 2007 (UTC)
- This isn't the Académie française, Wikipedia doesn't dictate what *should* be the usage, just records the way terms are actually used. When talking about "CW", radio people mean "ICW" sending Morse (or other forms of radio teletype) - and laser people mean a laser that runs as long as it's switched on. Context is critical in English, which is why machine translations so often fail. I think the article is reasonably honest in the way it uses the terms. --Wtshymanski 01:27, 14 April 2007 (UTC)
In the early days, transmission was by Spark, which could be directly received on a simple AM receiver as a buzzing sound. Later when Continuous Wave (CW) arrived, a Beat Frequency Oscillator (BFO) was needed to make the CW audible. In many early sets (particularly Military) the BFO knob was simply labelled CW. Hence today's confusion.119.18.11.19 (talk) 10:40, 2 April 2014 (UTC)
WikiProject class rating
[edit]This article was automatically assessed because at least one WikiProject had rated the article as start, and the rating on other projects was brought up to start class. BetacommandBot 09:46, 10 November 2007 (UTC)
Large bandwidth explanation
[edit]The article states "if the carrier wave is turned on or off abruptly, the bandwidth will be large" but there is no explanation why this is the case. There is a formula to calculate the amount of bandwidth but no reason given as to why this act should increase the bandwidth at all. Would it be possible to add a reason for this statement? -- Malvineous (talk) 06:38, 7 May 2012 (UTC)
- Just how deep do we want to get in this article? We could point at bandwidth and channel capacity and the Shannon–Hartley theorem but these go considerably beyond the scope of keying a carrier on and off. --Wtshymanski (talk) 13:21, 7 May 2012 (UTC)
- I'm not sure you need to invoke the Shannon-Hartley theorem here. This is analog bandwidth, and the relation between bandwidth and the sharpness of the on-off transitions comes directly from the Fourier transform. A signal with sharp transitions has a Fourier spectrum that spans a wider range of frequencies than a similar signal that has more gradual transitions.--Srleffler (talk) 03:23, 8 May 2012 (UTC)
- I'm all for a simpler explanation. By all means link to pages that explain the detail, but I think a short explanation would be helpful to quickly summarise the reason without having to slog through formulae. I'm thinking something like "...the bandwidth will be large due to the initial spike, which momentarily appears as a different frequency." This is probably inaccurate so hopefully someone else can correct it and place it on the page. It does however make me wonder why this is the case. Presumably if the carrier was switched on at the zero crossover point, you'd never get a partial waveform and the bandwidth would be no larger than during transmission? -- Malvineous (talk) 11:22, 12 May 2012 (UTC)
- We could do it that way. Something like It can be shown (by the Fourier transform) that any abrupt change in a signal produces high-frequency components. The appearance of these components during keying of a carrier momentarily increases the occupied bandwidth of the signal. These cannot be entirely eliminated but careful transmitter design limits the effect. or something like that. --Wtshymanski (talk) 17:02, 12 May 2012 (UTC)
- As concrete examples, we could produce a time-graph of a sine wave truncated at a non-zero point to show the sharp transition that would occur, then maybe another one showing it decaying smoothly to zero to show click suppression. This could make talk of fourier more understandable. --Nigelj (talk) 18:02, 12 May 2012 (UTC)
- I think both these suggestions would make the concept a lot easier to understand. You've got my vote! Thanks. -- Malvineous (talk) 23:31, 12 May 2012 (UTC)
- As concrete examples, we could produce a time-graph of a sine wave truncated at a non-zero point to show the sharp transition that would occur, then maybe another one showing it decaying smoothly to zero to show click suppression. This could make talk of fourier more understandable. --Nigelj (talk) 18:02, 12 May 2012 (UTC)
- I think the explanation should be simple, or we should not provide one here. We should not get off-topic with a detailed or complicated explanation. I am not sure whether this can be explained simply and correctly. Malvineous's explanation is not correct, and it doesn't matter if you switch on at a zero crossover point. A fast turn-on produces a large bandwidth regardless.--Srleffler (talk) 02:18, 13 May 2012 (UTC)
- But can you provide a simple explanation why? Saying it produces a large bandwidth "regardless" doesn't really make it any clearer. I still can't glean from the article the difference between a "fast turn-on" and a full cycle out of many at the carrier frequency. To me, a 1MHz signal has a million "fast turn-ons" every second, so I don't understand why the first one is any different to the other 999,999. -- Malvineous (talk) 08:33, 13 May 2012 (UTC)
- Technically, it's to do with the sharpness of any discontinuity in the waveform. I was wrong above to emphasise 'a sine wave truncated at a non-zero point' without explaining better. Fourier analysis is about finding the series of sine waves that, if added together, would create the waveform we are looking at. If the waveform has a sharp, angular feature, then it would take a lot of higher-frequency partial sine-waves added up correctly to (re-)create it. The less sine-wave-like the feature, the more other frequencies will be necessary. The almost-right-angle corner where a zero-crossing point instantly becomes flat would instantaneously require infinite bandwidth. A truncated partial-rise suddenly dropped to zero would also require infinite bandwidth, but with even more energy in many of the higher frequencies. Truncating a sine-wave power graph that is already dropping to zero would put less energy into the extreme frequencies. Now, no one truncates a 1 MHz carrier with a sharp corner, or the resulting 'click' would occupy a bandwidth that is very large compared to 1 MHz. In fact, the clicks that irritate other users on a CW band are probably no wider than a few hundred hertz, up to perhaps 1 kHz in extreme cases. The real time-domain graph of this happening would look like the 1 MHz transmitted wave decaying gently down to zero over about 1 ms, or 1,000 cycles, or more. If the click bandwidth was reduced to ± 10 Hz, then the decay would have to take about 0.1 s, or 100,000 cycles of the carrier wave - each a little smaller than the last. (I say ± because this is equivalent to a double-side-band modulation and the same click information is repeated above and below the carrier in the frequency domain.) Thinking this through, I don't think the graphs I proposed above would be much use - the realistic ones would just look like a tapering grey blur, and any showing a sine wave with an angle in it would actually be quite unhelpful. (P.S. some people may be muddling this zero-crossing business with the way mains power is sometimes switched to avoid contact burn in large power installations. This is a completely different application, and yes, those large contactors (switching 50 or 60 Hz) do produce considerable HF noise that needs to be further suppressed.) --Nigelj (talk) 09:06, 13 May 2012 (UTC)
- Thank you so very much for this last paragraph, Nigelj! I was frustrated about WHY this happened until I read your breakdown of the wave transforming with higher frequency to "ramp up" to the prescribed frequency. This totally blew my mind and I appreciate it very much. As someone who came to this page to get a handle on how exactly a radio creates and fools with waveforms to transmit information (i.e. a new ham), this was perfect. If there were some way to put this simplified explanation (or something very similar) into the part about spark-gap RF interference, the knowledge-gap (haha) would be filled. Again, thanks a million for your unpretentious and simple answer to a rather complex question. 66.91.204.11 (talk) 13:36, 16 December 2012 (UTC)
- Thank you for your kind comments. I would be happy to put a tidied up version of the above into article-space (hopefully with refs too, if I can find them). The trouble is that I don't think the best place is this article on CW communication. I don't really know where it should be, as I don't know of many other articles in this area. there's Fourier analysis - linked above - but that's a rather abstract and mathematical place for these practical points. Equally, this is rather technical for this more general piece. Suggestions welcome. --Nigelj (talk) 14:18, 16 December 2012 (UTC)
- Thank you so very much for this last paragraph, Nigelj! I was frustrated about WHY this happened until I read your breakdown of the wave transforming with higher frequency to "ramp up" to the prescribed frequency. This totally blew my mind and I appreciate it very much. As someone who came to this page to get a handle on how exactly a radio creates and fools with waveforms to transmit information (i.e. a new ham), this was perfect. If there were some way to put this simplified explanation (or something very similar) into the part about spark-gap RF interference, the knowledge-gap (haha) would be filled. Again, thanks a million for your unpretentious and simple answer to a rather complex question. 66.91.204.11 (talk) 13:36, 16 December 2012 (UTC)
- Technically, it's to do with the sharpness of any discontinuity in the waveform. I was wrong above to emphasise 'a sine wave truncated at a non-zero point' without explaining better. Fourier analysis is about finding the series of sine waves that, if added together, would create the waveform we are looking at. If the waveform has a sharp, angular feature, then it would take a lot of higher-frequency partial sine-waves added up correctly to (re-)create it. The less sine-wave-like the feature, the more other frequencies will be necessary. The almost-right-angle corner where a zero-crossing point instantly becomes flat would instantaneously require infinite bandwidth. A truncated partial-rise suddenly dropped to zero would also require infinite bandwidth, but with even more energy in many of the higher frequencies. Truncating a sine-wave power graph that is already dropping to zero would put less energy into the extreme frequencies. Now, no one truncates a 1 MHz carrier with a sharp corner, or the resulting 'click' would occupy a bandwidth that is very large compared to 1 MHz. In fact, the clicks that irritate other users on a CW band are probably no wider than a few hundred hertz, up to perhaps 1 kHz in extreme cases. The real time-domain graph of this happening would look like the 1 MHz transmitted wave decaying gently down to zero over about 1 ms, or 1,000 cycles, or more. If the click bandwidth was reduced to ± 10 Hz, then the decay would have to take about 0.1 s, or 100,000 cycles of the carrier wave - each a little smaller than the last. (I say ± because this is equivalent to a double-side-band modulation and the same click information is repeated above and below the carrier in the frequency domain.) Thinking this through, I don't think the graphs I proposed above would be much use - the realistic ones would just look like a tapering grey blur, and any showing a sine wave with an angle in it would actually be quite unhelpful. (P.S. some people may be muddling this zero-crossing business with the way mains power is sometimes switched to avoid contact burn in large power installations. This is a completely different application, and yes, those large contactors (switching 50 or 60 Hz) do produce considerable HF noise that needs to be further suppressed.) --Nigelj (talk) 09:06, 13 May 2012 (UTC)
- But can you provide a simple explanation why? Saying it produces a large bandwidth "regardless" doesn't really make it any clearer. I still can't glean from the article the difference between a "fast turn-on" and a full cycle out of many at the carrier frequency. To me, a 1MHz signal has a million "fast turn-ons" every second, so I don't understand why the first one is any different to the other 999,999. -- Malvineous (talk) 08:33, 13 May 2012 (UTC)
- We could do it that way. Something like It can be shown (by the Fourier transform) that any abrupt change in a signal produces high-frequency components. The appearance of these components during keying of a carrier momentarily increases the occupied bandwidth of the signal. These cannot be entirely eliminated but careful transmitter design limits the effect. or something like that. --Wtshymanski (talk) 17:02, 12 May 2012 (UTC)
- I'm all for a simpler explanation. By all means link to pages that explain the detail, but I think a short explanation would be helpful to quickly summarise the reason without having to slog through formulae. I'm thinking something like "...the bandwidth will be large due to the initial spike, which momentarily appears as a different frequency." This is probably inaccurate so hopefully someone else can correct it and place it on the page. It does however make me wonder why this is the case. Presumably if the carrier was switched on at the zero crossover point, you'd never get a partial waveform and the bandwidth would be no larger than during transmission? -- Malvineous (talk) 11:22, 12 May 2012 (UTC)
- I'm not sure you need to invoke the Shannon-Hartley theorem here. This is analog bandwidth, and the relation between bandwidth and the sharpness of the on-off transitions comes directly from the Fourier transform. A signal with sharp transitions has a Fourier spectrum that spans a wider range of frequencies than a similar signal that has more gradual transitions.--Srleffler (talk) 03:23, 8 May 2012 (UTC)
Watch this http://www.youtube.com/watch?v=Z5rKTagEsro
Notice how for the first few cycles of the driving pendulum, ALL the other pendulums start swinging the same small amount. Energy from the initial swing is transferred to every other pendulum regardless of its resonant frequency. It is only after several swings that the different resonant lengths become significant - only the pendulum with the same frequency as the driver swings in step, each swing of the driver reinforce the next swing of its partner - the other pendulums don't build up be cause they are not in step, but neither are they completely stationary.. This is frequency selection - it isn't a 100% like people think. Rather than the other pendulums NOT picking up any movement, resonance means the resonant one builds up. And this takes several swings. — Preceding unsigned comment added by 220.244.93.215 (talk) 10:43, 18 July 2013 (UTC)
Start a discussion about improving the Continuous wave page
Talk pages are where people discuss how to make content on Wikipedia the best that it can be. You can use this page to start a discussion with others about how to improve the "Continuous wave" page.