Talk:Crystal radio

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search


I object to the use of the perjorative "Naive" in regards to the simple circuit presented. It implies a preference or judgement that has no place in an encyclopedic article.

For example, getting 8-year-olds to build a working radio in less than an hour requires some simplification, and the the circuit presented does that. Perfect for the application.

Thbusch (talk) 16:53, 27 August 2009 (UTC)

I agree, it does imply judgment, but there is nothing wrong with sound engineering judgment, it is something we need more of.
... and perfect? Actually, that circuit does come pretty close to naive by adding a part (variable capacitor) that does more harm than good, it reduces the tuning range over the simpler slide tuner circuit. Many people add the variable capacitor because they are unaware that most inverted L antennas already have the needed capacitance. I am very sympathetic to your motive of making this article assessable to young people, but eight-year-olds are not going to understand how a variable capacitor works anyway, a simpler, slider tuner radio would be easier to build and understand.
That circuit is just plain hard to justify from any angle. John (talk) 03:04, 31 August 2009 (UTC)

Broken Links[edit]

2 of the first 3 links are broken, 1 to the 'enthusiasts', the other to using modern components for 'hi' performance.
Just after I added some text to the cites too! ADVICE always check a link works before trying to 'improve'it!
Dug out a few links that may be relevant. Some are obviously commercial sites, for those who are interested.

-- (talk) 09:28, 6 December 2009 (UTC)

Expanded 'How it works' section[edit]

Added subsections under 'How it works' describing the different parts of a crystal set, addressing some of the topics discussed on this page, such as impedance matching and selectivity. --ChetvornoTALK 06:40, 12 July 2010 (UTC)

This revision has some problems. It adds a capacitor across the coil, which will not work to tune the BCB, and is not the simple circuit, it does not address the antenna as the capacitor that tunes the inductor. We need to restore it to the previous version. John (talk) 06:22, 18 August 2010 (UTC)
Which part are you referring to specifically? --ChetvornoTALK 07:36, 18 August 2010 (UTC)
Much of the added subsections under "How it Works". This material is assuming a variable capacitor can tune the BCB with a coil, and it is not practical. We don't want to mislead people with this. See "The Naive Circuit" section. Electrically short antennas are capacitive and coil tunes the antenna in a single pole circuit, a variable capacitor cannot be used in this circuit. John (talk) 18:51, 18 August 2010 (UTC)
All of the diagrams in this section were sourced from actual circuits. Keep in mind that this is a "How it works" section, for nontechnical people. It's important to emphasize for them that a capacitor is always necessary in a tuned circuit, whether provided by the capacitance of the antenna or a descrete component. So the text mentioned that some circuits use the capacitance of the antenna alone, but most of the circuits I chose used tuning capacitors. And you're certainly mistaken if you think that practical BCB crystal sets didn't use tuning capacitors. There are literally hundreds of circuits of crystal sets that were tuned capacitively. See here I didn't have time to check how they solved the tuning range problem, but it looks like many use tapped coils and cover the band in separate sections. I think it's important to have a historical POV and show a variety of the circuits that were actually used. Your point about tuning range is important, and deserves a section, but I think removing all circuits with tuning capacitors is going too far. --ChetvornoTALK 02:38, 19 August 2010 (UTC)
Sets do use tuning capacitors, but it is not physically possible for practical sets to use them the way this diagram shows. Does what I said above suggest removing all sets with tuning capacitors? I do not. That diagram shows a variable capacitor and an antenna, and is not correct. I am very familiar with that site, and I am extremely familiar with how this problem is resolved: by not using this circuit. It is a common mistake and this article needs to avoid it. Please become familiar with this issue before editing this. I will be happy to discuss it with you. BTW, I am not sure hundreds of circuits means something works. Please do not put this figure back. I would like to make the "how it works" section correct. John (talk) 02:56, 20 August 2010 (UTC)

This edit war on the article needs to stop. Discuss the matter here with regard to Wikipedia's policies and guidelines, then change the article in accordance with compromise and consensus. Discuss the reliability of the sources and the representation of the cited sources in the article text and diagrams, not each other's intelligence, abilities or personal views. Having said that, here is my input: First, the scan of the 1950s article by L.B. Robbins looks authentic to me. Does anyone have any evidence that it is a modern con with technical errors introduced by someone else? If not then it seems to be a WP:RS reliable source and may be used to support the article. Second, I have made that circuit and so I know it works. This is WP:OR and so cannot be used in the article, but should be enough to stop the bickering. (I was trying to re-create a crystal radio made by my grandfather during WW2 and so wound a coil like his on a wooden former, but used a schottky diode as a detector and an audio amplifier instead of Hi-Z headphones. My grandfather's design was capacitor-less with a sliding earth-tap on the coil, connected by sandpapering away some insulation lacquer after winding it. Adding a small trimmer capacitor increased the selectivity and 'peaked' the reception slightly, but still only the most local medium-wave station was really receivable). So, the circuit is authentic, and it works. What's the problem with mentioning it in this article? --Nigelj (talk) 08:29, 20 August 2010 (UTC)

P.S. A completely different inductance is needed with the capacitor compared to without it, for the same long wire antenna and same reception band. I can dig out my old experiments and count the turns if anyone's interested, but it's largely irrelevant as it's WP:OR, however it shows the principle is sound. --Nigelj (talk) 08:37, 20 August 2010 (UTC)

Right you are, Nigelj, no more EWing. The argument is not over whether the circuit works; I think John would agree it does. The problem is that the circuit is impractical for an actual radio because it can't tune the entire broadcast band. John thinks that for this reason it shouldn't be included in the article, I think it should. John, I understand the issue. The AM broadcast band requires a tuning range of 3.09:1. So the tuning capacitance requires a range of the square of this: 9.5:1. Since the minimum capacitance is the antenna capacitance, the tuning capacitor must be at least 9.5 times the antenna capacitance. For a 15 meter antenna, which has a capacitance of about 135pF that's about 1300pF. It's difficult to find air variables that large, so with a standard capacitor the radio won't tune the whole band. So what? These are technical details. The circuit does in fact work. Many people build crystal radios just for the educational value and don't care whether it covers the entire band.
More importantly, the purpose of the schematic John deleted is not to present a practical radio, but to illustrate for novices the operation of a crystal radio with the simplest, most easy-to-understand circuit. Since the tuning capacitor is mentioned in the text, it has to include a tuning capacitor. The circuit I used is the one most commonly employed in the literature to introduce crystal radios, and is very widely seen: 1, 2, 3, 4, 5, 6. It says in WP:DUE that "due weight" must be given to different aspects of a subject "in proportion to their representation in reliable sources". Considering the wide appearance of this circuit in crystal radio literature, I think it probably should be included.
I have no objection to mentioning the drawbacks of this circuit in the article, in fact I insist on it. But I don't think there is a better circuit for illustrating how crystal radios work. John, I have given 6 references above, plus the 2 in the article, on the wide use of this circuit. Where are your references backing up your claims that it is unusable? --ChetvornoTALK 10:08, 20 August 2010 (UTC)
Tuning across the whole MW band is irrelevant. The poor overall Q-factor due to the inefficiency of a (short) long wire antenna, the almost complete lack of impedance matching at every point (antenna -> tuned circuit; tuned circuit -> headphones), and probably a poor earth, means that the selectivity of the receiver is woeful. If you have a relatively local AM station, that is all you will copy, with all the others appearing as noise, wherever you tune in the band. All you really change is the s/n of the local station compared to all the others heard together. I found better selectivity in some of the low shortwave bands (4 MHz, 7 MHz?) using a smaller inductor, if I remember correctly. I found a variable inductor (sliding tap) essential in the absence of being given a precise 'recipe' including number of turns, length of antenna, short connection to ground spike or buried pipes, etc. --Nigelj (talk) 15:10, 20 August 2010 (UTC)
Shouldn't we keep this article from becoming an article about tuned circuits, it is how simple crystal radios work. It is easy to get carried away. The way most simple crystal radios work is to tune the BCB by tuning a short (capacitive) antenna with a coil only. The diagram does not reflect that. Adding a capacitor across that coil simply aggravates the tuning problem, doing nothing to help it, and is going in the wrong direction. Also, we have articles on resonance and Q that we can reference for details. John (talk) 19:51, 20 August 2010 (UTC)
Adding a capacitor improves the Q and therefore the sensitivity and selectivity of any xtal set (I assume by masking the combined R & C of the antenna with a purer C term). Capacitors may have been expensive and rare in the 1920s and 30s, unavailable under Nazi occupation in the 1940s, but in fairly standard usage by the 1950s if that reference is anything to go by. --Nigelj (talk) 21:29, 20 August 2010 (UTC)
I agree. John, there are hundreds of crystal circuits and most of them do use a tuning capacitor of some kind. I have researched xtal sets back to the 20s and have added more than 80 references to this article. I can give you dozens of examples. This article is not just about "simple" crystal radios, but many "simple" circuits do use a capacitor, and solve the tuning range problem easily by taps on the coil or a loading capacitor in the antenna circuit. Besides the lower Q, the main problem with omitting a tuning capacitor is that the tuning range varies with the capacitance of the antenna. If a longer antenna is used, the coil may not tune to the top of the band.
And you're missing the point of this particular image. It needs to be a simple introductory circuit people can refer to while reading the "How it works" section. The majority of readers will be nontechnical people who don't know what a "tuned circuit" is. If the example circuit lacks an explicit tuning capacitor, it will be confusing. The section already has a pictorial schematic of a coil-only circuit, right at the top. --ChetvornoTALK 21:57, 20 August 2010 (UTC)
For many years now, there have been toy radios, sometimes sold for $1, that are LC based crystal sets. The small capacitors are pretty cheap. I have even seen ones built into a pen. (That I don't think would write.) But yes, without C is probably fine, too. Gah4 (talk) 01:56, 12 March 2019 (UTC)

Removal of mention of capacitance in 'How it works' section[edit]

I notice that, in addition to deleting the image, all mention of capacitance has been removed from the "Tuned circuit" bullet point at the top of the 'How it works' section. This is grossly false, and misleading to nontechnical readers, since it defines a tuned circuit as an inductor alone: "[A tuned circuit] . . . consists of a coil of wire called an inductor or tuning coil used to tune in different stations." Helllowwwww???? As the text made clear before it was eviscerated, a tuned circuit in a crystal radio always consists of an inductor and a capacitance, but some circuits use the capacitance of the antenna. The original description was thoroughly sourced. I don't want to be accused of continuing the above edit war, but unless somebody comes up with a compelling reason this dreck should stay, I'm changing it back. --ChetvornoTALK 22:24, 20 August 2010 (UTC)

I agree, it was clearer before (e.g. [1]) and I don't see any reason why the capacitor version should be expunged from the article. Maybe it can be better still with the compromise of introducing the capacitance inherent in a 'short' (detuned) antenna, and the possible use of this as part of the tuning 'tank', with disadvantages, when a capacitor is unavailable or impractical. I'm not sure where this new material should go, but somewhere in the 'How it works' section. --Nigelj (talk) 22:53, 20 August 2010 (UTC)
Changed it back --ChetvornoTALK 23:05, 20 August 2010 (UTC)
What I read now is better in my opinion. I did not intend to remove 'all mention' of C, if I did (?) I got confused. However, there are still substantial issues with this discussion of resonance, and how the antenna's L and C get tuned with reactances in crystal sets. There is still wrong stuff being stated as fact; this topic is not trivial and kinda beyond the scope of most readers of this article. I wish y'all would fix it or leave it to the antenna folks in the antenna article. I would suggest consider omitting stuff rather than misleading people. I hate to criticize without offering the solution, but I simply don't have time now. John (talk) 05:46, 21 August 2010 (UTC)
Sorry, I should have realized the capacitance thing was just an oversight. I hope you'll explain about the errors you mention when you have time. BTW, I appreciate you raising the issue of the tuning range of the "naive circuit", even if I disagree with your solution. This article is kind of a specialized subject and I'm glad there are knowlegeable editors working on it. --ChetvornoTALK 07:22, 21 August 2010 (UTC)
No problem, I will try but it may take me a few weeks at least. If someone else wants to take it on, that'd be nice; my hope was to leave it alone in this article but it keeps coming up. This is involved because an antenna acts like both coils and capacitors and they depend on the frequency. An electrically short antenna has too much C and needs L in the radio to tune the whole system of radio & antenna. It also affects Z matching since X is part of Z = X + R. The previous version of Fig 1 would have had to add too much L to tune with C, and the C would have needed way more range than any physically realizable variable cap could muster. There are references for this issue in some of the old 1920s textbooks but thats another research project.John (talk) 16:39, 23 August 2010 (UTC)
Yes, they used loading coils and capacitors to tune out the reactance of the antenna. I didn't go into that in the "How it works" section, in the interest of keeping it short and accessible for general readers, but maybe something should be said about it. Do you think the existing text states things that are false or misleading? Where? --ChetvornoTALK 23:25, 23 August 2010 (UTC)
To mention a couple, I think I see problems with the treatment of Capacitor in the Tuned Circuit section and the the Crystal Detector section is completely wrong in stating that detection efficiency is related to low forward threshold. John (talk) 14:16, 26 August 2010 (UTC)
Yes; technically there is no sharp diode "threshold" but only the exponential IV curve. But I didn't want to go into the Shockley equation and the resulting reciprocal relation between a diode's forward current and its small-signal AC resistance. The idea of a "threshold" for conduction is an approximation commonly used for diode's IV curves. It's an easy way for novices to visualize the detector's sensitivity problem, and especially why a bias voltage can improve sensitivity, by getting rid of the threshold. Crystal radio detectors' lack of sensitivity is often explained that way, as shown in the references. If you can come up with a better way, I'm all ears. --ChetvornoTALK 17:48, 25 August 2010 (UTC)
I think the best way to illustrate the problem, found in early books, is to give a graph of a diode's IV curve with the applied sinusoidal voltage against the horizontal axis, and show graphically how it creates a very poorly rectified output current waveform. Maybe such an image could be gotten from an early book and put in the article. --ChetvornoTALK 17:35, 25 August 2010 (UTC)
You could make a real curve for a 1N34 with pSpice along with the real detected signal curves. But more importantly, there is no sharp diode threshold, but further, there is no diode threshold at all, this notion is not even approximately right. This comes as a shock to most engineers, but the diode curve has no knee (NONE) at all, it is a complete optical illusion. Mathematically the diode equation (and real diodes conduction) is an e-to-the-x form curve (with slight variation, corrupted slighted by "n" the so-called efficiency parameter), and that has no point of max curvature (knee), it just does not exist. Don't take my word for it, check the second derivative of a real 1N34 plot, or plot this and look at it on different scales, the "knee" changes. I realize the notion of a knee is very popular (to the point it is emotional to many), but utterly fictitious. It misleads people to think germanium diodes will always detect better than silicon. Not true, it completely depends on the impedance of the radio tuner and transformed antenna. A lot of crystal set design is substantially suboptimal because of this misunderstanding; being popular doesn't make it technically right. It doesn't even make sense, though college professors actually teach this nonsense in top universities. Reality is that, for signals weak enough for the curve to be approximated by a linear segment, the inverse slope of the IV curve at the operating point (AKA detection Z) needs to match the Z of the radio, again not a trivial issue. It is a tough point to communicate, and few will believe it, I would suggest just deleting the wrong stuff in the article and reference some Agilent papers or something. John (talk) 14:16, 26 August 2010 (UTC)

The illustration incorrectly describes the slider[edit]

The slider in the illustration is on the right, not the left. —Preceding unsigned comment added by (talk) 13:22, 18 July 2010 (UTC)

Fixed it. Thank you very much for catching my screwup. I must have had a touch of dislexia :) --ChetvornoTALK 16:20, 18 July 2010 (UTC)

The detector[edit]

The explanation of how the crystal detector is wrong. I have made several attempts to correct (perhaps not very well) but my edit is simply removed. There does not seem to be any point in correcting errors if the corrections are simply removed and the incorrect explanation retained. Davidpetertucker (talk) 17:09, 30 January 2011 (UTC)

Crystal detector

This explanation would only work is the signal measured several volts peak to peak.

Since the signal in only micro volts p-p how can a diode possibly rectify such a signal?

The real explanation relies on the fact that the diode is a nonlinier conductor and this mices the sidebands and carrier frequencies resulting in the recovery of the modulating frequency (among others). The purpose of the bias is to move the opperating point to the part of the diode's curve which maximises the amptitude of the desired signal.

— Preceding unsigned comment added by Davidpetertucker (talkcontribs) 09:55, 2 February 2011 (UTC) 

The latest revision is better because it avoids the false myth of "a knee in the diode curve," and is more small signal oriented. A technical explanation of this can take the frequency domain approach of collapsing sidebands, but that is very abstract for many. There is also a very clear time domain explanation that requires a very clear understanding of detection. (talk) 17:25, 9 November 2011 (UTC)

I wrote the original explanation, that used the "knee of the IV curve" language. I was aware that this was an inaccurate explanation, but I thought that "the signal voltage has to overcome the diode's forward voltage drop" was a more concrete, visual way of explaining to a general reader why biasing the junction improves sensitivity (the same wording is used on many crystal radio websites, see refs). I couldn't think of a way to explain the actual nonlinear curve understandably. Davidpetertucker and others rightly pointed out that this was a bogus explanation. I reverted your edits; I felt at the time it was more important to keep it simple. recently did a good job of correcting the explanation. I understand the desire to correct the "myth of the knee" and present an accurate picture of small-signal rectification in a xtal set, but its inclusion makes the description of the how the diode works awfully "fuzzy". --ChetvornoTALK 05:47, 10 November 2011 (UTC)

Wrong Units[edit]

"energy of only 10−16 W/cm2"

W/cm2 is a power density NOT energy (which should be Joules). Someone put this right, please. (talk) 10:23, 11 May 2011 (UTC)

Fixed. But you can fix things, too. Wikipedia:Be Bold used to be a guiding principle around here. --Wtshymanski (talk) 13:52, 11 May 2011 (UTC)

Need to delete Construction section[edit]

Do we need a section about how to make a radio that is ineffective at tuning? John (talk) 04:43, 28 June 2011 (UTC)

It wasn't really about construction and operation, was it? I"ve moved it to the discussion of tuned circuits; it's an interesting point and this was exactly the explanation I needed as a kid to explain why I could only get CFRW (1470) and never CBW (990) on any of the crystal sets I made or had given to me. --Wtshymanski (talk) 13:46, 28 June 2011 (UTC)


Was this actually used in the day, or some contributor's original research? I've never seen a book or Web page describing this technique of setting a cat's whisker, and One thing that troubles me is how one is expected to hear anything in the headphones with a buzzer running within arm's reach. --Wtshymanski (talk) 22:32, 10 November 2011 (UTC)

nevermind. [2]] says the WWI BC14A artillery spotter receiver (no tubes!) had a tuning buzzer. --Wtshymanski (talk) 22:44, 10 November 2011 (UTC)
Yes, they were actually used. Check the reference at the end of the sentence, that's what its for. --ChetvornoTALK 23:49, 10 November 2011 (UTC)
Just because most Wikipedia references are junk doesn't mean I should always ignore them. A bad habit of mine. --Wtshymanski (talk) 15:32, 11 November 2011 (UTC)

File:Modern crystal radio set.jpg Nominated for speedy Deletion[edit]


An image used in this article, File:Modern crystal radio set.jpg, has been nominated for speedy deletion for the following reason: All Wikipedia files with unknown copyright status

What should I do?

Don't panic; you should have time to contest the deletion (although please review deletion guidelines before doing so). The best way to contest this form of deletion is by posting on the image talk page.

  • If the image is non-free then you may need to provide a fair use rationale
  • If the image isn't freely licensed and there is no fair use rationale, then it cannot be uploaded or used.
  • If the image has already been deleted you may want to try Deletion Review

To take part in any discussion, or to review a more detailed deletion rationale please visit the relevant image page (File:Modern crystal radio set.jpg)

This is Bot placed notification, another user has nominated/tagged the image --CommonsNotificationBot (talk) 10:00, 15 June 2012 (UTC)

Crystal earpiece impedance[edit]

The article describes the impedance of the crystal (piezo) earpiece as being ~ 1 Mohm. This is the DC value. IIRC, the AC impedance is still quite low (~5kOhm), so a coupling transformer is desirable. Have I got this right? I haven't changed the article; someone better informed may wish to. — Preceding unsigned comment added by (talk) 02:29, 10 August 2012 (UTC)

Ground connection[edit]

In order for a radio to work it is necessary to have a ground connection or the radio needs a very large capacitance to act in lieu of the earth.

This requirement is similar to that for a power delivery system. The current from a power plant arrives at your home, it passes through various devices in your home and then travels to the ground wire in the circuit box and from there to a ground wire buried in the earth under your home. In a sense the earth itself is the conductor that returns the current to the power plant where there is a massive grid of conductors buried in the ground. The earth is not a great conductor but there is a lot of it and so it works. There is no wire that returns the current to the power plant. The earth could be replaced with a gigantic capacitor but that would be prohibitively expensive. Similarly, a radio needs either a ground connection or to be connected to a large capacitance that acts like the earth. The use of early crystal radio sets in aircraft and such was problematic due to the vibrations to which the radio was subject not due to grounding problems. I am not certain about the grounding of aircraft but will speculate that the tremendous amount of air through which a moving craft is passing will act like a sufficient ground/capacitance. Zedshort (talk) 16:16, 12 September 2014 (UTC)

I suspect Wtshymanski understood that; he may have been thinking of counterpoises. As you alluded to, a lot of early crystal sets used capacitive grounding via a counterpoise instead of an earth ground; that is what early aircraft and zeppelin wireless sets used. 1, p.312 From the early 1900s there have been portable crystal sets carried on the body which could not have earth grounds. For example, in the photo at the top of the page, the crystal radio the boy is listening to does not have a ground wire, but is grounded sufficiently by the capacitance of his body through his hand holding the set. --ChetvornoTALK 16:56, 12 September 2014 (UTC)
Cell phones. Airplanes. Spacecraft. How much capacitance is there between a satellite in synchronous orbit and the Earth? A Mars rover and the Earth? (For that matter, light is just really really high frequency radio waves and propagates just fine over intergalactic distances.) The Tesla model is inappropriate - radio receivers work with the free space field and any use of earth is for convenience. At the frequencies used for AM broadcast, it's inconvenient to build an antenna that collects enough power to operate an earpiece, so you use the earth as part of the system - but it's not really correct to think of Tesla-style currents running through the Earth from transmitter to receiver. --Wtshymanski (talk) 23:59, 12 September 2014 (UTC)
What are you basing the above upon? Is it just opinion? A ground is shown in most all diagrams of radios. Why are they shown? What purpose does a ground have? In the case of a rover or spacecraft, they too have enough capacitance to function as a ground. Your injection of a counterpoise into the definition makes no sense. A counterpoise is used in the case that a proper ground cannot be made at the transmitter site. Zedshort (talk) 04:42, 13 September 2014 (UTC)
Zedshort, were you addressing Wtshymanski or me? A crystal radio does not have to have a ground, it could use a dipole antenna; it's just that with the long wire antennas needed, a dipole would require a second long wire, doubling the length of the antenna, to give the same reception. A monopole, due to the ground reflection, has 3 dB more gain than an equivalent dipole. Another issue is that AM stations broadcast in vertical polarization, so a horizontal dipole isn't as efficient as a vertical monopole. A crystal radio with a monopole antenna without a ground can also work, because the radio's case and the hand holding it can function as the other half of a dipole antenna. It will be much less sensitive but is probably adequate for the strong signals in an urban area. --ChetvornoTALK 10:14, 13 September 2014 (UTC)
A dipole antenna does not use a ground, but a monopole antenna, which is what was used with crystal radios, requires the other side of the transmitter or receiver to be connected to some type of conductive ground plane, which functions as a reflector, not a radiator. 1, p.142 It doesn't have to be connected to the Earth. The size of ground plane required depends on the frequency, it should ideally extend at least a quarter-wavelength from the antenna. The MF and LF broadcast freqs received by crystal sets require a large ground plane, at least 1000 ft in dia., so the Earth is the most convenient conducting plane available. If an Earth ground connection cannot be made, a counterpoise, a wire screen capacitance to Earth, can be used.1, p.142, 2, p.523 "Counterpoises" were widely used during the wireless era for receivers as well as transmitters. 3, p.157, 4, p.217, 5, p.278, 6, p.311-312, 7, p.45 However early radio researchers didn't really understand how they worked. If a counterpoise is too small or too far away from the Earth, its voltage will not remain constant and it will radiate (or in a receiver pick up radio waves), simply functioning as the other half of a dipole antenna. That is actually how the whip antennas in modern walkie-talkies, CB, and portable FM radios work. 8, 9, p.53, 10, p.48-49 The whip is nominally a "monopole" antenna, but the other side of the transmitter or receiver is merely "grounded" to the radio's chassis or circuit board. So the chassis and whip function as the two halves of an asymmetrical dipole antenna. This is also how small portable crystal radios without grounds work. --ChetvornoTALK 09:04, 13 September 2014 (UTC)

This photo shows a woman in 1922 with a crystal radio in a parasol. It has a wire antenna in the parasol, but is not grounded. The radio itself and the woman's body function either as the "ground" or the other half of the antenna. If they have enough capacitance to ground they function as a counterpoise. If not, they function with the parasol as the two halves of a dipole antenna; the incoming radio wave induces opposite phase oscillating voltages in the antenna and the radio's case including the attached earphone wires, etc. which constitutes the input signal applied to the tuned circuit. --ChetvornoTALK 09:04, 13 September 2014 (UTC)

Well, this has been a truly fascinating exchange and collection of factoids. Now, would someone please explain the presence of a ground connection in diagrams of radios without using the circular reference of a counterpoise, and do so in twenty words or less. Thank you. Zedshort (talk) 14:03, 13 September 2014 (UTC)

Sorry, I may have had too much coffee last night. What it boils down to is: The oscillating current from the wire antenna, after passing through the radio, needs some conductor with a high capacity to flow into. The Earth is ideal, but it doesn't have to be the Earth. --ChetvornoTALK 18:47, 13 September 2014 (UTC)
So, in order for a current to slosh back and forth within a circuit, the charges needs something to slosh into and out from that has lots of space to allow the sloshing, that thing being a capacitor. Whether the antenna is connected to the ground directly via a conductor into the ground or coupled to the ground by a counterpoise makes no difference, provided the capacitance is sufficiently large relative to the need. But of course, in writing the article, that idea needs to be expressed. Should the article describe the "ground" as a very large capacitance? If it is described as I did, as a large reservoir, that might work. The ground wire analogy with a physical connection back to the transmitter obviously fails in the case of spacecraft communications. Zedshort (talk) 20:05, 13 September 2014 (UTC)
The simple statement is that crystal radios need good antennas because the power that drives the headphones comes from the antenna. The more power the antenna delivers, the better. A half-wave dipole is a good antenna, and it does not need a ground (but for practical reasons might need a balun). Over-the-air powered RFIDs can use dipole antennas. AM transistor radios also used ferrite loopsticks; they capture the magnetic field and do not need a ground. Consequently, a good ground is not a requirement. Back when crystal radios were popular, the operating frequency was low, the transmitted power levels were high (that pesky atmospheric noise required it), and people often used a crappy, untuned, single-ended, long-wire antenna. Grounding those antennas improved them. If the ground were perfect (which it isn't), it would make an image antenna and cause current to flow in the antenna wire. Simplistically, if there were no connection to ground (and no capacitance to ground), then no current (and no power) would flow from the single-ended antenna. Glrx (talk) 01:17, 15 September 2014 (UTC)

Impedance Matching[edit]

I have made a quite serious edit to this section. Given the interest in crystal sets evident in this talk page, it is likely that what I have done will prove to be controversial, so I shall try to explain. Firstly I have corrected the statement of the max power transfer theorem, so that it reads correctly. Impedances should not be equal for max power transfer, but one should be the conjugate of the other. I do not expect any disagreement over this. However, it seems to me that the only respect in which this theorem can be applied to a crystal set is that the impedance of the power source (i.e. the aerial) should be matched to that of the the load (the earphones). Ideally, the equivalent resistance of the tuned circuit at resonance should be severely mismatched to the source impedance of the aerial in order to minimise resistive losses in the tuned circuit, which can only be wasteful. If the tuned circuit were power matched to the aerial, maximum power would be transferred to the resistance of the tuned circuit, which is not what is required. It is for this reason that I have deleted all references to the idea of matching the aerial to the tuned circuit. Transforming the equivalent resistance of the aerial has significance in relation to selectivity, and this concept is best treated separately. Both topics (impedance matching for max power transfer to the headphones, and aerial resistance transformation for improved selectivity) are complicated by the fact that so many different crystal set circuits are used. In the circuit which accompanies the "Impedance Matching" section, it is clear that the left-hand slider can be used for tuning, and the right-hand one for impedance matching, and that the right-hand slider will also affect the Q of the inductor, and therefore the selectivity. In more complicated circuits where a variable capacitor is used for tuning, and the aerial is connected to a variable tap on the inductor, the useful effect of the different tap positions will be to improve selectivity when the tap is low, since then the Q of the parallel tuned circuit will be at its greatest. In this type of circuit, the resistance of the headphones will lower the Q of the parallel tuned circuit, so for greater selectivity, the resistance of the phones should be high. In this circuit, power transfer to the phones will also be affected by the tap position, so that the tap control can be used either to improve selectivity, or to achieve best sensitivity, depending on requirements. g4oep — Preceding unsigned comment added by (talk) 11:43, 4 January 2015 (UTC) G4oep (talk)

I don't agree with your idea that maximum power is produced by mismatching the impedance of antenna and tuned circuit; neither do any of the sources cited in the article. The aerial is connected to the crystal/earphone circuit through the tuned circuit. Although there are many different circuits, the load resistance the aerial sees is some parallel combination of the resistance of the crystal/earphone circuit and the parasitic resistance of the tuned circuit (scaled by the impedance transformation of the antenna tap on the coil). You can't separate those two resistances. To transfer max power to the earphone, the antenna must be matched to the impedance it sees, which is the parallel combination; that's what the maximum power transfer theorem says. The max power theorem doesn't discriminate between resistances in the circuit; maximizing power to the earphone also maximizes power lost in the coil. --ChetvornoTALK 21:02, 31 March 2015 (UTC)

Hi - I have looked at the paragraph you refer to, and although there is not much there to object to, I feel that the ideas involved in this section could be presented more rigorously. The paragraph starts by stating that the aim is to transfer power to the headphones, then states "Therefore, in better receiver circuits, to match the antenna impedance to the receiver's impedance...". The point that I tried to express in the material you deleted is that the first objective is not necessarily achieved when the second is fulfilled. True, it is necessary to match the receiver's aerial and earth terminals to the actual aerial/earth system in order to maximise power absorbed from the aerial, but then this absorbed power must be transferred to the phones with as little as possible lost in the tuned circuit. The resistive component of the input admittance of the receiver (considered as a parallel equivalent circuit) can be represented by 2 parallel components - 1) the transformed equivalent parallel resistance of the coil (or of the tuned circuit as a whole- Rc)) and 2) the transformed resistance of the phones/detector combination (Rp). For maximum power in the phones, Rc should be large compared with Rp since in a parallel equivalent circuit, the same voltage is applied to both resistors, and P is inversely proportional to R (P = V^2/R). This condition (Rc >> Rp) can be represented by saying that the antenna should be matched to the phones, and mismatched to the resistance of the inductor. That was what I wrote, and I feel that it should not have been deleted without proper justification. The resistances of the coil, and of the phones are indeed separable, and the design and adjustment of the circuit can be altered to control the fraction of the available power which is dissipated in each simply by adjusting their relative magnitudes. These magnitudes are under the control of the various transformation ratios. This idea is completely false: "maximizing power to the earphone also maximizes power lost in the coil". I have altered the text again. I expect that someone will quibble that the points I have made are not referenced, but it should be obvious to anyone who understands ac circuit theory that what I have written is correct; trust me - I am an engineer. If you don't understand circuit theory, don't change it. g4oep.

The issues are much more complex than transformer matching. I reverted to the sourced version. Glrx (talk) 01:29, 7 April 2015 (UTC)

Modulation fix wanted[edit]

In the section about "use as a power supply" it states that AM stations modulate only 30 percent. This is only partially true. Aircraft navaid stations modulate their IDs 30% to leave room for voice modulation over the ID. But broadcasting stations routinely modulate 100% and even 125% in the positive direction. Can someone fix this please? (talk) 01:12, 10 January 2015 (UTC) Chip Veres 1-08-15 ]

They might have peak modulation that high, but as a power source it is average that counts. With no modulation, an AM transmitter runs at 1/2 of power that it does at full modulation. (That is, half the power in the carrier, half in side bands.) Also, carrier power (no modulation power) is 1/4 of Peak_envelope_power. Gah4 (talk) 04:40, 12 March 2019 (UTC)

Crystal Radios CAN demodulate FM signals![edit]

Intro page says "By the nature of their operation, crystal radios can only demodulate amplitude modulation (AM) signals, not frequency modulation (FM) or digital signals."

Not true...

Crystal Radios CAN demodulate FM signals very easily. There's a number of HOWTO DIY pages on this.

See: for one example. or or — Preceding unsigned comment added by (talk) 03:53, 31 March 2015 (UTC)

I agree. Slope detection is a time-honored way of demodulating an FM signal with an AM radio, although I don't know how much this technique was used with crystal radios. --ChetvornoTALK 06:21, 31 March 2015 (UTC)
Well, I would be careful... I don't see anything in the three references above that actually says it has been successful in practice (although I haven't read every word in each of them - perhaps someone will quote the relevant bit(s)). First, the reception bandwidth of many practical crystal sets is very wide, due to lack of good filtering or tuning. This means there isn't much 'slope' anywhere to perform the FM to AM conversion. Secondly, the power capture of a traditional crystal set depends on a very long (e.g. hundreds of feet) random wire antenna. These are not much use at VHF, where most FM is transmitted, and would not capture more received power than a one or two meter tuned dipole. I don't know that such short reception antennae, even if tuned, would ever pick up enough power to hear directly unless one was very close to a transmitter. (VHF transmitters don't put out megawatts, as they know the range will be limited by line-of-sight anyway). I'd like to see a quote from an actual WP:RS before stating that crystal radios are practical on FM bands. --Nigelj (talk) 19:41, 31 March 2015 (UTC)
Good points. Even if it has been done just to prove it was possible, that's a long way from saying it is a practical technique. Without more specific sources I don't think FM reception should me mentioned in the article. --ChetvornoTALK 20:26, 31 March 2015 (UTC)
There is discussion, I believe in the FM article, on a narrow-band FM broadcast system preceding the current system, and that could be demodulated by AM receivers. But we don't have transmitters like that, so it doesn't apply now. Gah4 (talk) 02:59, 12 January 2019 (UTC)

I've researched this question and concluded that reception of normal wideband FM is also possible with some crystal sets up to 30km from a transmitter. I've mentioned this fact and a link, but have found many more references during my research. We could probably use some wording to mention some of these as well, like their designs, efficiency, techniques etc. Main discussion of an optimized crystal discriminator design:

Slope detector vs. FM discriminator response curve:

Previous design:

Some others (also referenced above and from other places):

-bkil (talk) 22:46, 3 July 2020 (UTC)

Recent change to explanation of detector[edit]

Recently the description of the crystal detector in the Design section bullet points was changed from:


  • A semiconductor crystal (detector) which rectifies the alternating radio frequency. The crystal does this by allowing current to pass through it in only one direction, blocking the other half of the oscillations of the radio wave. This rectifies the alternating current radio wave to a pulsing direct current, whose strength varies with the audio signal.

I think the previous version was better and should be restored. The detector is the most important part of the radio. I think it is important for readers to understand what it does; it demodulates the radio signal, separating the audio modulation from the carrier wave and producing the audio signal which is converted to sound waves by the earphone. The new text, in addition to being repetitive, doesn't make this clear. --ChetvornoTALK 22:16, 7 May 2016 (UTC)

I agree the previous version quoted above is better than the latter. I also think that AM detection with a diode could be better explained than either. Perhaps someone with a recent university level textbook on the subject could do so based on an actual, you know, cited source? --Nigelj (talk) 19:20, 9 May 2016 (UTC)
I changed it from "extracts" to "rectifies" as I am certain that a diode does nothing more than to rectify. I think I have had the discussion once before about how the audio waveform is actually "extracted" but it was not really settled. I am very certain that the "extraction takes place in the earpiece. The earpiece is an electro-mechanical device that is incapable of responding to the very high frequency half wave form produced by the diode. It can, however, respond very slowly to the many, many, many pulses of electrons passed through to it from the diode in the form of the rectified and modulated waveform. The earpiece is a spring-mass-damper system, but also partly electrical. Since it has mass, its response time is very, very long compared to the carrier frequency input. It cannot respond, but can be slowly moved by the many impulses from the carrier and hence the audio is produced in its "attempt" to follow the carrier. Unfortunately, I do not have a good reference for precisely that. If you have good references that say that the diode in fact creates the audio signal rather than a pulsating modulated waveform, I would like to read it. I could be wrong. If the carrier is 600 kHz the period would be 1.7x10^-6 sec or 1700 nano sec or 1,700,000 pico seconds. I think that Schottky diodes have a very quick response time, as low as the pico second range in which case it could reproduce the carrier frequency quite accurately and the argument that it would "smear out" or average out the input into a audio waveform doesn't work. I am honestly interested in this. Sorry about my heavy handed editing but I am not shy about rephrasing things that seem wishy-washy. One thing I am pretty sure of is that people are very, very vague when it comes to an accurate description of exactly how that part of the circuit works. I reiterate. The earpiece is an electro-mechanical system that cannot reproduce the carrier but can respond slowly and hence reproduce the "audio waveform" that is implied by the modulated carrier. Zedshort (talk) 22:48, 9 May 2016 (UTC)
@Nigelj: The explanation down in the Crystal detector section is a little better (but also needs to be improved), and is supported by citations. I give 7 additional sources below.
@Zedshort: Yes, we discussed this earlier on my talk page. I agree with a lot of your edits, they make the article easier to understand. But on the subject of the detector, electronics and telecommunications sources are clear. The diode detector (envelope detector, AM demodulator) has been used throughout telecommunications for 100 years, it is thoroughly analyzed, and there is nothing "fuzzy" or "wishy-washy" about our understanding of how it works. By rectifying the radio signal it extracts the modulation from the radio frequency sidebands of the carrier wave and translates it back to its original frequency (called the baseband), producing an audio signal. This can be proved by connecting a spectrum analyzer to the input and then the output of a crystal detector in a radio. At the detector input the analyzer will show no audio frequency components, while at the output audio frequencies will be present. This is also shown mathematically in many textbooks. The detector output will have radio frequency pulses from the carrier wave present in addition, but these are unneeded.
The earphone converts the audio frequency alternating electric current to sound waves. It has no other role. It is true as you say that in some crystal radios it functions as a low pass filter to remove the carrier pulses in the detector output. However in many circuits a bypass capacitor performs this function instead. Regardless, the point is that this is not a necessary function in the circuit. The carrier components in the detector output are much higher in frequency than the audio and don't interfere with the audio signal. Even if the earphone reproduced the radio frequency carrier pulses in addition to the audio, they would be in the ultrasonic range and we couldn't hear them, so the radio would sound the same.
Below are sources supporting my wording:
--ChetvornoTALK 09:36, 10 May 2016 (UTC)
Thanks much for all that. If the sources say the rectifier "extracts" the audio signal then that is what the article should say as we are not here to do research. But one last question. If the earpiece is not included in the circuit, would the audio signal be present there, i.e. would its waveform be displayed on an analyzer? I suppose if such a test should be performed, the equivalent reactant of the earpiece would need be included in the circuit to simulate its electronic qualities. I will follow and read as much as I can comprehend of the links you provided. Thanks much. Zedshort (talk) 12:41, 10 May 2016 (UTC)
Although the first statement is better, I'm not sure about either.
If there is an amplifier, then the RF signal amplitude can be a couple of volts, the diode will be a good rectifier, and it will be a reasonably linear envelope detector.
If there is not an amplifier, the signal levels may be low enough that the devices operate not as on-off rectifiers but rather square law detectors.
I've never seen an authoritative discussion of crystal radio detectors. Non-RS conclusions seem to be mixed.
Glrx (talk) 20:27, 15 May 2016 (UTC)
@Zedshort: The earphone is in series with the detector (in most circuits) and as you say, its impedance is usually necessary to make the detector operate correctly. But if it was replaced by an equivalent load impedance made of a resistor/capacitor/inductor combination, the audio signal would still be present in the circuit. --ChetvornoTALK 11:14, 16 May 2016 (UTC)
@Glrx: I agree, a discussion of linear vs square-law operation would be good to include in the Detector section. But on a more basic level, the article needs to acknowledge that it is the detector which demodulates the signal, which is what this sentence is about. --ChetvornoTALK 11:14, 16 May 2016 (UTC)
Hausmann p 61 shows square law detection. Glrx (talk) 15:11, 16 May 2016 (UTC)

Why not a bridge rectifier?[edit]

I might be being dumb and missing something fundamental, but wouldn't substituting the simple diode for four of them in a bridge rectifier provide a signal of twice the strength? After all, the idea for the envelope detection setup is to remove the inverse side of the original wave; if we instead invert it, we get twice as much positive signal. Yes?

Obviously this would be a massive pain to try and implement on an original oxide-coated-conductor and cat's-whisker type radio, but a semiconductor based model could use it... (talk) 20:14, 9 July 2016 (UTC)

The major issue with the diode detector used to rectify the weak signals in a crystal radio is the forward voltage drop of about 0.3 V (in germanium) required to turn it on and begin rectifying. A bridge rectifier has two diodes in series in the signal path, and thus has twice the voltage drop. I remember seeing a discussion of the bridge rectifier in crystal radios somewhere on the web, and the conclusion was that it was not better. --ChetvornoTALK 23:50, 9 July 2016 (UTC)
Yes. Even for low voltage power supplies, the loss is too much. Consider the 3.3V supply common for computers now, and lose 1.4V (two Si diode drops) instead of one. But two diodes and a center-tapped transformer give only one diode drop. Can you use a center tapped coil for a crystal radio? Power supplies now use a synchronous rectifier switching FETs on and off. Gah4 (talk) 04:45, 12 March 2019 (UTC)

Can we find a better photo for the lead?[edit]

The photo shows a boy wearing what appears to be a modern device of some sort. It is not really informative. Perhaps we can find a photo of an actual crystal radio and its associated circuit diagram. Zedshort (talk) 13:51, 19 December 2016 (UTC)

That is an actual crystal radio; that's what modern ones look like. There aren't very many pictures of modern crystal radios in Commons. I thought this one had several good features; it shows actual use by a person instead of just showing a crystal radio sitting on a desk; the piezoelectric earphone is visible plugged into the boy's ear; and it shows how small and portable they can be. The black antenna wire is also visible. --ChetvornoTALK 19:33, 19 December 2016 (UTC)
The File:Crystal radio enthusiast.JPG is odd. There's lots of boy, little radio, and he could just as well be listening to an MP3 player. Furthermore, I doubt many modern boys are listening to crystal radios these days. File:Crystal radio advertisement.png covers the same ground, the radio is a bit more prominent, and it's more historically accurate. I'd replace the boy with family ad pix. The article has plenty of schematics and pictures of radios, but it doesn't have a good homemade crystal radio pix for the lede. I want a cross of the inductor-tuned File:Crystal radio.agr.jpg set but with an iconic cat's whisker File:Kristallradio (3).jpg; a quick search comes up dry. Here's a pictorial of a capacitor-tuned set that's close to what I want. Pictures such as File:Poste a diode 1.JPG look too modern and don't even show the crystal. Glrx (talk) 21:13, 19 December 2016 (UTC)
Crystal radios have no power source and unless we know for a fact that that device is not powered it should be removed. Too much boy, not enough radio. Zedshort (talk) 23:56, 19 December 2016 (UTC)
I agree that the boy in the picture could just as well be listening to an MP3 player, and the whole picture looks posed and unrepresentative of how a real crystal radio radio might actually be used. The essential components of a crystal radio need to be much more visible. The current picture is rather misleading, in that real-world crystal radios typically have a ground connection and a more substantial antenna in order to operate. Reify-tech (talk) 00:36, 20 December 2016 (UTC)
Actually it is a pretty representative picture of modern crystal sets sold to young people and shows exactly how they are used. Many modern crystal sets are portable like this and do not have a ground connection but just an antenna wire (the black wire visible dangling from his hand). They are used in strong signal areas and probably the hand holding them serves as a capacitive ground. When I was young I had a crystal set very similar to this one and it had no ground wire. Here are some pictures of novelty crystal radios [3], [4], [5] shaped like rockets, etc. Note they have an antenna clip but no ground connection. --ChetvornoTALK 10:19, 20 December 2016 (UTC)
@Glrx: I think a historical picture as you describe would be good, but I think there should also be a picture of a contemporary crystal radio up top. This article is not History of crystal radios. Crystal radios are still made today, and many of the general readers who come to this page (I'm thinking young geeks) may not care about the history, they may only want to know about modern sets they themselves could build or buy. The article currently has 16 pictures of historical sets (did you look at the Gallery?) and only 2 pictures of contemporary sets. No, diodes are not much to look at compared to the old cat's whisker detectors, but they are what are used in modern sets. --ChetvornoTALK 10:19, 20 December 2016 (UTC)
@Chetvorno: I don't buy into the notion that we must show pictures of modern crystal sets or even novelty crystal sets. I think the modern crystal radio is pretty much irrelevant in western society and perhaps elsewhere, too. There might be an argument that it is used where MP3 players are too expensive or batteries are extravagant. Ten years ago I read an article that hand-cranked radios were important luxuries in Africa. No batteries required, and it would be the source of entertainment at a village party.[6] It's pretty easy to share a radio with a powered speaker. Maybe a crystal set can be important in similar circumstances, but I'd want some reliable sources that tell me that. Certainly such a radio doesn't need to be in the shape of a rocket ship. Are those villages too remote or sheltered to get good crystal reception?
As a kid, I built some crystal sets, and they had poor performance. They needed strong stations; they did not need a volume control because they did not provide enough audio power to worry about. I could only get a few stations. Adding a transistor provided more gain, but the selectivity was still poor. A transistor superhet with its 9V battery and AGC was a much better AM radio. Still, the underlying performance is crappy AM. I gave up listening to AM music stations a long time ago. There's a reason that FM caught on.
I think the article is historical. The crystal radio is important for what it provided in the early days of radio and for its path to modern semiconductors. The 1N21 and 1N34 aren't that much different from the exotic rock and metal whisker detector. I don't see it as a relevant product today. Sure, there are crystal radios for sale, but I don't think WP is interested unless there's evidence of current impact. Wikipedia is neither a how-to guide nor a buyers' guide; that is not its mission. For commercial perspective, I can buy a crystal radio at Walmart for $20; I can also buy a CD player with AM/FM radio for $20. WP wants to tell readers what they are and how they work, but WP isn't here to describe how to build them or which one to buy.
Glrx (talk) 21:01, 21 December 2016 (UTC)
I'm not suggesting the article tell readers how to build them or which to buy, just that it cover its subject adequately. Of course crystal sets are not used for serious radio listening today - I never said they were. The crystal sets that are produced today, as with most since WWII, are made as educational toys and kits for young people, to learn about radio. That is their "relevance" today. That is why it's good to have a picture of a young person listening to one. There is still plenty of interest in crystal radios: in addition to the many assembled radios for sale there are kits [7], [8], [9], books [10], [11], websites telling how to build them [12], [13], [14], [15] and DIY videos [16], [17], [18], and they are included in high school physics textbooks [19]]. There are also the clubs devoted to constructing crystal receivers [20], [21]. Your suggestion to arbitrarily exclude mention of all this activity from the article seems to stray into WP:POV editing.
I agree that most of the article should be devoted to the history, and it's a fascinating history. But I think that, in addition to the 17 pictures of historical sets, there should be one or two token pictures of modern crystal radios. Ideally I'd like to see a picture of a modern student putting one together; not as a how-to, but just to show the crystal radio's place in modern science education. But I can't find one. The kid listening to a crystal radio (Crystal radio enthusiast.JPG) does illustrate the main use and market for modern crystal sets, so I think the article should include it. --ChetvornoTALK 04:45, 22 December 2016 (UTC)
Unless the "modern" example is clearly not powered it does not belong here. The photo of the boy is not imformative at all. Zedshort (talk) 19:41, 24 December 2016 (UTC)
It is clearly not powered; it is a crystal radio, here is a similar one. Powered radios (and MP3 players) do not have an external antenna wire (black wire); they use internal ferrite antennas. The earpiece is clearly larger than the moving-magnet earbuds used with modern players; it is a piezoelectric earpiece needed for a crystal set. --ChetvornoTALK 19:53, 24 December 2016 (UTC)
You seem to have an inordinate interest in guarding that photo. Do you have some personal association with it? Zedshort (talk) 14:51, 28 December 2016 (UTC)
I have no "association" with it; like I said, we need a picture of a representative contemporary crystal radio, and this is one of the few available. If there were a better picture out there it would be fine with me. Why do you dislike it so much? This happens to be what modern crystal radios mass-marketed to children look like. The fact that you couldn't recognize a modern crystal radio and thought it was a powered radio is a good argument that we need such a picture. --ChetvornoTALK 18:21, 28 December 2016 (UTC)
I think it is simply not informative enough. Zedshort (talk) 18:39, 28 December 2016 (UTC)
Actually, there are a couple of pictures that might be better. How about one of these?
They address the "too much boy, not enough radio" issue. I think I prefer the 2nd one, even though the 1st looks to be of more modern vintage. --ChetvornoTALK 18:56, 28 December 2016 (UTC)

From our German colleagues, a homemade set:

Capacitor tuned; up-close-and-personal diode.

Also, a nice crystal:

Glrx (talk) 20:17, 28 December 2016 (UTC)

I wouldn't mind your File:Detektorempfaenger3.png, but it will be pretty incomprehensible to general readers without labels for the parts. It's also somewhat dated, it looks to be of 1940s vintage. I have no objection to its inclusion, but I think we also need a picture of a more modern crystal radio showing the entire radio, preferably including the earpiece and antenna wire, like the ones above. --ChetvornoTALK 13:59, 29 December 2016 (UTC)
@Glrx: There's room at the top for a 2nd picture and I agree with you that the article should have a historical set up there too. What do you think about one of these? --ChetvornoTALK 04:10, 27 July 2018 (UTC)
It's tough to page all this back in. I'll say no to all of them.
The first picture is the most informative, but it is a poor, washed out, image. I have trouble using it. I like the detector, but the resolution is so low that crystal looks like a binding post and the coil wires merge to a sheet. There's no context for the antenna and headphones.
The second image is a box with dials. It does not illustrate the crystal set idea. There could be a battery or a power cord, the object is massive.
The third image is another box. The headset location is almost cartoonish. The image is also horribly out of focus.
The fourth image has a speaker on the right. Speakers are not consistent with crystal sets.
Glrx (talk) 18:31, 27 July 2018 (UTC)

I agree that the current lede photo is unsuitable. It shows nothing about the topic and is much later that the important use of such radios, which was before WW II and certainly before transistors. I would suggest File:Kristallradio.JPG. It clearly show the cats whisker detector and the crystal and the headphones. There are several other views of this item as well. Images of modern crystal radios belong in a separate section on later usage.--agr (talk) 19:38, 27 July 2018 (UTC)

@Glrx: I agree they all have drawbacks, but it seems to me any one would be better than no picture. I can photoshop the photos, increase the contrast of the first one and crop the irrelevant speaker from the last one. What do you think of agr's picture? --ChetvornoTALK 03:50, 28 July 2018 (UTC)
agr: This has been discussed above. I agree there should be a historical radio at top, and your photo File:Kristallradio.JPG would be fine with me. But there is room for 2 or 3 photos at top and I think one of them should be a modern (post WW2) radio. There is still much interest in contemporary crystal radios, as educational kits [22], [23], [24], books [25], [26], websites telling how to build them [27], [28], [29], [30] and DIY videos [31], [32], [33], and in school physics textbooks [34]], and clubs devoted to constructing them [35], [36]. I think this activity should be represented. Also young readers may not be interested in the history, just crystal radios they can build or buy. --ChetvornoTALK 03:50, 28 July 2018 (UTC)
I added the Kristallradio.JPG image to the top for now. I have no problem including a children set. I'd like to see the later uses section expanded, with a link to the DX contests, for example, and maybe a gallery of its own. I also added a link to Arrow Electronics to the original picture. I remember when they were a single store on Cortlandt St. in Manhattan; they are now a Fortune 500 electronics distributer.]]

Sliding contact for tuning the inductor coil[edit]

Assuming the wire of the coil is insulated, how is a sliding contact able to work? ZFT (talk) 22:35, 20 January 2017 (UTC)

The wire used in such coils was enameled magnet wire, insulated by a thin layer of enamel. The enamel was not hard, but soft and viscous. The pressure of the spring contact sliding up and down the wire windings rubbed the enamel off at the point where the spring made contact, exposing the bare copper. Or possibly the manufacturer used a solvent to remove the enamel from the area on the surface of the windings where the contact was made.
BTW, such sliding contact coils were not used in quality radios because the sliding spring usually made contact with two or more of the fine wires at a time, thus short-circuiting them together. These turns acted as a short-circuited transformer winding, and the high current induced in them by the magnetic field caused power losses. Quality crystal sets used tapped coils, with the tap selected by a multiposition switch.--ChetvornoTALK 23:23, 20 January 2017 (UTC)
Yes, the coil is wound with enameled wire, but then the enamel is sanded off where the slider would connect. It's an easy operation, I did it as a kid, and it does not short the coils because the enamel is still intact between the turns. Variac construction is similar. Glrx (talk) 18:44, 26 January 2017 (UTC)
But the sliding contact tends to make contact with two adjacent turns, particularly with the fine wire used in radio tuning coils. --ChetvornoTALK 04:14, 28 July 2018 (UTC)
The wire on sliding variable inductors is not that thin; it must stand up to the abuse of the slider. The magnetic path / coupling in an air wound coil is poor. Look at the magnetic field generated by the entire coil, and then consider how much of that field is intercepted by the shorted coil. A single shorted turn will have little impact on a long solenoid. A variometer is a better solution. The tapped coils on crystal sets were commonly used for impedance matching rather than tuning; there would be too many taps unless one used one switch for 1 turn steps and another switch for 10 turn steps. I've never seen that done for crystal sets (it is done for early crystal-controlled VHF radios and also antenna matchers), but I'm sure somebody tried it.
The magnetic path is good in a Variac, so a shorted turn is death, and dead zones during rotation are undesirable. The Variac plays a subtle game with the resistance of the wiper so the turn isn't shorted but rather loaded:
Glrx (talk) 18:42, 28 July 2018 (UTC)

Removal of term 'cat whisker detector'[edit]

Editor LSMFT has removed all use of the term "cat whisker detector" from this article. I think this should be reverted.

The primitive semiconductor detector used in early crystal sets, which consisted of a fine wire touching a piece of crystal, was called either a "crystal detector" or "cat whisker detector" in literature of the time. The problem is, the semiconductor diode used in modern crystal sets is also sometimes called a "crystal detector", so a different term is needed to unambiguously distinguish the antique detectors. This article, following usage in most modern sources, called it a "cat whisker detector". LSMFT changed these terms to "crystal detector", claiming that 'cat whisker detector' was "not historically accurate"

LSMFT is wrong. In the first place, inventor Greenleaf Whittier Pickard named the device "cat whisker detector" when he invented it. Second, numerous wireless books of the time called it a "cat whisker detector". Third, even if they didn't, numerous modern reliable sources, technical and historical books and websites, call it a "cat whisker detector" (in modern sources the alternate form "cat's whisker detector" is usually used). This is probably the most common term for it today, used to distinguish it from various devices used later in the history of electronics, such as the germanium "crystal detectors" used in radar sets in WW2. Here are sources supporting these points:

--ChetvornoTALK 21:44, 1 May 2018 (UTC)

Hello, Editor Chetvorno. The 14:26 1 May revision of the article did NOT remove all use of the term "cat-whisker detector". It was included, WITH CITATIONS, in the "Crystal detector" section of the article. The term WAS NOT deleted from the article completely. Please carefully read the 14:26 1 May revision.

Thank you, LSMFT (talk) 17:39, 2 May 2018 (UTC)

No, not completely deleted, but only one use was left, buried in the "Crystal detector" section. The word "whisker" is used in two captions; readers are not going to know what that means if they don't happen to notice the term "cat whisker detector" in that section. I think the term "cat whisker detector" needs to be more widely used to distinguish the early detectors. The problem is the term "crystal detector" is overinclusive. --ChetvornoTALK 03:29, 3 May 2018 (UTC)
I have always heard them described as cat's whisker, I suppose maybe cat whisker, though usually soon indicating that it is actually metal and not a part of a cat. Gah4 (talk) 01:42, 12 March 2019 (UTC)

square law detector[edit]

The article has a red link to square law detector. The article on decibel, and especially the case, especially in optics, where detectors give an output voltage proportional to input power, caused me to look for a square-law detector article. Gah4 (talk) 03:01, 12 January 2019 (UTC)

I personally don't think it merits a separate article, because there is not really a separate type of detector called a "square law detector". Square law detection is just the mathematical regime that a rectifying detector like a diode operates in when a small-amplitude signal is applied to it; the DC output is dominated by the 2nd order term in its Taylor series so the output is proportional to the square of the signal amplitude (as opposed to at large amplitudes, at which the output is proportional to amplitude) [37]. But I think a section on "Square-law detection" explaining it should be added to Diode, Crystal detector or this article, and the article should link to it. --ChetvornoTALK 11:51, 12 January 2019 (UTC)
If the concept is mentioned in a few articles, then it is a separate idea and a separate article could be appropriate. In general, a nonlinear device can have its transcribing function expressed as a Taylor series. The even power terms generate a DC component. Semiconductor diodes have an exponential transcribing function. FETs are often characterized as drain current is the square of the gate-to-source voltage. So the article could cover the math and give examples. More specifically, the square law detector can function at frequencies where it is difficult to get gain, such as optical frequencies. The early crystal detectors were before vacuum tubes / radio frequency gain devices. Glrx (talk) 16:45, 12 January 2019 (UTC)
In optics, and much of quantum mechanics, there is much discussion of amplitude and intensity, where intensity is the absolute square (it might be complex) of the amplitude. In simpler system, power is proportional to the square of voltage or current, without dependence on Taylor polynomials, but just ohm's law. Absorbed optical power, which might generate a photocurrent, is proportional to the square of the amplitude of the EM (light) wave. But, yes, also in the case where of the second order Taylor series approximation. Gah4 (talk) 01:50, 12 March 2019 (UTC)

Did Bose's crystals function as bolometers?[edit]

Bose's galena detector from his 1901 patent.

@Gah4: The reference is Lee, p.5-6. Bose's detector experiments are described in Sarkar, p.296-297, 482, see also Crystal detector#Bose's experiments. Bose's galena detector consisted of a tiny galena crystal with a point contact, mounted either in a horn antenna or at the focus of a lens, with a DC current from a battery through it. Microwave radiation caused an increase in current through the crystal, which he detected with a galvanometer. The DC forward bias would have moved the operating point up on the diode IV curve, making it unlikely that the junction was functioning as a rectifier. Lee says the current change was due to change in resistance of the crystal due to heating, as in a bolometer. One indication that it was not functioning as a rectifier is that Bose noted the detector also responded to light and ultraviolet. However other sources either aren't aware of or don't make this distinction, and just say Bose invented the first 'semiconductor detector'. --ChetvornoTALK 00:36, 12 March 2019 (UTC)

Having not read the patent, I am only guessing, but it might be that is what Bose believed, semiconductor band structure not having been understood yet. Still without understanding band structure, germanium and then silicon point contact rectifiers were built, as well as I know Schottky barrier semiconductor rectifiers. These were used for WW2 radar to 10GHz. I suspect that heating/cooling can't go to 10GHz. If the patent does say it that way, the article should indicate that it is what the patent says, and not what modern semiconductor theory says. I will look at the reference. Gah4 (talk) 01:37, 12 March 2019 (UTC)
Looking at the reference, it seems to be quoting the patent. It continues to the not-shown page 9. As above, I think we can indicate we are describing what Bose believed. Gah4 (talk) 01:37, 12 March 2019 (UTC)
First, detecting radiation by its heating effect can work at any frequency; sensitive bolometer detectors are used up into the optical frequencies. On the other hand it seems doubtful whether a crude lead sulfide contact could function adequately as a Schottky rectifier at the 60 GHz frequencies Bose used; even modern Schottky diodes are limited to 50 GHz.
There are other primary sources on Bose's detectors besides his patent; Bose published detailed papers on his experiments, which are described in Emerson and Sarkar, p.296-297, 482. These say Bose didn't detect rectified current from his detectors; he passed a DC current through them and measured the change in resistivity. There is no indication that rectification occurred, and Bose did not claim it. That is what Lee [38] is saying. Semiconductor rectification was unknown until Greenleaf Whittier Pickard discovered it in 1904. I don't think we can say definitively that rectification NEVER occurred, but no reputable scientific source that I know of credits Bose with discovering rectification.
I think the article should simply say that Thomas Lee claims that Bose's detectors did not function as rectifiers as later crystal detectors did, but were thermal detectors. --ChetvornoTALK 19:29, 16 March 2019 (UTC)

Bose did not seek out a patent for this. Bose ignited gunpowder and rang a bell at a distance using electromagnetic waves proving that communication signals can be sent without using wires. He sent and received radio waves over distance but did not believe it to be useful. EyesoftheOperation (talk) 09:53, 18 October 2019 (UTC)

Nikolas Tesla invented the crystal radio![edit]

This article is mostly bogus. It is more or less someone's opinion and not factual. Nikolas tesla 1892 with a type of commercial patent later. Marconi came after Late 1890s- Bose ignited gunpowder and rang a bell at a distance using electromagnetic waves proving that communication signals can be sent without using wires. He never sought to patent anything radio related because he felt they were technically useless. Tesla contributed more in the field of frequency communication with derivatives based on the simple radio oscillator he had put together, than any other inventor at that time. EyesoftheOperation (talk) 09:52, 18 October 2019 (UTC)

@EyesoftheOperation: I agree the History section is not very good. However, the claim that Tesla invented the crystal radio is totally wrong, not found in any reliable histories but only WP:FRINGE WP:PSEUDOSCIENCE websites. Tesla had nothing to do with the development of crystal detectors (see the thoroughly sourced Crystal detector#History section). Tesla invented the inductively coupled resonant transformer in 1891, which was used in both transmitters and receivers in the famous "four-circuit" radio design, later employed in crystal receivers throughout the early 20th century. But it was Oliver Lodge, John Stone Stone and Karl Ferdinand Braun that first applied the circuit to radio receivers and transmitters. Braun and Marconi received the 1909 Nobel physics prize for this. Tesla only used it in his wireless power experiments, and did not develop a radio system until after these others;[39][40][41][42] and see the "Tuning" sections in the thoroughly sourced Radio receiver#History and Spark-gap transmitter#History. The electronic oscillator was invented by Elihu Thomson, William Duddell, Valdemar Poulsen, Edwin Armstrong, and Alexander Meissner; see the thoroughly-sourced Electronic oscillator#History section. --ChetvornoTALK 09:12, 20 October 2019 (UTC)