MDS975 logo
Crystal Sets
MDS975 logo
Home  |  Contact  |  Site Map  |  Radio Stations & Memorabilia Amateur Radio

Crystal Sets Introduction  |  Resistor & Capacitor Conversion Tables

Crystal Sets (Part1)  |  Build Your Own Crystal Set  (Part 2)

Spider's Web Crystal Set  (Part 3)  |  Crystal Set By Kenneth Rankin  (Part 4)  Crystal Radio Links


Picture 1 - The Complete Experimental Crystal Set

THE POPULARITY of the crystal radio arises from its simplicity, and the fact that it needs no power supply. The circuit here allows for easy experiments with tuning, aerial and diode coupling, and frequency coverage. Wrong connections can cause no damage to any components.
A Crystal Set is more often than not used for the reception of medium and long wave radio, but short wave reception is also quite feasible.  It will normally be  possible to receive some of the stronger international radio stations.

This is adapted from an article that appeared in the 1970's in Everyday Electronics, and gave me almost endless hours of fun!

The basic circuit is shown in Picture 2 below.  The coil L1 can be air cored, or have a ferrite rod placed in its winding.  The variable capacitor C1, in conjunction with aerial-earth capacitance, tunes the circuit to resonate with the wanted radio station frequency.  The diode D1 "detects" or demodulates the radio signal so that the programme is heard in the earpiece.
This basic circuit can be modified in various ways to obtain better performance. 


As most constructors will be using a Crystal Earpice to listen to the crystal set it is essential that a 47k Ohm resistor is connected across the earphone terminals (TB1/1 and TB1/2 in the diagram), i.e. in parallel with the earphone, otherwise results will be very quiet.

A High Impedance headset of 20k Ohms (20,000 Ohms) may give even better results, but these are very difficult to obtain , so unless you happen to already own such a headset the Crystal Earphone with 47k resistor will be the only option.  An ordinary magnetic earpiece or Walkman headphones will not work with a crystal set.


Construction is of a 'breadboard' type using a wooden board of about 165 x 130 mm.  A 12-way block connector, TB1,  is used to connected together the components and this is screwed onto the wooden    board.   The use of a block connector provides an easy method of connecting the components together and then subsequently rearranging them as the experiments progress.

Tuning capacitor C1 is screwed to a bracket made of some scrap metal which is then also screwed firmly down to the baseboard, see Picture 1 above.  Thin plywood screwed to the front edge of the baseboard would also provide a suitable method of fixing the tuning capacitor to the base.  A knob with pointer is fitted to C1, and a scale is drawn and fitted behind this.

Except for C1, all connections are made by the terminals of the 12-way terminal block as shown in Picture 4.  Loosen the screws with a small screwdriver, insert the bared ends of the wires, and tighten the screws. The various locations on the terminal block, TB1, are also shown in the circuit diagram, Picture 2.


Crystal receivers need a long wire aerial preferably strung outside and about 25m long, or as long as is possible to install.  If this is outside it should be high and clear of earthed objects as this will improve performance.
An earth is absolutely essential for a crystal set to work properly.  The earth lead can be run to an earth rod or spike that is buried to a depth of about 1 meter into damp soil.  Or it may be soldered to a bare metal can which is buried in damp soil. 

It is feasible, though not recommended, that the earth lead can be connected to the earthing terminal of a hi-fi system or even to the bare metal case of a personal computer that is plugged into an earthed mains outlet, but is switched OFF.

Stranded, insulated wire, or purpose made aerial wire can be used for the aerial and earth leads.

Picture 2 - The Basic Circuit
Picture 3 - Photo Of The General Layout

INDUCTORS (The Tuning Coils)

The following four coils are suggested for initial use as L1 :

Coil 1:  Make a thin card tube to slide on a 10mm diameter ferrite rod, and on this tube wind about 105 turns of 32 s.w.g. enamelled copper wire, side by side.  Secure ends with sticky tape.
Coil 2:  Make a similar coil to to coil 1 having about 15 turns of 24 s.w.g. enamelled wire on the card tube.  Loops of cotton will help hold the ends in place.
Coil 3: Wind 9 turns of 20 s.w.g. bare tinned copper wire on an object about 20mm in diameter. Remove and stretch to separate the turns, to obtain a coil about 25mm long.
Coil 4: Make a similar coil to coil 3, but with 5 turns.
The Ferrite Rod

It will be necessary to have a ferrite rod of about 60mm to 75mm long available.  Coils 1 and 2 will provide reception of medium wave and the longer short wave bands.  Coil 3 should cover about 3 - 10MHz shortwave with the ferrite placed in it, or about 6 - 18MHz with the ferrite rod removed.  Coil 4 should cover about 6 -13MHz with the rod in, and about 9 - 20MHz without the' rod.

It will be noted that as the ferrite rod is inserted, any particular signal has to be re-tuned by opening Cl. This arises because the ferrite increases the inductance of the winding, so less parallel capacitance is needed for the same resonant frequency.


Tune in a m.w. transmission using coil 1 which gives good headphone volume. Place a microammeter or multi-range meter on a sensitive range in series with the headphones.  A reading of 50-100uA or more may be obtained, depending on aerial, earth, earphone resistance and resistor value, coil and detector efficiency and strength of signals at your locality.
Placing the ferrite rod in the coil and re-tuning should boost the meter reading to some extent. Surplus or other detector diodes can be tried by substituting them in turn and noting the meter reading. Improvements to the aerial (or earth) will also show up as a rise in meter reading.
If experimenting with a crystal earpiece, which gives no direct current circuit, the meter may be clipped across the phone leads, i.e. D1 cathode to earth.

Picture 4 - Baseboard Layout Of The Crystal Set


The aerial loads the tuned circuit heavily when connected directly to the top of the tuned circuit, as in Picture 2.  This damps the tuning action and it can be found that stations spread out all over the dial, which is unsatisfactory.

The series capacitor, C2 connected in Picture 5(a) reduces the loading and thus improves the sharpness of the tuning.  A variable or pre-set capacitor of about 250pF maximum is most suitable. for this role, though it is possible to experiment with a variety of fixed value capacitors in this range also.
Connecting the aerial to a tapping on the coil, as in Picture 5 (b) also sharpens tuning.  It may also increase volume. Try about 2 turns from earth for coil 4, or 4 turns from earth for coil 3.

Another method is to have a coupling primary, as in Picture 5 (c).  This consists of a second coil, with about one third the turns of the original wound on top of the existing coil.
You can even combine these methods to find what arrangement best suits the aerial in use.
The diode can be disconnected from the end of L1 and taken to a spare position on TB1 for example location TB1/9.  You can then run a flying-lead fitted with a crocodile clip from this position, connecting it to various tappings on the coil as required as in  Picture 5 (d).  This method also reduces loading on the tuned circuit.

Coils with spaced turns of bare wire are readily tapped.  For other coils, small loops can be made every ten turns or so, and crocodile clips can be attached to these when selecting tapping points.

Picture 5 - Alternative Methods Of Aerial Coupling


For shortwave reception, a good efficient outdoor aerial is certainly recommended.  Evening listening in the region around 5 - 9MHz in  often proves to be the most fruitful.

Since there is no amplification, as with a valve or transistor receiver, certain frequencies will seem to be completely dead at particular times of day.  So if the crystal receiver works satisfactorily on medium wave and longwave, but no shortwave signals are heard, check again in the evening, or after dark, when conditions are different.


365pF or 500pF Air Spaced Tuning Capacitor
OA47, IN34, OA81, OA90, OA91, IN94 or similar point contact small signal Germanium Diode
* The OA47 will be of particular interest since it has the lowest forward bias voltage of any of these diodes which will make the crystal set somewhat more sensitive and therefore louder. The US equivalent of the British OA47 is the IN34.
High Impedance Headphones (20,000 Ohms)
or Crystal Earphone
12-Way Plastic Screw Block Terminal
Also Required:
47 k Ohm Resistor for Crystal Earphone:
Enamelled Copper Wire: 32 and 24 s.w.g. for L1:  20 s.w.g. tinned wire for L1:  Ferrite Rod 10mm diameter x 75 mm long: 25m of wire for aerial:  Wire and rod or spike etc for earth:  Wood for base e.g. 10mm x165mm x 130mm:  Scrap of metal of thin plywood for C1 bracket/front panel:  Knob:  Crocodile clip(s)

Adapted from an article in Everyday Electronics magazine, November 1981, By F.G. Rayer.


Circuit Diagram

1/   Using old coils from old bulb radio for MW and LW band. Though it would be straightforward to wind the coils - one for Long Wave, one for Medium Wave and a coupling coil.  Variable capacitor is 2 x 500pF only one half is used: 500pF.  For the crystal earphone a resistor of about 82k ohm in parallel is required.  This set also uses two Ge diodes as a multiplier in the quest for for higher audio signal output.


2. If signals are not strong signal in your location, then the above circuit design can be considered.  A simple transistor amplifier is used. A variable resistor M22 is used for better sensitivity which can be adjusted for poor signals.  This crystal radio is aversion from cca 1960 - 1970 y.

                        Set Photo
Rat's finished Crystal Set with additional amplification - very neat!

'Minilabs' Crystal Radio

BELOW:  Ian Tomlinson kindly sent in a photograph of the box that contained the kit for his John Adams Toys 'Minilabs' Crystal Radio. 

It is a very simple circuit consisting of the coil (inductor) with a sliding contact that provides variable tapping points, a diode and crystal earphone.  All that is added is the aerial and earth. There is no variable tuning capacitor for simplicity and to keep costs down.

The coil provides the inductance required for tuning into a certain frequency (wavelength).  These days a variable "tuning" capacitor is normally wired in parallel across the inductance (coil) in order to vary the resonance of the tuned circuit and therefore enable to easily tune into various transmitters on different frequencies.  This crystal is tuned varying the number of turns on the coil (ie varying the inductance) by tapping off at different points using the sliding contact ("ball").

The crystal earpiece, or high Z headphone, is connected between the output of the detector diode (the other end from the coil) and earth.  The volume from a crystal earpiece may be considerably improved by connecting a resistor of - somewhere between - 4.7 k and 47k ohms in parallel with the earpiece.  A crystal earpiece cannot directly allow current to flow through it and the parallel resistor therefore allows current to better flow through the circuit.

'Minilabs' crystal set
'Minilabs' crystal set by John Adams Toys

A discussion on configurations for Crystal Sets by Felix Scerri VK4FUQ

This discussion, by Felix Scerri VK4FUQ, was posted at this address which no longer appears on the web (error 404) so here it is reproduced:

Crystal Set design is one of my passions closely allied with my obsession for audio and high fidelity.

My main interest in crystal sets, apart from the wonder of a radio receiver that does not require a power source, is the potential excellence of the recovered audio quality from normal AM broadcast stations.

Personally, it is one of my great laments that most people have never heard how good wideband AM can sound. A high performance crystal set or similar TRF approach is, in my opinion,the only way to do it. There are a few people around who have heard the audible results of my efforts,and can only agree.

I have often wondered,given the ultimate simplicity of the crystal set, being essentially a tuned circuit,a diode detector and some form of output device, what it takes to achieve optimum performance. What follows are my thoughts on the matter.

Crystal Set optimisation, is in my opinion, all about reduction of circuit losses. Essentially this means high "Q" tuned circuits and high quality detectors. Efficient output devices also help too. But as we will see, there are some trade-offs required as well. A high "Q" tuned circuit is always benefical, as a high "Q" tuned circuit has lowest RF losses,highest potential selectivity,and highest voltage at resonance, which is very useful for the diode being fed from the tuned circuit. Variable capacitors, even the "modern" miniature variable capacitors (although the older air dielectric units, as used in old valve receivers are more desirable) for various reasons,are generally quite efficient, and a higher "Q" coil will produce the most worthwhile improvements.The best (highest "Q") coils are wound with "Litz" wire, which is a multistranded woven wire with all strands insulated from each other. The performance of Litz wire wound coils is spectacular, unfortunately, although I know Litz wire is still being made, from personal experience, it is VERY rare in Australia.

Efficient coil design can be quite complex and all my coils are wound on ferrite rods. There seems to be,at least for ordinary single wire windings (close wound), an optimum wire thickness for optimum coil "Q". I have determined .315 mm winding wire to be about optimum for simple (single wire) coils on ferrite rods. Thicker wire is NOT better, believe it or not.

Lacking Litz wire, an interesting winding approach I have developed is to use two slightly thinner wires wound as a bifilar winding connected together at the beginning and end of the coil, yields considerably higher "Q" compared to a simple single wire winding. I have found 0.25 mm winding wire optimum in this application.

Whilst high "Q" coils are beneficial from the RF point of view, there is a possible downside. If one is interested in maximum selectivity and sensitivity, there is no problem, but remember highest "Q" results in a narrowed audio band-width as a simple consequence of band-width. For high fidelity applications this could be a disadvantage under some circumstances, although there are clever ways around this.

Regardless of ultimate coil "Q", selectivity is a major issue with crystal sets generally. Here another trade-off is evident. For the maximum voltage into the diode, connecting the diode to the high impedance end of the coil (i.e. the top) yields the greatest voltage but the selectivity is usually terrible, because of severe "loading" by the diode circuit. For this reason, tapping well own the coil improves selectivity at the expense of signal volume (reduced voltage). Once again there are ways around this. As described in my "Double Tuned Crystal Set Tuner" article in "Amateur Radio" magazine, March 2002, the use of two separately tuned coupled resonant circuits allows top connection into the diode without compromising overall selectivity, thanks to the use of a second tuned circuit which is fed from the external antenna. The whole network forms a double tuned input bandpass filter and in practice this approach works very well. For single coil crystal sets I recommend the use of an un-tuned "antenna" winding adjacent to the "hot" end of the main coil, preferably adjustable (old paper reels from sewing cotton threads are ideal). This allows the degree of coupling to be optimised under actual listening conditions. The double tuned set up is best, yielding superb selectivity, but the un-tuned antenna coil arrangement also works quite well, especially if the diode is tapped well down the main coil.Tapping halfway works well.

The other method of performance improvement involves the use of the most effective detector system possible. Here things get very interesting. In fact the temptation is to use more complex circuitry, but that gets away from the charming simplicity of the crystal set. As an example, my own crystal set tuner has at times mutated into a TRF tuner complete with FET RF preamplifiers, active(powered) detectors and other enhancements. These modifications do work well, but loses the simplicity of a basic crystal set. In actuality, a simple diode detector can work extremely well, subject to some qualification. Diodes like to work with a reasonable level of RF input voltage. Audio distortion can result under conditions of low signal level, due to diode transfer curve non linearity and other factors, such as the widespread use of broadcast station "processing". The actual type of diode makes a difference. The 1N34A germanium diode is very popular for crystal set use, although in my experience just about ANY germanium diode will work, although it is worth trying different specimens. Some are definitely better than others. Even from a pack of twenty 1N34A's from the same source, some were definitely better than others. Measuring the average value of rectified output voltage across the diode load resistor will show which diodes are best. By the way, I regard a diode load resistor as being mandatory. I find a value of about 47K about right, especially if a crystal earpiece is being used or the crystal set is being used as a tuner feeding an audio pre-amplifier and following amplifier. If using high impedance magnetic type headphones, the headphones provide the diode DC load.

Another type of diode that is very interesting, is the hot carrier diode. There seem to be a lot of different hot carrier diodes around these days. There are even hot carrier diodes now being sold as "germanium diode equivalents". I have tried them and they do work acceptably well, but they are not quite as good as genuine germanium diodes such as the 1N34A. Typical UHF mixer hot carrier diodes, such as the 1N5711 will not work well in crystal set service simply because their "turn on voltage"is too high, similar to silicon diodes such as the 1N4148/914 series, which require a lot of RF input to function adequately as RF detectors, however a simple technique can be used to turn hot carrier diodes such as the 1N5711 into superlative detectors.

I guess we are cheating a little, because the technique is to use a little voltage bias supplied via a 1.5v battery, through a simple potentiometer voltage divider arrangement, with capacitor (for DC isolation) fed into the diode from the tuned circuit. With applied adjustable bias, I find the 1N5711 diodes absolutely superlative detectors under ANY signal strength conditions. I find the detection quality also superlative, with a clarity and low noise profile unmatched by any other diode arrangement. In my opinion, hot carrier diodes, running bias,are the best detectors overall.

Regarding other detector arrangements, the diode "voltage doubler" is often recommended, however my own experiments with the doubler arrangement have been inconclusive and slightly disappointing overall. I have found no real advantage in their use over a simple (one) diode detector, believe it or not.

Yes, they do work, but they're nothing special, at least in my opinion.

Any comments on this general subject of crystal set optimisation would be welcome.

73's Felix Scerri VK4FUQ.
22nd July 2002


IMPROVED VERSION OF THE ABOVE CONCEPT !! New update from Felix Scerri February 2010:

New 'two FET  infinite impedance AM detector'

I've developed a new version of my old favourite FET 'infinite impedance' AM detector that I think sounds very nice.  I include a short audio of one of our local AM stations.  I picked this station as it is my reference 'torture test AM station' as they run very heavy 'processing' which normally sounds yuck with all my other (diode and non diode) detectors!  However it's quite clean with this detector.  What do you reckon?  I'll do up a circuit if you'd like to feature it in your TRF radio section.  A general draft article follows.
'A favourite non diode based AM detector that I've built and used many times over the years is the FET based infinite impedance detector, offering very good general AM detector performance, especially under weak RF signal conditions where diode based detectors do not perform well, especially in terms of audio distortion.
However one of the slightly strange things I've noticed about the simple FET based infinite impedance detector is the variable audio quality noted, even when using the same type of FET.  Some I've built have sounded good and others slightly fuzzy when used with an audio preamp and fed into a high quality audio system.  I've been giving this a considerable bit of thought of late and I've wondered if the audio distortion might be a result not necessarily of the detection process itself, but the FET stage in its guise as a 'source follower' audio stage which essentially, it is.
I have long been aware that as a simple audio buffer stage, the FET based 'source follower' can exhibit a considerable amount of audio distortion, and a technique I've long used to greatly reduce this audio distortion is to use a second FET in the source lead of the first FET as a 'constant current source' which serves to 'linearise' and greatly reduce audio distortion in the buffer stage overall.  So, to test the theory I built a simple one FET infinite impedance AM detector which worked well, but with just a hint of audio 'fuzziness' on received AM stations.  So I added a second FET in the source lead of the first FET wired as a constant current source, taking the output from the source of the first RF detector FET and the source resistor and RF bypass capacitor off the source lead of the second FET 'constant current source'.  The result, totally clean audio!  The theory seems proved!  I call this modified detector the 'Two FET infinite impedance detector'

))) Here's what is sounds like - click to play the audio file (((

Here is the circuit diagram:
Felix Scerri - high quality 2 FET AM
Felix Scerri's high quality 2 FET AM
                          detectorThis detector has been a real eye opener for me in terms of its excellent performance, especially considering its circuit simplicity.  Indeed  in the past I have designed other more complex FET based infinite impedance circuits that do not quite work as well in practical terms as this latest circuit, at least according to my well calibrated ears!

I do not have access to any precise test equipment but my well calibrated ears tell me this 'two FET infinite impedance detector' is a beauty, surpassing practically every other AM detector I've built even at low RF input, and that's rather a impressive claim and the audio quality when used as an AM tuner feeding a high quality audio system is quite remarkable.  Possibly the best thing about this detector is its excellent performance under weak signal conditions.  Diode based detectors also work beautifully, but the use of an RF stage to ensure detection over a linear portion of the diode's curve is mandatory!  This compound infinite impedance detector works beautifully on the sniff of a useable RF signal. 

Just add a high Q tuned circuit and that's it! 

Felix Scerri

A better FET for the 'basic' Infinite Impedance Detector:

Quite recently by accident, I've realised the MPF102 FET that I've long used in my FET based infinite impedance detectors is possibly not the best FET to use.  This was the reason why I developed the 'two FET' infinite impedance detector some time ago which works very well.
However I've found the choice of a more suitable FET works beautifully in the basic FET based infinite impedance detector circuit, which has appeared for many years in many editions of the ARRL Handbook.
I use the 2N5457 and others of the same 'family' may be equally suitable, but I haven't tried them!  However with a 2N5457 in place of an MPF102, the basic infinite impedance detector has became my AM detector of choice.  It works beautifully even at low signal input with lovely and clean low distortion audio along with a very high input impedance for good tuning selectivity.  It's a beauty!  The basic generic circuit is attached, courtesy of Rod Elliott's ESP website. 

73 Felix VK4FUQ
10 / 02 / 2012.
                            Infinite Impedance Detector
The basic generic circuit is attached, courtesy of Rod Elliott's ESP website
Felix Scerri  VK4FUQ

As often happens with me, my renewed interest in FET based 'infinite impedance detectors' of late has led to some interesting new research and I may have considerably improved the 'two FET infinite impedance detector' as a result.
My research suggests that although the use of a CCS (constant current source) reduces audio distortion in an audio stage, the value of the 'source resistor' in the CCS stage is somewhat critical for best results.
By using a potentiometer in lieu of a fixed resistor I have found that a resistance value of around 470 kohms cleaned up all overall audio distortion.  I used an MPF102 as the CCS in this circuit.  An interesting and worthwhile little circuit refinement. 

73 Felix VK4FUQ
21 / 02 / 2012.

A Minimum Component Count High Quality AM Detector

I was generally messing around with various circuit ideas and I came up with this AM detector circuit, a simple diode detector along with a FET stage.  It was an attempt to provide good performance along with minimum number of components.  Actually I've been pleasantly surprised at the excellent level of general performance and the best of all, it sounds great!
The circuit is quite conventional being a BAT 46 diode detector feeding an MPF102 FET buffer/ common source amplifier stage.  I would ordinarily use a FET source buffer stage in this application, but opted to use a simple low gain FET 'common source' amplifier stage instead, with excellent results.  I also used a BAT 46 Schottky diode instead of an ordinary germanium signal diode.
This was done for several reasons.  Firstly, germanium diodes are now very hard to find but in any case these 'germanium diode equivalent' Schottky diodes are actually a superior diode, having very low noise, almost zero back leakage and an essentially complete absence of carrier storage effects and very good weak signal sensitivity.  I call these diodes high fidelity diodes as they sound wonderful as RF detectors.

                              Component Count AM Detector by Felix
Circuit Diagram of the Minimum Component Count AM Detector by Felix Scerri

Minimum Component Count AM Detector
                              by Felix Scerri
Minimum Component Count AM Detector by Felix Scerri

Minimum Component Count AM Detector
                              by Felix Scerri
A Closer View

The high impedance of the FET's gate circuit is perfect for optimal buffering of the diode detector, something very important for good low distortion diode detector performance.  Apart from providing slight voltage gain, the use of the common source FET amplifier is a new idea, as this prevents the possibility of incidental RF rectification occurring in a FET source follower stage, which can happen.  A 1 uf plastic film capacitor may be added in series with the audio 'hot' output lead to block the DC offset out of the FET drain, if required.
Despite no additional RF stage ahead of the diode, audio quality on even relatively weak RF strength stations is actually very good, and of course the audio quality will be even better with increasing RF signal strength, something which will also increase the audio output level.  Just on this, for a long time I was somewhat negative regarding diode detectors, as one AM station locally (the strongest one) was always distorted when using a diode detector.
I strongly suspected a transmitter fault, but my complaints were ignored, until one day some time ago when all audio distortion suddenly disappeared!  Nothing was ever 'said', but I realised that my suspicions of a long standing transmitter fault were correct, after all!

73 Felix
02 March 2012

Voltage Doubler Detector

This is an AM detector circuit that I've long known about and which worked ok, but never seemed to work as well as it should have.  However I spent some time late last night trying to optimise the circuit, with some success.
It is a curious circuit being essentially a 'voltage doubler' originally developed for power supply applications, and its use as an AM detector is hard to analyse!  It seems that the component values in the circuit are somewhat critical for good performance and if not, the performance is rather 'ordinary'.  The circuit that I eventually came up with uses a 150 picofarad 'input capacitor' with a 150 kohm 'load' resistor and loaded into a FET common source voltage amplifier stage (as previously described) through a coupling capacitor with a 1 Mohm input resistance.

                              Doubler Detector

With these circuit values, it all works 'quite well'.  Give it a go!  It's an interesting AM detector with quite good 'sensitivity' and clean audio quality, and it seems to work well at low signal levels. 

73 Felix

16th April 2012.

Simple AM detectors:  What works best?  A practical experimenters viewpoint.
I have written a lot about simple AM detectors for use as tuners for feeding into an audio amplifier, and it has been a long time interest.  These days I use either diode based or 'infinite impedance' types of AM detectors.  In this location our 'local' AM stations are quite distant and are therefore quite weak in terms of signal strength.
As such I find infinite impedance detectors based on field effect transistors give consistently better results for tuner applications due to their lower apparent overall detector distortion.  Diode based detectors are quite 'fussy' as they require both optimal output buffering (AC/DC ratio) and an 'adequate' (beyond the diode knee) level of RF signal injection.
Diode based detectors will happily 'detect' at very low signal levels, however the (inevitable) audio distortion that results, can be extremely irritating to the ear!  Under these conditions I find infinite impedance detectors (even without additional RF preamplification and subject to individual FET characteristics), generally sound 'cleaner' and more pleasant to the ear.
FET's of course require a power source for operation whereas diodes are passive (un-powered) detectors (most of the time), however this is of no real advantage in a tuner application as an 'active' audio amplifier stage will generally be required anyway for audio level boosting, buffering etc.
In the end it will come down to a consideration of prevailing RF signal levels and other related circuit considerations at one's location.  If local RF levels are strong, a well designed diode detector will give excellent results.  If not, an 'infinite impedance' type of detector is most likely the better option unless one goes towards the option of additional RF preamplification prior to the diode detector. 

73 Felix

17th April 2012.  

ANOTHER DETECTOR by Felix Scerri !

I’ve tried diode based ‘voltage doubler’ (or more correctly ‘diode integrator’) AM detectors before with indifferent results, however the other day, just trying a few ideas I came up with this version that works rather well, with low audio distortion, high audio output and really ‘nice’ audio quality and the best of all, it seems to work very at very low RF input level. 

The two diode ‘voltage doubler’ detector using two BAT 46 silicon schottky diodes feed directly into a MPF102 source follower stage set at 1 Megohm input resistance.  The ‘input’ capacitor feeding the diodes from a tuned circuit is 68 picofarads.

I have the simple RF filter right on the output of the FET stage.  In that respect this circuit is vaguely similar to the old ‘Selstead-Smith’ valve AM detector of the past.  An interesting one!  I am very happy with its general performance.  Regards, Felix vk4fuq  11/04/2013
                              Scerri Detector

Felix Scerri VK4FUQ


G’day all, readers may recall the two FET infinite impedance detector I developed some time ago.  That circuit worked well, but some samples of the MPF102 regretfully produced distorted output.  However a recent discovery has resulted in an improved version that has truly exemplary performance. 

The MPF102  is in all honesty a device essentially designed for VHF applications, but is often used in simple audio applications albeit with occasionally indifferent results, due to general device parameter ‘spread’.  More or less by empirical trial and error I have found that a simple change in source resistor value to 100 k (from a much lower value), pretty well fixes everything. 

In a FET source follower circuit this is interesting as the output impedance is actually a lot lower than 100 k due to the action of the FET’s transconductance.  It is similar to the action of a bipolar transistor emitter follower circuit.  In fact a simple infinite impedance detector with a 100 k source resistor actually works very well and will be good enough for many applications, however the addition of a second FET as a constant current source does markedly reduce overall audio distortion in the stage overall, and produces very clean and low distortion audio quality and also reduces the output impedance considerably (good for tuner applications), so take your pick! 

Regards, Felix vk4fuq 02/06/2013.

A simple modification to the basic FET 'infinite impedance' AM detector that dramatically improves performance.

Infinite impedance detector -
                                  Felix Scerri January 2014

G’day all, over the Christmas break just messing around I worked out (mostly by accident), a simple modification for the simple FET based ‘infinite impedance’ detector circuit that dramatically improves weak signal performance and also greatly reduces audio distortion. 

Essentially by the addition of another FET ‘buffer’ stage, another source follower, capacitively coupled from the first detector stage.  The circuit is actually a simplified version of the circuit that I described in this link,  (figure 6)  and testing the two head to head, they both sound superb and the simpler version is actually somewhat easier to build.  I cannot get over how low distortion and ‘nice’ the recovered audio sounds.  It is a joy to listen to! 

Regards, Felix vk4fuq 12/01/2014.

Diodes for ‘weak signal’ crystal set applications.

As I am primarily interested in using crystal sets as AM tuners for feeding into a preamplifier/amplifier and loudspeakers, the actual type of diode can be relatively critical.  In a strong signal area, not so much, but in a weak signal area such as where I reside, definitely.  A diode with a good ‘square law’ performance (the area below the ‘knee’ in the diode curve), generally results in much cleaner and lower distortion than other good performing diodes, and believe me diode distortion under weak conditions, even with optimised diode ‘buffering’ used is very nasty sounding to the ear!

Testing many, many diodes in actual working crystal sets and listening critically to the audio output, it seems to me that the best diodes to use under weak signal conditions are the so called ‘gold bonded’ germanium diodes.  I have sampled many different gold bonded germanium diodes and they all work well in this specific application, although sometimes the rectified output voltage may not be as high as other germanium or silicon schottky diodes, however the audio quality is much cleaner and shows much less apparent distortion! 

I have tried OA5, OA47, IN141 and several ‘CG’ gold bonded germanium diodes with consistently excellent results.  Other ‘ordinary’ germanium diodes may also work well under weak signal conditions, but they will need to be tested individually to check actual performance in a working circuit.  One thing that I have noticed about germanium diodes is that due the ‘point contact’ nature of their construction, even diodes of the same specific type can exhibit rather different levels of performance! 

The otherwise very good BAT46 silicon schottky diode works extremely well at good RF input level but not so well at weak RF input, especially when heavy broadcast ‘processing’ (commonplace these days), is used.  New ‘gold bonded’ germanium diodes are probably no longer made although I am aware that they can be purchased through vendors over the internet.  Apart from that, they may be found in old gear.  As stated earlier, ordinary germanium diodes may be quite good but will need to be checked individually.  Diodes are complex things!

Felix (vk4fuq). 29/01/2014.

* * * * * * * * * * * * * *

Just when I thought that I’ve tried everything I realised that I’ve not tried a voltage doubler (diode integrator) detector with an RF stage in front.  It is an unfortunate fact that all diode detectors ‘need’ a good RF injection level (and buffering) or audio distortion becomes ‘bad’ (an understatement!) 

Consider yourself very lucky if you live in a strong signal area! (unfortunately I don’t).  Anyway I quickly built a couple of prototypes which worked ok but not as well as hoped.  Initially I tried using a source follower buffer (no voltage gain), but converting it to a common source amplifier (with appreciable voltage gain), seemed no better, which was strange.

After staring at my prototype for what seemed like forever with my magnifiers on, I (finally) realised my mistake.  A common source stage takes its output off the ‘drain’ terminal of the FET, not the ‘source’ terminal as it does with the FET source follower buffer stage.  I shifted one wire, and all worked as expected and it sounds fantastic!  All diode detectors (one diode envelope detectors, doubler detectors etc), need good RF input and a simple untuned FET RF gain stage works very well.  As to the sound, it sounds great.

73 vk4fuq 10/03/2014.

Felix Scerri - FET Crystal Set with
                              Voltage Doubler
Felix Scerri's FET Crystal Set with Voltage Doubler
A 'nice sounding' AM detector - update October 2015

Looking through these pages the other day I was slightly shocked to realize that I’ve completely forgotten about some of the circuits that I’ve contributed in earlier days!  They all work pretty well, however in my advancing older age and a little like my tastes in high fidelity generally, I’m starting to show a particular preference for ‘nice sounding’ bits of audio gear.

I guess that this also means low noise/low distortion too, although with AM broadcasting, at least in this country and probably elsewhere in the world too, the very common use of broadcasting ‘processing’ tends to make it hard for AM detectors generally, often resulting in a ‘hard/compressed’ sort of sound although still low in distortion, is not ‘nice sounding’, if that makes any sense!

Well of all the AM detectors that I’ve tried and/or developed, only one sounds ‘nice’ when confronted by heavy broadcast processing and that is an AM detector ‘based’ on a voltage doubler/diode integrator circuit (similar to the above circuit, actually the 68 pf capacitor should be changed to .1 uf, for slightly greater output).

It has taken me a very long time (years) to ‘optimise’ this circuit, but as it presently stands this is my favourite ‘nice sounding’ AM  detector, and gets most use for general high quality listening on the AM broadcast band.  It sounds really good!  It somewhat reminds me of an old OA47 gold bonded diode detector  from years ago before such abhorrent ‘processing’ became commonplace! 

Ah yes the OA47….now that is a lovely sounding detector diode! 

Regards, Felix vk4fuq. 17 / 10 / 2015.     

* * * * * * * * * * * * * *  

That's it for crystal sets.  I hope you try building one, it's easy and great fun!

See some useful links below....


No AM radio stations or transmitters in your locality or country? AM Medium Wave
                          Transmitter from Vintage Components

Has your local medium wave broadcast station closed or been moved to VHF/FM or Digital? Don't worry. You can still build and experiment with crystal sets and TRF radios by also buying or even building a simple low power AM transmitter. So, not only can you use your crystal sets but you can also run your own radio station that can be heard in and around your home - playing the music or programmes that you want to hear!

SSTRAN AMT3000 Superb high fidelity medium wave AM transmitter kits from SSTRAN. Versions available for 10kHz spacing in the Americas (AMT3000 or AMT3000-SM) and 9kHz spacing in Europe and other areas (AMT3000-9 and AMT3000-9SM). Superb audio quality and a great and well designed little kit to build:

SSTRAN AMT3000 low power AM medium wave

Other AM transmitters available:

Spitfire & Metzo Complete, high quality ready built medium wave AM Transmitters from Vintage Components:  Vintage Components offer a choice of the high quality Spitfire and Metzo transmitters:

SPITFIRE AM Medium Wave Transmitter with 100 milliwatt RF output power:
Spitfire AM transmitter from
                                Vintage Components

METZO AM Medium Wave Transmitter with built in compressor:
Metzo AM Transmitter fromVintage

AM88 LP  A basic AM transmitter kit from North County Radio.

Component Suppliers:  Links to electronic component suppliers here

Crystal Sets (Part1)  |  Build Your Own Crystal Set  (Part 2)

Spider's Web Crystal Set  (Part 3)  |  Crystal Set By Kenneth Rankin  (Part 4)   |  Crystal Radio Links

Home  |  Contact  |  Site Map  |  Radio Stations & Memorabilia   |  Amateur Radio

Crystal Sets Introduction  |  Resistor & Capacitor Conversion Tables © 2004 - 2015