I know that germanium diodes are trivial to find online, but as this is for a demonstration I'd rather not spend $6-7+ on shipping for a single 5 cent part for a project that's academic in exercise anyway. RadioShack has proven stereotypically useless in stocking germaniums.

I do have available to me jellybean components like the 741 and 324. I also have several varieties of N & P-channel FETs as well as BJTs. Is there some small and straightforward circuit I can use to emulate the low-voltage drop behavior of a germanium diode in a low (microwatts?) power application?

  • 5
    \$\begingroup\$ Schottky diodes have about a 0.25V drop. \$\endgroup\$
    – Kaz
    Sep 17, 2013 at 19:21
  • \$\begingroup\$ Depends on the exact application. Maybe an active rectifier around an opamp is an option? sound.westhost.com/appnotes/an001.htm and niu.edu/~mfortner/labelec/lect/p575_01b.pdf \$\endgroup\$
    – jippie
    Sep 17, 2013 at 19:25
  • 4
    \$\begingroup\$ On an historical note a 'crystal set detector' was a crystal of galena with a sharp wire probe (the whisker) - the germanium diode replaced this in later sets (after WW2). The 'practical whisker' was developed by a G.W. Pickard who tested over 30000 combinations of minerals and wire setups, how enterprising. His first detector, using a SILICON crystal was patented in 1906. \$\endgroup\$ Sep 17, 2013 at 22:00
  • \$\begingroup\$ @Kaz: but Schottky diodes have a quite high reverse leakage current what makes them unsuitable for a crystal radio \$\endgroup\$
    – Curd
    Sep 18, 2013 at 15:16
  • 1
    \$\begingroup\$ Well, all diodes have a low forward voltage at low current. For instance, 0V at 0A. :) \$\endgroup\$
    – Kaz
    Sep 18, 2013 at 19:06

4 Answers 4


As others (@Kaz) have noted, a Schottky diode may be a simple and cheap solution. I personally haven't seen a crystal radio made with them, but that can very well be because I have really never checked for such a circuit. By all means that should be your first try.

A germanium diode is best known for two properties:

  • Low threshold voltage
  • Relatively high resistance in contrast to silicon diodes, resulting in a more curved characteristic.

The low threshold voltage (essentially 0V!) can easily be reproduced with an active half wave rectifier as shown in the image below (found on Elliott Sound Products).

enter image description here

The operational amplifier is used to eliminate the (rightmost) diode's threshold voltage by inserting the diode within the feedback loop. The positive halve waves are amplified by -1 (\$A = -\frac{R2}{R1}\$), so essentially it is an inverting rectifier. With a sine wave you won't notice the difference as both half waves are symmetric.

The leftmost diode prevents the opamp from being driven in saturation (low rail) during the positive halve input wave. Subsequently the inverting input will act as virtual ground (V- = V+) which stabilizes the circuit.

This circuit only works reliably with a dual power supply as the opamp's output will be driven about 0.6V below ground.

  • \$\begingroup\$ Unsure if an active rectifier like this is low noise enough for your application, some germanium diodes I checked were marked 'very low noise'. \$\endgroup\$
    – jippie
    Sep 17, 2013 at 20:00
  • \$\begingroup\$ I like this, it accomplishes what I need it to. I'm assuming I can use 1N914 signal diodes for the two in the schematic? I'll give it a few days before accepting to see if there's anything else, but +1 for now. \$\endgroup\$ Sep 17, 2013 at 20:41
  • 2
    \$\begingroup\$ You are going to need a op-amp that is able to operate at the RF frequencies you're hoping to tune. \$\endgroup\$ Sep 18, 2013 at 4:44
  • \$\begingroup\$ @ConnorWolf Oh I didn't think about the RF frequency ... good point. \$\endgroup\$
    – jippie
    Sep 18, 2013 at 5:26
  • \$\begingroup\$ YOu haven't seen Schottky diodes in a crystal radio because they have a quite high reverse leakage current what makes them unsuitable for that purpose. \$\endgroup\$
    – Curd
    Sep 18, 2013 at 15:18

Note that the crystal-radio's Ge diode was required for listening to extremely weak signals from distant stations using no power supply.

To pick up the closest few AM stations, usually the diode need not be germanium. Well, unless you're down in the basement, or way out in the country far between cities. Or, if you're not using a ground with a longwire antenna.

Heh, you could always add an adjustable 0-1V battery supply using a 100K voltage-dividing pot, and place it in series with your 1N914 diode for forward biasing, then adjust the volts to maximize RF reception (perhaps 0.6 volts?) Add a 0.1uF bypass cap to route the RF past this DC bias supply? A little coin-cell should be more than enough here.

If a 1N914 diode doesn't do it, and if you don't want to use a ground+antenna, often you can fix things by using a ferrite loop antenna with extra-long ferrite core ...or by winding an old-style loop antenna, 1-meter diameter, need roughly 250uH inductance to match a 365pF tuning capacitor for 550KHz-1.5MHz . In a city with an AM transmitter within miles, such a resonator can develop several volts RF amplitude. Sometimes you can even charge a capacitor and use it to flash an LED. One guy in Chicago said he was seeing several volts at a couple amps, and could use a silicon diode and run DC solar-cell motors (this from an AM station less than 1KM distant.)

Cheat: watch the LC resonator's output with an oscilloscope. Tune it to maximize the RF amplitude, and if it's well above 1V p-p, then your detector diode doesn't need to be germanium.

Finally, is a professional signal generator available? Set it to 1MHz sine output, turn on AM modulation at about a KHz or so, and connect the output to a few-turns loop inductor perhaps a foot across (Heh, or string a 1-turn loop around the lab, or even out the window and around the entire building.) Use this "transmitter" to provide RF for designing your crystal radio. When you can receive a strong signal, crank the transmitter output way down, then redesign your radio to bring it back up again. After sufficient cycles of design improvements, shut it down and tune for ambient signals.

Don't fall for a misconception being propagated by crystal radio sites: they say that LC resonator is just a bandpass filter. Nope, wrong, and its purpose isn't to block other AM stations while passing just one. Instead, the resonator is part of an "electrically-short resonant antenna" configuration, where the 'EA' effective aperture is immensely enhanced by resonant coupling to incoming EM waves. In other words, disconnecting the LC resonator doesn't cause your crystal radio to receive all AM stations at once. Instead, it goes silent, because the antenna-wire's "electrical diameter" has decreased to nearly zero. With no resonator present, the too-short antenna no longer couples strongly to nearby EM fields, and has stopped absorbing EM energy. (The same antenna wire, whenever a high-Q resonator is connected, can intercept vastly increased milliwatts. It completely alters the fields surrounding any antennas shorter than 1/2 wavelength. It focuses the EM waves onto itself, somewhat like the "director" elements in a Yagi antenna.) Very cool physics, a classical analog of gas absorption lines, particle-collision resonances and even of Stimulated Emission (heh, does it display Rabi Oscillations when given sudden pulses?!!) See products based on this piece of little-known EM physics: Select-a-tenna, and Terk AM antenna. Check it out:

So, everyone always assumed that crystal radios were too simple to spend time investigating? They're too simple for engineer post-doc "science fair projects?" Guess again!


You are talking Active circuits here .This means that power is available .The Opamp active rectifiers would need a good fast opamp .Jellybeans like LM324 are far too slow.If you apply some foward bias current on the diode you can overcome the foward drop issue .When this is done the common Si diode say 1N4148 will work as well as the rare say OA81 Ge diode .This prebiasing has been done on early solid state radios before I was born .If you do not prebias you get nonexistant weak signal performance and horrible distortion at medium signals .The old vacuum tube detector diodes were high impedence devices that did not need prebias. It can be said that the contact potential did the prebias. Sure I have lots of Ge devices but this is a non commercial site and I recommend you prebias your diode .If you get this detector right you can get very good sound quality .Operation is better at high impedence due to lower distortion and less tuned circuit loading .If you have some old small signal RF Ge transistors floating around they should make very good diodes.

  • \$\begingroup\$ Prebiasing a Si diode is a good idea. All you need is a very large resistor (1 meg? I'm not sure) and a 1.5V battery. Or, you might just try using the small-signal Si diode instead of the Ge. If your radio signal is strong enough the Si diode will work. 1N4148 or 1N914 are likely suspects. \$\endgroup\$ Jan 29, 2017 at 1:16

An active component won't work unless you want to defeat the whole point of a crystal radio (that is, zero power source, other than the signal itself, required).

The germanium diode is used to rectify the tuned signal, the same way a signal diode would be used in an amplifying AM receiver (which is, in essence, the powered version of a crystal radio: it filters, rectifies and low-passes the signal so you can hear it, simple as could be).

The Wikipedia article discusses what they used to rectify the signal before modern germanium diodes. There are some interesting solutions for making prehistoric diodes, though I wouldn't stake my senior project on relying that they would work.

You may want to try out any of the small signal diodes they may offer at your local parts supplier (I also have a deep, bitter hatred for Radioshack). At a few cents, it's worth the experiment if this is an academic exercise. Maybe they can order some 5 cent germanium diode so you don't have to pay shipping? Lots of retailers will let you order through them and they just eat the shipping cost to their store.


Not the answer you're looking for? Browse other questions tagged or ask your own question.