by Andrew Ellis, NO6E
The diode detector on my first "crystal set" radio was just that -
a Galena (Lead Sulfide) crystal with a fine "catwhisker" metal contact. One felt around until one found a good spot where diode action would "detect" AM signals. It seemed, and it was, remarkably crude. Still, there was the something-for-nothing appeal of a radio that picked up signals (more or less) without any power other than the signals themselves.
The upgrade from the Galena crystal was to a packaged-in-glass commercial diode. I used the granddaddy of signal diodes, the 1N34A. Like most semiconductors of the day, it was made using Germanium.
In the decades that passed, Germanium semiconductors were replaced almost 100% with those made from Silicon. Silicon was cheaper, more reliable and just all-around better in almost every way. The 1N34A gave way to the ubiquitous silicon signal diode of today, the glass-enclosed 1N914/1N4148. In hundred-quantities they're dramatically cheap, well under $.02 each.
Still, there a few places where Germanium diodes are useful, both as radio-frequency detectors and in some intentional-distortion generators used in music processors. Germanium diodes have a lower forward-voltage drop than Silicon. Fairness demands, though, that I acknowledge that Shottky (or "Hot-carrier") diodes, typified by the commonly-used 1N5711-series, also offer low forward voltage drops. They're more expensive than '914's, but only about $.40 in small quantities.
When I decided to build a field-strength meter (a piece of test equipment that shows the relative amount of energy from nearby transmitting antennas), I decided to use two of the old 1N34A diodes at the detector. How much could they cost, after all?
Somewhere north of $2.00...each. If you could find them at all.
I checked eBay, my go-to source for things that seem too expensive from regular dealers. I found an offer of 20 for $2.59, so I bought them.
What's going on? Is there a worldwide Germanium shortage? More likely, it just has to do with popularity. Engineers tend to use the same parts over and over in projects. The ones they choose may not be the most cutting-edge versions, but they're common. Commonly-used parts often are "second-sourced," meaning that more than one manufacturer makes them. That's vitally important when you order parts for your already-designed project and the normal supplier you use suddenly lists availability of one of your parts as, say, "2Q17," meaning "In the second quarter of next year." Chances are you need it much sooner than that.
So, when you look at, say, a small audio amplifier kit and it uses the hoary LM-386 chip instead of something newer, which would also be cheaper and better, that may be the reason. At my last job before retirement, I dragged us into using the stereo TDA2822 series amplifier instead of the '386. Today, the '2822 is listed as "obselete," though still available. The noisy and troublesome '386 series soldiers on. Even if they don't use the "O-Word," manufacturers often include the ominous phrase "Not Recommended for New Designs" on their data sheets. The designer ignores such warnings at her/his peril.