The germanium point contact diode, and almost every semiconductor device using germanium, is now obsolete. There was a time when almost every television or radio would have contained one or two of them, but the world has moved on from both analogue broadcasting and discrete analogue electronics in its lower-frequency RF circuitry. [TSBrownie] is taking a look at alternatives to the venerable 1N34A point-contact diode in one of the few places a point-contact diode makes sense, the crystal radio.
In the video below the break, he settles on a slightly more plentiful Eastern European D9K as a substitute after trying a silicon rectifier (awful) and a Schottky diode (great in theory, not so good in practice). We’ve trodden this path in the past and settled on a DC bias to reduce the extra forward voltage needed for a 1N4148 silicon diode to conduct because, like him, we found a Schottky disappointing.
The 1N34 is an interesting component, and we profiled its inventor a few years ago. Meanwhile, it’s worth remembering that sometimes, we just have to let old parts go.
Why would I move away from it for my own projects? I have a significant stock on hand, and have since I was in uni, adding to it periodically as chance permitted. I added more to it a few months ago at a yard sale.
Several other components that I stock up on opportunistically, including several types of germanium transistor and tunnel diodes, as well. When I go belly up, I hope someone appreciates it all as one lot, but most likely it will be broken up by a reseller and some crapcanned by an unappreciative soul.
Oh well.
You could will it all to somebody who might appreciate it, like a ham radio group or some audio fanatics
It is, but that requires that it be followed. If I die tomorrow, the executor of my estate should handle it ok. To put it simply, if I end up incapacitated and in hospital, due to the USA’s wonderful health care system, despite my having insurance (at WAY too high a monthly cost, despite it being `free’), they get first dibs, and the junkers come in to clear my home out. I have as much protection in place as possible, the electronics being a minor part of it, but short of being wealthy, I really have no control.
Which is why you use a hack that is better than just a will. You establish a revocable trust in your name with a pour over will, move all your assets but a checking account into the trust, and direct and constrain the actions of the other trustees under various states of incapacitation or death. Before you complain about having to pay for health care, I suggest you trying living in the places where it is “free” so long as you can wait forever for an appointment or treatment. You don’t have to travel very far to see that the “free” health care was paid for by doubling sales tax on everything, and the average wait for a new “family doctor” who acts as a gatekeeper for other services is 3 years.
“Before you complain about having to pay for health care, I suggest you trying living in the places where it is “free” so long as you can wait forever for an appointment or treatment. ”
yes to a shortage of Drs accepting new patients, but as a permanent Canadian I can tell you that the vast majority of healthcare needs here are met competently and quickly. Even if you don’t have a family Dr and have to go through a walk-in clinic.
You can always leave, y’know.
Now I’m gonna go do inventory of my germanium diodes. I frankly don’t care what happens to my stuff after I die, as long as it’s not a burden to my peeps.
It’s not clear what he is measuring, but it seems the signal generator (1.5 V peak, 3 V p-p ?) is probably driving the cathodes of the diodes and the negative peaks are rectified and held by the scope probe’s ~ 15 pF capacitance (and 1 MΩ). Notice the ‘rise’ time is slow (this the decay of the peak rectified signal and dominated by the scope 15 pF // 1 MΩ).
Look at the waveform baseline — it appears to be ~ 60 mV above the signal.
It’s not clear that this indicates the P9K is ‘better’ — where is the baseline for channel 1 ? it just appears that there is more capacitance in on than the other. Notice that the (negative) peaks of the waveform are at the same level. Just looking at the p-p amplitude doesn’t tell anything (a resistor in this setup would give 3.0 V !)
What you want in a crystal radio is a high ratio between forward conductivity and reverse conductivity at low signal levels. For AM, diode capacitance isn’t significant. Better to measure forward and reverse DC current at (say) 0.1 V, 0.2 V, 0.3 V, to 0.5 V or so.
At least it’s still (and always will be) pretty easy to make a germanium point-contact yourself. Not great for miniaturization
Because I read Hackaday, I know there are people out there making artisanal vacuum tubes and building home chip fabs. This seems easier, maybe even I could do it.
yeah i was surprised to hear they’re unobtainium…made me look in my pile and i have a few left from a 10-pack i bought 25 years ago. and if i have them — yup — then ebay has them. presumably when they become truly rare, someone will start manufacturing them just like polaroid film and floppy disks :)
the thing is, i haven’t built a crystal radio in 35 years and when i think about it, the few times i’ve used a germanium diode recently i was just playing around blindly and truly what i wanted was a schottky. that is, they’re still so plentiful in my house in particular that i have used them where they don’t make any sense to. really goes to show more about how much of a novice i still am at analog circuits :)
now that i know they’re rare, i’ll always try the schottky first when i’m casting about for a lower Vf, and i’ll probably never use a germanium diode again
Check the forward voltage drop — something over 25 years ago, I bought some “1N34A”s that showed a 0.6-0.8 V drop, indicating that they were actually silicon diodes in disguise, and so totally unusable for my application.
No void. No relevant application.
Schottky diodes make excellent rectifiers in power circuits due to their low forward voltage drop and fast response times. However, they also have significantly higher capacitance and reverse leakage current; so it makes sense why one wouldn’t make a good detector in a crystal radio circuit.
Schottky power diodes make excellent power rectifiers, they’re designed to be so.
I’ve used Schottky signal diodes like the 1N6263 and some more esoteric (but still surprisingly cheap) variants as detectors and they work reasonably well, enough that I didn’t feel the need to rpelace them.
They often have far better reverse leakage current (0.2µA for the 1N6263 vs. 100µA for the 1N34A), the junction capacitance is also pretty good on a bubblegum part like the 1N6263 at ~2pF vs. 1pF for the 1N34 so I’m not sure that’s the whole story.
I actually wonder if it could be the ‘poor’ diode specs of the old germanium parts that makes them a good detector?
Here’s where you can find the 1N6263W diode in the SOD123 SMT package and the SPICE model:
https://www.diodes.com/part/view/1N6263W/ [SOD123]
https://www.diodes.com/spice/download/506/1N6263W.spice.txt
https://www.digikey.com/en/products/detail/diodes-incorporated/1N6263W-7-F/814969
To use the SPICE model in LTspice:
1. Schematic > Edit > SPICE Directive > Paste the two-line model below into rhe box.
.MODEL 1N6263W D(IS=1.82u RS=2.80 BV=60.0 IBV=200n
+ CJO=2.65p M=0.333 N=1.70 TT=1.44n)
2. Right-Click the Diode symbol’s name (D) and change it to 1N6263W which matches the name in the model.
3. Run the simulation.
ST Microelectronics sells the 1N6263 in the axial DO-35 package. The SPICE model is the same for either package:
https://www.st.com/en/diodes-and-rectifiers/1n6263.html#overview
https://www.digikey.com/en/products/detail/stmicroelectronics/1N6263/603533
Herr brain,
You are right re power Schokttys. However be advised that RF and uW devices are totally different. Do a search for Low Barrier units. These will have 0.25 to 0.3 forward voltage at 0.5 mA. Very low Capacitance too..Their reverse voltage is only 3 to 10 volts and they are very susceptible to static or obviously lightning induced failure. I can’t tell the difference in sensitivity between a 1n34a, a 1n60 and a shoktty like bat6804 or bat1504w. The latter part is offers 0.35pf at 0 volts bias. All are about 50 cents in small quantities from digikey.com.
Seriously? That’s a terrible way to measure the switching performance of the diodes. As best I can tell he’s not using a shunt across the scope inputs, which means that he’s just charging the input capacitance of the scope. Besides, I did see anywhere in the video where he says a Schottky diode was worse, and the HaD author didn’t bother to say why she thought it was either. Schottky diodes in general have a sharper knee than germanium didoes, and that’s desirable. And there are Schottky diodes designed for multi-gigahertz applications that have very low capacitance with FAR less variability in other specs than point contact germanium diodes do.
HaD still has no edit capability. That should have been “I didn’t see” …
I didn’t come here to read about didoes. What? Nevermind.
I agree with azDave. Schottky diodes (sometimes called “hot carrier” diodes) that are designed for RF/microwave applications have excellent characteristics for a crystal radio. I once used an axial one to measure the leakage around the door seam of a microwave oven, with the leads acting as a dipole antenna and a cheap digital voltmeter as the detector.
A 1N34 has about 1 pF of capacitance (typical), though sloppy layout and design techniques could increase that substantially. Microwave Schottky diodes will have considerably less than that.
For a drop-in replacement, one could try a 1N5711. Digikey has a ton of them in stock for $0.25, quantity one, which drops to less than a nickel each when you buy a thousand.
If you need better specs, MACOM has a number of possibilities.
Weird, there are plenty of 1N34A diodes on eBay. AliExpress also has them. So why would anyone need to look for alternatives?
On related note, aside from making a crystal radio, there is no real need for germanium diodes. And true crystal radios should use galena anyway…
I have bought from a number of sellers and all have been Schottky diodes, not the genuine thing.
Because they’re pretty much all fake.
A quick search on ebay gives 60 results, of those 60 55 are obviously not poit contact diodes, there are only 5 listings which actually have pictures of point contact diodes and of those 5 I wouldn’t like to guess how many are actually 1N34A parts or are actually some other Germanium point contact diode (which may well work perfectly fine)
I grew up with OA95 germanium diodes to make crystal radios etc. and still have a few of them in stock. These are still in production it seems. Not sure how it compares to the 1N34 though.
When testing diodes for use in xtal sets it it vital to do so at the expected operating current. Schottky diodes have a lower forward voltage drop than Germanium when measured for datasheets (typ. 1mA). But as you lower the current the forward voltage of Germanium falls much more the Schottky and at xtal set levels (picoAmps at best) the Germanium is way lower then the Schottky. The video fails to show the test conditions other than frequency.
That characteristic of a germanium diode (a sloppier knee) makes it a generally worse switch as a detector. And that also happens to be one of the characteristics of a germanium diode that suffers from significant production variations. I spent my entire career in semiconductor manufacturing and there were good reasons we stopped making germanium devices.
If one simply takes the time to read and understand the electrical parameters of the two devices, one will very quickly–without much effort, whatever–find the difference(s) between a Schottky diode and a germanium diode which determine why a Schottky diode absolutely can NOT be used as a detector in low-signal-level, high-impedance applications.
Low forward-voltage-drop–with no consideration of any of the rest of the Schottky diodes’ characteristics–is, absolutely, not one of them.
This happens everywhere in every area everyday, parts and things being put to EOL. And people get mad every time. Myself included, wanted to make a project with INTEL Edison and none to be found on the internet.
I mean, let’s look forward, people.
If there is a market, someone will make them. The fab tech required is not very expensive or complex. The only application requirement I can think of is the lower threshold vs silicon. Can’t think of a legitimate use case except novelty though.
Speaking about germanium generally, rather than just the 1N34A, germanium has a higher mobility than silicon, particularly hole mobility. This means faster devices.
I tried just about every schottky diode for use in crystal radios and nearly all of them were far worse than 1n34a’s except for the BAT15s and BAT63s which make excellent replacements for the 1n34a germanium diodes in crystal radio applications. They are louder than the germaniums. The only disadvantage is that they have relatively low reverse voltage handling of a few volts (they act like zeners), which would cause audio distortion at signal levels strong enough to light an led (but at that level just about any small signal diode will do).
Perhaps you should try some Germanium Tunnel diodes as detectors…