1N34A diodes. Cheap vs expensive.

[mozz]

Original picture i can see the corrosion on the leads, usually old stock does that. I have hundreds of old germanium diodes and also hundreds of old silicon, all from the 50-60's from my dads tv shop. I don't think i have a schottky here to compare. I'm going to pull out my old Hickok germanium transistor tester. It measures a lot of parameters for transistors and i know there is a section in the book on diodes, will have to see what i can measure. There are usually color coded bands but i have seen many japanese diodes i scrapped from radios that i know are germanium but are clear glass.

[Rob Strand]

The bottom of this page has some measurements of  the diode "slope" parameter "n".   The results are done at very low currents.

https://www.lessmiths.com/~kjsmith/crystal/dpiii.shtml

You can see the n for Schottky's tends to have n < 1.2 and germaniums have n > 1.2.  If you test at higher currents you should see a similar pattern.  "n" affects the softness of the clipping.

There's also this type of diode (the geranium diode) which tends to flourish this time of year.  The only diode that doesn't smell bad when the smoke gets out.

[bluebunny]

"What did you get for your Valentines gift?"

"A diode."   

[anotherjim]

Ok, but testing when you get them is a little late in the game. It would help to know how to judge the listing.

We know that some suppliers consider Schottky to be equivalent/replacement for obsolete germanium.  If you search for a germanium part, it could well offer Schottky in this belief.
We also know that you can't trust the photo of the offering to be correct. Very often it's any picture of any diode that was handy.
We probably do know what the price of the offering is. If it's at mass-produced silicon prices, then that's probably what it is.

[mark2]

Anyone using the ones NTE sells, either direct or through digikey? ~$0.70/each. Seem to be new

I got some recently and their Vf is fine, but I haven't done any further checking or testing.

I'm wondering if they're schottky.

[rankot]

Anyone using the ones NTE sells, either direct or through digikey? ~$0.70/each. Seem to be new

I got some recently and their Vf is fine, but I haven't done any further checking or testing.

I'm wondering if they're schottky.

Those seem to be Ge.

[iainpunk]

we can discuss whether or not diodes are true Ge or if they are fake/schottky... but i wonder 'if' and/or 'how much' it actually matters.
in a clipping situation, i couldn't tell you in a blind test if they are Ge or schottky, i used to be convinced i could, but after doing blind testing, i can't.
some germaniums sounded harsher and some softer, while the schottky diodes were generally more middle of the road, but they too had types that are softer or harder.
i don't want to stirr up a heated debate like those you can find on gear forums, but i like to hear other peoples experiences and opinions.

cheers, Iain

[PRR]

Anyone using the ones NTE sells, either direct or through digikey? ~$0.70/each. Seem to be new....

NTE has been doing business for 42 years. Their stuff has very long shelf life; until last year I could go into Greenbrook Electronics and find ECG/NTE parts from the last century. I don't know that old diodes are surely fresher from DigiKey or NTE itself.

[PietS]

I finally got my orders from China in and did some measurements. I made a simple test rig with a 9V battery, a 10k pot + 40 ohm resistor as a voltage divider. It's not good to test higher currents but I was mainly interested in the slope and smoothness of I-V curve at lower values anyway.

The moderately priced 1N34A mentioned before are called 1N34A-M in the figure. I tested 3 of those but the values were very similar so I just averaged them.
I got 2 expensive 1N34A, that had the cat whisker and all, in the figure they are called -E1 and -E2.
I also got a some cheap "1N34A" that were not see-through but orange (-C). I bought them together with some "germanium" 1N60 that had similar packaging except for some small print on the glass. The V-I curve of all these diodes were the same, so probably all the same diodes but a few with the print rubbed off. I also tested a few BAT41's and silicone 1N4148's for reference.

The moderately priced 1N34A and the cheap 1N34A and 1N60 are probably all Schottky diodes as their V-I curves are at low voltages and steep. The expensive 1N34A's had a different curve as the others, but the slope was very flat and the red one (-E1) really had lower current at the highest voltage. At 600mV the conduct similar amounts as a silicone diode which seems wrong. They are probably real 1N34A but just not very good ones   

[iainpunk]

i want to get my hands on diodes which have that red curve, might give some cool semi-wave-folding grind especially if the curve goes down further at that rate.

to bad you got some bad / semi-fake diodes, but the real question is; which sound better to you?
or do they just have different characters which are good for different types of distortion? even better yet, how do they sound together in asymmetric clipping???

cheers

[Rob Strand]

The moderately priced 1N34A mentioned before are called 1N34A-M in the figure. I tested 3 of those but the values were very similar so I just averaged them.
I got 2 expensive 1N34A, that had the cat whisker and all, in the figure they are called -E1 and -E2.
I also got a some cheap "1N34A" that were not see-through but orange (-C). I bought them together with some "germanium" 1N60 that had similar packaging except for some small print on the glass. The V-I curve of all these diodes were the same, so probably all the same diodes but a few with the print rubbed off. I also tested a few BAT41's and silicone 1N4148's for reference.

The orange package ones are likely to be Schottkys.    I've got a lot real germaniums which sit around the same voltages as the Schottky until you get to 1mA where it starts to deviate and by 10mA the difference between the germaniums and Schottkys is clear.

The -E1 and -E2 curves look very different.  Perhaps more different to what I generally see.   I'd have to pull out some very old stuff to see if any were like those.   Unfortunately I thew out most of my old original Ge's a *long* time ago.

FWIW, your 1N4148 curve looks a bit different to what I'm used to.

The BAT41's looks a bit high as well.
Data sheet is 260mV @ 100uA and 360mV @ 1mA.
My BAT46's were 200mV @ 100uA and 270mV @ 1mA.
You curve is 260mV @ 100uA but is aiming at 540mV @ 1mA;  looks like quite a bit more series resistance, like 180 ohms!

[Gus]

Anyone using the ones NTE sells, either direct or through digikey? ~$0.70/each. Seem to be new....

NTE has been doing business for 42 years. Their stuff has very long shelf life; until last year I could go into Greenbrook Electronics and find ECG/NTE parts from the last century. I don't know that old diodes are surely fresher from DigiKey or NTE itself.

Don't forget Masters

[PietS]

i want to get my hands on diodes which have that red curve, might give some cool semi-wave-folding grind especially if the curve goes down further at that rate.

to bad you got some bad / semi-fake diodes, but the real question is; which sound better to you?
or do they just have different characters which are good for different types of distortion? even better yet, how do they sound together in asymmetric clipping???

cheers

True, the real question is how they sound. I'll just try them all and see what I like best.

[PietS]

The moderately priced 1N34A mentioned before are called 1N34A-M in the figure. I tested 3 of those but the values were very similar so I just averaged them.
I got 2 expensive 1N34A, that had the cat whisker and all, in the figure they are called -E1 and -E2.
I also got a some cheap "1N34A" that were not see-through but orange (-C). I bought them together with some "germanium" 1N60 that had similar packaging except for some small print on the glass. The V-I curve of all these diodes were the same, so probably all the same diodes but a few with the print rubbed off. I also tested a few BAT41's and silicone 1N4148's for reference.

The orange package ones are likely to be Schottkys.    I've got a lot real germaniums which sit around the same voltages as the Schottky until you get to 1mA where it starts to deviate and by 10mA the difference between the germaniums and Schottkys is clear.

The -E1 and -E2 curves look very different.  Perhaps more different to what I generally see.   I'd have to pull out some very old stuff to see if any were like those.   Unfortunately I thew out most of my old original Ge's a *long* time ago.

FWIW, your 1N4148 curve looks a bit different to what I'm used to.

The BAT41's looks a bit high as well.
Data sheet is 260mV @ 100uA and 360mV @ 1mA.
My BAT46's were 200mV @ 100uA and 270mV @ 1mA.
You curve is 260mV @ 100uA but is aiming at 540mV @ 1mA;  looks like quite a bit more series resistance, like 180 ohms!

It was pretty quick and dirty and I'm sure my simple test rig gets unreliable with higher currents. Up to 200-300 uA it's probably ok, but I wouldn't trust the accuracy above it. The point of adding the BAT41 and 1N4148 was to have something to compare to but more in a relative way than in an absolute way.

[Gert]

I just had a closer look at the data-sheet characteristics of small-signal germanium diodes once made by Intermetall, Siemens, Telefunken, and Valvo/Mullard between the 1950s and 1980s. The considered diodes are from the AA and OA series. Similar germanium diodes from other manufacturers shouldn't differ significantly from these types, though.

First of all, it is important to note that there are actually two basic types of germanium diodes, namely point-contact diodes (which, physically, are Schottky barrier diodes) and ordinary junction diodes. The germanium point-contact diodes can be further subdivided into the following two subtypes:

– Germanium point-contact diodes of the non-welded contact type
– Germanium point-contact diodes of the welded contact type

As the internal construction of these two subtypes is virtually identical, a visual distinction is next to impossible—yet their electrical characteristics are quite different! Provided that my considered set of 20+ diodes is sufficiently representative, germanium point-contact diodes of the non-welded contact type, which seem to be of primary interest here, should nevertheless be safely distinguishable from silicon junction and Schottky barrier diodes and also from other types of germanium diodes.

To the best of my knowledge, all germanium diodes from AA111 to AA134, the AA137/138, and the AA140/142 are point-contact diodes of the non-welded contact type, as are all OA types from OA50 to OA95 and from OA150 to OA179, plus a few more. Similarly, all 1N types from 1N34 to 1N75 (except 1N53), from 1N86 to 1N90, from 1N111 to 1N117, and from 1N119 to 1N128 belong to this class, too. Examples of germanium point-contact diodes of the welded contact type (also denoted as gold-bonded diodes in some data-sheets) are the AA135 and AA143/144, whereas the AA136 and AA139 are germanium junction diodes.

In short, identifying germanium point-contact diodes of the non-welded contact type basically requires cautious measurements of the forward currents at 500 mV and 1 V forward voltage drop and of the leakage current at 2–5 V reverse voltage, all at 25°C junction temperature. Besides, all diodes in DO-35 or similar glass cases (like the 1N4148) can be completely ruled out anyway; all small-signal germanium point-contact diodes I've ever seen use considerably larger packages such as the DO-7 glass case.

The measurement temperature of 25°C is an important factor here since the forward and in particular the reverse characteristics of germanium diodes vary significantly with the temperature. For instance, the leakage current of the collector-base junction of germanium transistors at medium reverse voltages approximately doubles with every 10°C increase in junction temperature!

To get around such temperature effects, any internal heating of the junction must be kept as small as possible during the measurements. For the forward-current measurements, this means that the active measurement time must be kept short (e.g., by using pulsed operation with no more than 2% duty cycle) and the device must be given enough time to cool down after each measurement. Moreover, in order not to damage the device under test, all currents must be strictly limited to 15–20 mA, no matter whether the above two forward voltages can be met or not.

With this, the typical forward current of a small-signal germanium point-contact diode of the non-welded contact type should then lie between, say, 3 mA and 15 mA at 1 V and roughly 25% of this value at 500 mV; this rather flat forward characteristic is probably responsible for the very specific sound when such diodes are being used for smooth clipping. And when, in addition, the leakage current at reverse voltages of 2–5 V is safely greater than a few hundred nano-ampere, there is a pretty good chance to have indeed such a germanium point-contact diode at hand.

In contrast, germanium point-contact diodes of the welded contact type (gold-bonded diodes), germanium junction diodes, and most silicon Schottky barrier diodes usually reach forward currents in excess of 10 mA already at, or even well below, 500 mV forward voltage drop, whereas silicon junction diodes will just enter the conduction region at this voltage. Additionally, the leakage current of silicon junction diodes at reverse voltages of 2–5 V will usually be much smaller than 100 nA.

One minor, remaining problem with this approach, however, can be that all these old germanium diodes revealed significant variations in their actual characteristics, so that a definitive distinction can nevertheless be rendered difficult in some cases.

I hope this helps to clarify this basic issue.

In a forthcoming post, I'll address the in-circuit emulation of germanium point-contact diodes of the non-welded contact type using dedicated circuits with two or more silicon Schottky barrier diodes and a few additional resistors. This will probably be a more economical and also more reliable way to deal with this whole subject.

[Rob Strand]

There were some pics around of diodes which have the longer glass casing like germaniums but they are zeners.   For example old 1N746A through 1N75xA and 1N4370A to 1N4372A.  I still have quite a few of these.   

The difference is they don't have the wiggly wire inside like the true germaniums shown reply #25,
https://www.diystompboxes.com/smfforum/index.php?topic=126386.msg1208968#msg1208968

They have a solid internal structure, like this, (same look as Reply #35).



Also when you measure them in the forward direction it should be quite obvious they measure like a silicon diode.

[Gert]

There were some pics around of diodes which have the longer glass casing like germaniums but they are zeners.

That's not the point. What I wanted to express is that there are no genuine germanium point-contact diodes in such small glass cases as, for example, the DO-35 case.

The difference is they don't have the wiggly wire inside like the true germaniums shown [in] reply #25

The fundamental drawback with this wiggly wire (or 'cat whisker') is that this is, of course, a necessary condition for point-contact diodes, yet not a sufficient condition for point-contact diodes of the non-welded contact type—and these are the diodes which are of interest here!

By the way: I was a bit lazy in this respect when describing the three types of germanium diodes. In fact, only the point-contact diodes of the non-welded contact type are Schottky barrier diodes, physically. The welding process forms a tiny P region in the N-type crystal around the original point contact which, in effect, transforms the point-contact diode into a junction diode with completely different electrical characteristics. See:

– https://en.wikipedia.org/wiki/Diode#Point-contact_diodes
– https://www.richis-lab.de/Diode05.htm

Since semiconductor junctions can be light-sensitive, many of these old, glass-encapsulated germanium devices, diodes as well as transistors, were painted black to protect the interior from light. But even if one scratches off the black painting and finds such a wiggly wire, it's still next to impossible to distinguish between welded and non-welded types. At least something like a microscope would be needed to see artefacts of the welding process in the close vicinity of the point contact (last image on Richi's Lab page); non-welded types do not have such a welding bead. All in all, this is not really a practical option.

The only sufficiently reliable method to identify germanium point-contact diodes of the non-welded contact type is therefore to perform the described measurements, even if the given range of forward currents (i.e., 3–15 mA at 1 V forward voltage drop and roughly 25% of this measured value at 500 mV) is just a best guess of what ought to be found. Slightly lower values might be possible for some diodes, though.

[jsa_man]

Are these diodes original? As far as I know, TME is a serious company and I don't think it messes with fakes, and I was curious if these diodes are really 1n34a because I know they can't be found anymore. https://www.tme.eu/en/details/1n34a-nte/diodes-others/nte-electronics/1n34a/