What is the deal with those 1n34a diodes that look like silicon diodes?

[soggybag]

I'm seeing lots of diodes listed as 1n34a but they look like typical 1n914 or similar diodes? Are these really germanium, or are they silicon work alikes, or just bogus?

[bushidov]

Though not 1N34A germanium diodes, I have run into 1N68 germanium diodes that are in a DO-35 package. I came across these in my Behringer HM300 pedal, which is a clone of the BOSS HM2. They use those germanium diodes for the cross-over distortion diodes, which on the BOSS were germanium ones, so I was curious what a currently produced pedal was using in place of those, and they were 1N68's. But at quick glance, they looked like 1N4148s.

[garcho]

Where do you see them? I wonder if sometimes schottky diodes get passed off as germanium for the low V_F.

[j_flanders]

Check this thread from earlier this year on the subject:  "1N34A diodes. Cheap vs expensive."
https://www.diystompboxes.com/smfforum/index.php?topic=126386.0

[antonis]

Check this thread from earlier this year on the subject:  "1N34A diodes. Cheap vs expensive."
https://www.diystompboxes.com/smfforum/index.php?topic=126386.0

What Jack said : http://www.muzique.com/news/fake-ge-diodes/

Never experienced a genuine Ge diode with a forward voltage drop lower than 450mV or so..

[mozz]

? Many GE tests a lot lower than that on a Dca55.

[Mark Hammer]

It works in both directions.  Whenever I'm pulling diodes from the diode drawer, I have to test anything that looks germanium, because some have been pulled/scavenged from old boards I had picked up.  They may look like germanium, but they come up with Vf in the 600mv range, or even higher.

[amz-fx]

There is a chart on my page where I measured several different types with a Peak DCA75 semiconductor tester:

http://www.muzique.com/news/fake-ge-diodes/

regards, Jack

[Phend]

Using a magnifying glass or good eyes, does it look like this inside, with that little curved wire ?
Plus they are fat, not 1n914 dinky things.

(Help, correct me if I am being misleading since I am not a diode expert, but do believe germ diodes look like these pictures, if not, the diode in question is probably not, of course this visual only works if you can see inside)

[Rob Strand]

It's already been discussed in a few threads, like this one,

https://www.diystompboxes.com/smfforum/index.php?topic=126386.msg1211471#msg1211471

If the diode looks like a 1N4148 then it's going to be a Schottky and not a germanium.

Measuring the voltage is drop at high-ish current like 10mA is th best way I know of to split the difference.  Schottky's have a low voltage drop and germanium's a higher voltage drop - see prev thread for details.

[PRR]

..... if I am being misleading.....

The package is not the part. Ge diodes were packed in fuse-bodies in 1946, metal with leads by 1950, glass came later in several sizes. The glass package sure hung around (modified) in the Si era. I'd say a Ge could be in *almost* any package.

Any attempt to classify by voltage must say what current.

[Rob Strand]

The package is not the part. Ge diodes were packed in fuse-bodies in 1946, metal with leads by 1950, glass came later in several sizes. The glass package sure hung around (modified) in the Si era. I'd say a Ge could be in *almost* any package.

Any attempt to classify by voltage must say what current.

But winding the clock forward 70 years,  I think if you bought a Ge diode today it's safe to say it's going to be a "wiggly" wire in a glass package like Phend's pics.   If it's not like that I would be very suspicious.   If it looked like a 1N4148 it's almost certainly going to be a Schottky (ie. "Fake Germanium").

You need 10mA to be able to tell the difference from voltage measurements.   Anything less than 10mA starts to be less reliable.
[AFAIK, Jack's DCA75 tests are at 5mA.]

[PRR]

> You need 10mA to be able to tell the difference from voltage measurements.

Most guitar fuzzes put much less than 10mA to the diodes. So there's no difference?

All the basic diodes will shift 60-66mV per decade. At a more likely 1/10mA they will be 180mV-200mV down. Half-junctions (Schottky) shift less, don't they?

[Rob Strand]

Most guitar fuzzes put much less than 10mA to the diodes. So there's no difference?

All the basic diodes will shift 60-66mV per decade. At a more likely 1/10mA they will be 180mV-200mV down. Half-junctions (Schottky) shift less, don't they?

I threw-up a question like that in the old thread.   If we can hear a difference we should be able to measure some characteristic.

Most Silicon diodes end up being more than the theoretical 60mV/decade, perhaps around 100mV/decade.   Transistor base junctions and Schottky tend to be more around the theoretical 60mV/decade mark.     The difference comes from the non-ideality factor 'n'.   For transistors it's around n=1 and for most diodes it's around n=1.8 (1.8*60 mV/decade = 108 mV/decade).

Germaniums are a bit wonky but off hand they tend to be 100mV/decade and up ( n>1.8 ).   At higher currents, like 10mA, the resistance starts to come into play for germaniums.   When you start adding resistance it actually becomes a lot harder to fit the curve than you expect as either n or the series resistance can contribute to the shape.   To make matters worse both n and the series resistance depend on the current (FWIW, something not modelled by spice).   You might pickup a difference by measuring the slope at low currents.

One place where Schottky's differ is the reverse-recovery characteristic but you need a good pulse generator, fast oscilloscope and correct 50 ohm terminations and cables in the test set-up.    That doesn't explain why we here the difference though.

What you hear is probably reflected in a distortion or intermodulation test.   I could easily hear the difference side by side.    IIRC the Schottkys had a "grainy" sound and the germaniums an "open" sound.    However, with those samples I could measure some difference in the DC characteristics.

[Steben]

  Diode types are futile, only design will get you what you want. 

... Running away from the reeks and bats ....

[soggybag]

Help me understand the issues here. Those substitute 1n34a diodes are not a suitable replacement for a Ge clipping diode, they might be suitable as clipping diodes, or we should check them by ear?

[Rob Strand]

Help me understand the issues here. Those substitute 1n34a diodes are not a suitable replacement for a Ge clipping diode, they might be suitable as clipping diodes, or we should check them by ear?

Real germanium 1N34A's are still out there.   The "fake germaniums"  are Schottky diodes.    The question boils down to when does it make a difference substituting a Schottky for a germanium.

For clippers in a distortion pedal, where we hear the diodes,  it's safe to say you would hear a difference if you compared the diodes side-by-side.   If you had a germanium in series with a silicon diode and replaced that with a Schottky diode in series with a silicon diode then the differences would be less.

The characteristic that germanium and Schottky diodes share is the low voltage drop.   There are many circuits which only care about the low voltage drop.   If you replaced the germanium with a Schottky in these circuits it would work fine, almost certainly better than substituting a silicon diode which has a higher voltage drop.

[PRR]

> Real germanium 1N34A's are still out there.

And Small Bear and others still have buckets of Ge transistors. With luck, two diodes each. Steve also sorts-out low-gain transistors, and might even have a box of half-dead transistors. I'm not aware of any 3-leg Schottky as available as transistor-radio parts, so they gotta be genuine. We are not going to run out of Ge diodes.

[Rob Strand]

And Small Bear and others still have buckets of Ge transistors. With luck, two diodes each. Steve also sorts-out low-gain transistors, and might even have a box of half-dead transistors. I'm not aware of any 3-leg Schottky as available as transistor-radio parts, so they gotta be genuine. We are not going to run out of Ge diodes.

I wonder if they sound like germanium diodes.  The BE drop of a germanium transistor is somewhat lower than a germanium diode.  (You can also connect B to C and use B+C for one diode terminal and  E for the other.)

According to Gert Willmann's germanium models the non-ideality factor 'n' vary from 1.2 to 2.5.     However that's a curve fit so the 2.5's might be a result resistances.

Not a lot of clippers a using Ge transistors as Ge diodes but that shouldn't stop people give it a go.    Quite a few out there using silicon transistors - these will sound different to silicon diodes since the transistors are more like 'n' = 1 compared to the diode's n=1.8.

[soggybag]

Thanks for the discussion. I'm building a couple of EQD Life pedals, I'm using the Parentheses Fuzz PCB:

https://www.pedalpcb.com/docs/Parentheses.pdf

It uses 4 or Ge diodes. The first half of the circuit is a Green Ringer. The two diodes here are listed as Ge but I'm pretty sure I can sub something else and it should sound the same.

The second section is sort of a Rat clone with a switchable diode clipping options. One option is a pair of Ge diodes. From the discussion here the take away suggests the imitations won't sound the same.

All that aside the Ge clipping option is my least favorite choice it's a lot quieter. I feel like I should experiment with a few diodes I have on hand and see what sounds good there.