half wave rectification and Ge diodes

UZILSD · Post by **UZILSD** » 14 Jun 2009, 11:51

RnFR wrote:well here's the circuit. it's basically a tweaked bosstone. as you can see it's not using normal diode clamps. i tried zeners and silicons and they sound like ass. i guess i might have to add another stage if i don't feel like picking through diodes. the smaller parts count turns me on though. oh, and no, i don't think you can't stack them like this.

Uhm... wait a moment. I did'n run a Spice emulation, but at a first glance increasing V drop of the diode won't increase volume.
In your circuit, it forms an half wave rectifier. The signal that passes to the output is the part of signal that is higher than V drop... it is not a diode clamp. An higher V drop means less and gated output here!

rrhoads33 · Post by **rrhoads33** » 14 Jun 2009, 16:17

UZILSD wrote: In your circuit, it forms an half wave rectifier. The signal that passes to the output is the part of signal that is higher than V drop... it is not a diode clamp. An higher V drop means less and gated output here!

That's exactly how I see this thing too. The only ways to get a higher output level are an additional amplifying stage or if you added a second diode antiparallel to the first one. But this would change the sound, because the signal would be symmetrical again...

I've already tried some stuff with serial diodes like yours in the signal chain, but it always sounded terrible, so i let it be...

Cheers

RnFR · Post by **RnFR** » 14 Jun 2009, 17:35

UZILSD wrote:
RnFR wrote:well here's the circuit. it's basically a tweaked bosstone. as you can see it's not using normal diode clamps. i tried zeners and silicons and they sound like ass. i guess i might have to add another stage if i don't feel like picking through diodes. the smaller parts count turns me on though. oh, and no, i don't think you can't stack them like this.

Uhm... wait a moment. I did'n run a Spice emulation, but at a first glance increasing V drop of the diode won't increase volume.
In your circuit, it forms an half wave rectifier. The signal that passes to the output is the part of signal that is higher than V drop... it is not a diode clamp. An higher V drop means less and gated output here!

i know it looks strange, but all i know is what i've done on my breadboard works, but only with Ge diodes. maybe it has something to do with leakage? and diodes with a higher Vf do end up with a higher output. i have actually tried it and measured the diodes. and it sounds great!

UZILSD · Post by **UZILSD** » 14 Jun 2009, 18:43

RnFR wrote:
UZILSD wrote:
RnFR wrote:well here's the circuit. it's basically a tweaked bosstone. as you can see it's not using normal diode clamps. i tried zeners and silicons and they sound like ass. i guess i might have to add another stage if i don't feel like picking through diodes. the smaller parts count turns me on though. oh, and no, i don't think you can't stack them like this.

Uhm... wait a moment. I did'n run a Spice emulation, but at a first glance increasing V drop of the diode won't increase volume.
In your circuit, it forms an half wave rectifier. The signal that passes to the output is the part of signal that is higher than V drop... it is not a diode clamp. An higher V drop means less and gated output here!
i know it looks strange, but all i know is what i've done on my breadboard works, but only with Ge diodes. maybe it has something to do with leakage? and diodes with a higher Vf do end up with a higher output. i have actually tried it and measured the diodes. and it sounds great!

Well, if so my final word is: choose the diode that sounds better to you, once you're satisfied with the sound throw in an amplifier stage at the end. It could be Ge transistor based, or even Si or FET since it is not intended to further distort the signal but to amplify it.

RnFR · Post by **RnFR** » 14 Jun 2009, 18:52

yeah, that's what i wrote up above, and what i'll probably go with if i can't find the right diodes. with a normal Ge diode, the output is still at least unity, but i would prefer to have a decent volume boost so another stage will probably have to be. this thing really sounds great-just a massive wall of fuzz. give it a chance if you've got some time.

Ripdivot · Post by **Ripdivot** » 14 Jun 2009, 21:56

UZILSD wrote:Try stacking 2 or 3 Ge diodes in series, it should increase V drop without altering tonal response, since all diodes are Ge. Give it a chance.
If it doesn't satisfies you, just add another stage and forget messing with diodes/trannies. I know that low parts count is appealing, but at this point if you got a sound you like and you only want more volume add an amplifier stage.

I Agree, I don't see why you can't series two germanium diodes in this circuit...it should work fine

RnFR · Post by **RnFR** » 15 Jun 2009, 02:42

i believe i tried it and didn't get any sound. i'll give it a shot again and see how it goes.

RnFR · Post by **RnFR** » 15 Jun 2009, 05:06

well i did some more testing and came up with some very interesting results. first, let me explain exactly what was happening when the Ge diode had too low of a Vf. the signal would clip and compress, than there would be a pause, and then the noise floor would rise up. with the higher Vf Ge's, there is less compression, and this annoying release wouldn't happen. with 2 series Ge's, the signal would pass, but the threshold of the compression would remain unchanged, and i also heard no change in clipping. when i tried a couple different zeners, indeed there was lower output, and some silicons ended up gating.

now here is where it gets cool- i put 2 Ge diodes in PARALLEL, and the threshold of compression went up! i ended up increasing the number of diodes to 3, and i eventually got a compression threshold and higher output that i was happy with. the clipping characteristics sounded basically the same, with just a bit of a change in tone. i tried the same trick with some zeners, and although it did work, the effect was considerably less and i didn't bother seeing how many diodes it would take to get to unity.

so now the question is, what is it about these high Vf Ge diodes that makes this happen? my uneducated guess is that it has to do with the increased leakage that Ge diodes have as opposed to other diodes. and Ge diodes with a higher Vf may have a higher leakage as well. of course i really have no idea what i'm talking about here, but it's the first thing that came to my head that seemed to make sense. of course there are many other measurable characteristics of diodes that i am pretty ignorant of, so i could be way off. the cool thing is that i seem to have found my solution. now i'd like to understand it.

soulsonic · Post by **soulsonic** » 15 Jun 2009, 07:29

I think you really need to try it with a BAT41 or similar Schottky signal diode. I've used them in a similar way before, and they worked really well in this way. I'll try it when I knock it up later, probably today.

RnFR · Post by **RnFR** » 15 Jun 2009, 07:50

ok, it looks like i'm kinda blurry about leakage in diodes. i'm going to do some reading.

i think i did try a schottkey. i got signal through nice and loud, but it clipped kind of strange and sounded sort of ratty in the decay. i'll try some more later.

UZILSD · Post by **UZILSD** » 15 Jun 2009, 13:32

RnFR wrote:ok, it looks like i'm kinda blurry about leakage in diodes. i'm going to do some reading.

i think i did try a schottkey. i got signal through nice and loud, but it clipped kind of strange and sounded sort of ratty in the decay. i'll try some more later.

The "rattyness" in the decay is normal: as the signal goes down at V drop level it goes "chopped", only the peak of the waves break the barrier and it this add extreme "crossover" distorsion with odd harmonics.
It's for this reason that I think that an higher V drop in the diode just worsens the scenario. Maybe Ge diodes, besides having lower V drop, have a less steppy break curve and they sounds more musical.

Anyway, I had an idea: if you want to minimize the gated and distorted decay you could add the amplifier stage before the diode. Doing so, you'll have a stronger signal braking thru the diode.

I'm waiting for further development, I always interested in a new fuzz design!

RnFR · Post by **RnFR** » 15 Jun 2009, 16:00

with the Ge diodes it sounds great, so i doubt if i will have to go that route. i'll try some more schottkeys later to see if i get any different results though. this thing has a ton of gain before the rectification already, but hey it couldn't hurt to give it a shot!

but wait a second- with the schottkey that i have already tried, if i raised the gain before the rectification, wouldn't the crossover distortion still happen? just later in the decay?

-----

i think i should split this thread into the workbench section. there's just ain't no part sourcing goin on here!

UZILSD · Post by **UZILSD** » 15 Jun 2009, 16:23

RnFR wrote: but wait a second- with the schottkey that i have already tried, if i raised the gain before the rectification, wouldn't the crossover distortion still happen? just later in the decay?

Exactly.

If you want to get rid at all of that you could bias the signal passing into the diode at a voltage around the V drop, so you can avoid crossover distortion at all.
In your circuit the diode rectifies the negative half wave, the signal goes under 0V so it would be more complicated to bias... if you reverse the diode it will rectify the positive half wave and all you need would be a simple voltage divider biasing (with a trimpot to fine tune it). BUT the sound may change, the positive half is probably different from the negative one... it is common in fuzzes. Trying doesn't hurt, anyway.

IggY · Post by **IggY** » 15 Jun 2009, 17:42

RnFR wrote:do i really need to explain this again?

no. One day you`ll understand why such questions were asked!

BTW:

FWIW: It`s Schottky, Walter H. Schottky
http://en.wikipedia.org/wiki/Schottky_diode

RnFR · Post by **RnFR** » 15 Jun 2009, 17:45

that sounds like good idea, i'll try and mess with it later, but i think i'm pretty happy with where this is going now. i think all i really need to finish this thing off is some filtering of the high end. i might have to adjust the .02 and 470pF caps to to roll off some more high end, or perhaps an RC filter after the diodes. all of this after i fix my boss's guitar of course! off to get some coffee, then back to the bench!

RnFR · Post by **RnFR** » 15 Jun 2009, 17:54

IggY wrote:
RnFR wrote:do i really need to explain this again?
no. One day you`ll understand why such questions were asked!

BTW:
FWIW: It`s Schottky, Walter H. Schottky
http://en.wikipedia.org/wiki/Schottky_diode

who are you, my kung fu teacher? why don't you just say what you mean?

earthtonesaudio · Post by **earthtonesaudio** » 15 Jun 2009, 19:41

I think you're right about leakage currents, especially from your comment about paralleling the diodes.

Have you tried using LEDs? Try some standard (not ultra- or super-bright) red LEDs, I bet you get a similar effect, but with a higher threshold for the weird compression stuff.

Or perhaps try paralleling different types of diodes, different thresholds might mean the decay behavior is spread out over a larger range.

RnFR · Post by **RnFR** » 16 Jun 2009, 00:27

well if our local flux capacitor guru agrees with me than i must be on the right track!

good to know i'm not too far off base. the last tests that i did were actually with multiple diodes of different types, and i just got to LED's, but i had to go to work so it will have to continue later.

so the question seems to be- is there a connection between Vf and leakage?

so LED's have a decent amount of leakage as well then? good to know i have more options on the table. i also did some more adjustments of frequency gain and filtering, so that stuff might change as well. overall it's sounding pretty cool either way. extra thick, and fuzzy. very doomy!

earthtonesaudio · Post by **earthtonesaudio** » 16 Jun 2009, 00:58

RnFR wrote:so the question seems to be- is there a connection between Vf and leakage?

I dont't think so. But I do think it has something to do with the materials used.
First, every diode has a characteristic reverse breakdown voltage, and every diode has some reverse leakage. Your typical 1N400x rectifier might have a reverse breakdown of 700V, you can get zeners in nearly any voltage from hundreds down to 3ish volts.
Germanium diodes, as you might expect, vary widely in both reverse breakdown and reverse leakage. Then there's the fact that the conduction "knee" is very soft, so that at low currents, the Vf is low, and as the current increases, the Vf gradually increases also. Other diode types have sharper knees, such as Schottkys or zeners.

Here's something you could try to figure out which effect is predominating. "Make" a low-leakage diode by taking a standard n-channel JFET, tie source and drain together to make a cathode, and use the gate as the anode. Compare it with a standard Si signal diode, and a Si power rectifier diode. This might help you determine what effect is making the sound you're after.

Then after all that, is the fact that diodes are also small-value capacitors. Paralleling them increases the capacitance value, letting more signal through. Isn't life great!

RnFR · Post by **RnFR** » 16 Jun 2009, 23:11

im not sure i understand what you are saying the reverse breakdown has to do with this whole equation. ii know that most of the characteristics on a diode datasheet are the max rating for voltage, current, temp, etc. but i'm not aware how those will fit into something like how they clip or rectify. is there some connection here? also i was aware of the different conduction knee points of diodes, but i don't think there is any proper way to measure this feature, is there? it doesn't seem to be something that is listed on a datasheet either. i'll try and do the experiment when i get a chance.

i could definitely hear the different capacitance of the diodes i tested as well. some of them seemed to cut a ton of low end, while others(Ge) seemed to pass plenty- especially when parallel.