Strange results from using 1N34A diodes in a TS-based circuit.

[skiraly017]

Just playing around today and tried using 1N34A's as the clipping diodes in a TS-based circuit. I was surprised to find that they yielded almost no drive at all. In fact it drive control seemed not to work/make a difference when turned either way. I didn't install them backwards or anything, it was just like a clean boost with the tiniest bit of grit. When I stacked a LPB-1 in front of it, I could hear something driving but it was very, very faint. Almost like blending out the drive on a Sparkle Drive. Anyone have any thoughts on this? Should I be looking at different germaniums to use as clipping diodes? Thanks.

[Paul Perry (Frostwave)]

Try stacked diodes (that is, put two 1N34A's in series for each diode).
Also, if the Ge has a smoother curve, it is going to sound less harsh & thus possibly it will subjectively sound quieter.

[skiraly017]

Stacked up to three on each side and the pedal just started to come to life. Gain is still low. I can imagine a chain of diodes the length of my arm befroe this thing really starts to cranks up. Too bad, because what I can hear sounds nice.

[pyrop]

Stacked up to three on each side and the pedal just started to come to life. Gain is still low. I can imagine a chain of diodes the length of my arm befroe this thing really starts to cranks up. Too bad, because what I can hear sounds nice.

I'm thinking maybe use a 1M drive pot?
Worth a try?

paulw

[brett]

they yielded almost no drive at all

Yep, that's completely normal.  LEDs give the most headroom and heaviest distortion, Ge diodes the lowest headroom and mildest distortion.

[m-theory]

I put a 6 position rotary on my recent build, to select between different variations.  The Ge options (asym and very asym) are both highly disappointing.  Very little gain, very little output.  In fact, the output of both of those options is less than the stock T/S option, but with far less gain. 

I heard very good results with both asymmetrical and very asymmetrical variations of Si, and an asymmetrical variation of red LED.  I'm going to replace the Ge options on mine with an asymmetrical yellow LED, and a very asym variation of LED.  It might be worth experimenting with a single Ge coupled with maybe a pair of LED's and a signal diode. 

[Paul Perry (Frostwave)]

If you have a clean boost or preamp of any kind available, put that in front of it & see what happens.

[Dirk_Hendrik]

Try stacked diodes (that is, put two 1N34A's in series for each diode).
Also, if the Ge has a smoother curve, it is going to sound less harsh & thus possibly it will subjectively sound quieter.

Ok,
By stacking the diodes one lifts the voltage on which the diodes start to conduct.
In a project I'm working on(not a TS but with diodes in the feedback loop nevertheless)  I figured that before the board layout was done. (I ran into this problem some 14 years ago already after making a very weak fizzy distortion)  For that reason I added board space for anothher series diode to be able to do that lifting. In the current situation I use 1N4148's to get the required voltage levels and have a OA91 Ge diode in series with them for the clipping. Results are very good.

But,
All other parts in the project are pure Germanium*. Therefore, from a Mojo perspective, the adition of silicon in the circuit screws the "all germanium" approach. My original thinking was in the manner that the Si diodes are for the lifting and the clipping comes from the Ge diodes..... But. Is that the case or doe the Si's affect the sound just as well? I hear Y'all say: "fire up that iron and give it a go but there is the "*" below.....


So, for the "*"
I've got a small quantity of OOS (used) GE AC188 trannies here which were going to be used for this project. Since there's guite a variation I decided to build (amplifier) stage by stage and tune them correct. So.. Off I went built the first stage and tested it. Gain was far below one.... (no need to go on to the next stage btw) Hmmm. soldered in another trannie.... similar. Replaced with an Si one. Gain of about 20! perfect.  Spent the rest of the morning verifying leakage and measuring Hfe's  and got to the conclusion they were definitely OK from that perspective...(annoyance!!), breadboarded a rangemaster in a few minutes and Guitar tested it.... perfect. Great boost!

To wrap up without more details...
If your generator (more ideal for biasing transistor stages than a guitar) puts out 100Khz instead of one (accidental push of button) a Si will work... a Ge has lost all interest 
One afternoon "well spent". Never going to make that mistake again. Gawd I felt stupid.... but.. I've kinda had it with soldering for today.

[Transmogrifox]

I am a fan of the 1N400x series.  I have some 1N4001's in mine, and I think it sounds perfect to my ears.

Of course, perfect guitar tone is all a matter of taste.  The guy who sold me the OD pedal kinda looked down his nose at me, like, "Yeah you can buy it if you want, but it doesn't distort much...".  He was a glam metal warrior.

[Elektrojänis]

Just playing around today and tried using 1N34A's as the clipping diodes in a TS-based circuit. I was surprised to find that they yielded almost no drive at all. In fact it drive control seemed not to work/make a difference when turned either way. I didn't install them backwards or anything, it was just like a clean boost with the tiniest bit of grit.

To me, it seems like you might have a short between output and negative input (feedback path). It could be a solder bridge or a defective diode. That would make the clipping stage a non-inverting buffer and the diodes and the gain control would not have any effect on the sound.

The boost (if really any) and a tiny amount of grit could come from the tone control stage.

What happens if you take the diodes away from the circuit?

[Dirk_Hendrik]

You get a cleanbooster with a lot of gain. Better have a tubeamp following or you will not like how it sounds. (actually, this is the "hot" mode on a TX9DX)

[Elektrojänis]

You get a cleanbooster with a lot of gain. Better have a tubeamp following or you will not like how it sounds. (actually, this is the "hot" mode on a TX9DX)

Yes, I know. My question was for debugging skiraly017's pedal. If it does that then he has a shorted diode or removed some other short when taking out the diodes. If not, there is a short somewhere.

I suspect a short circuit in the feedback loop because in my experience TS-style circuit with GE diodes will still have distortion/drive and the drive pot will work. The output level will be a bit lower.

[skiraly017]

Just playing around today and tried using 1N34A's as the clipping diodes in a TS-based circuit. I was surprised to find that they yielded almost no drive at all. In fact it drive control seemed not to work/make a difference when turned either way. I didn't install them backwards or anything, it was just like a clean boost with the tiniest bit of grit.

To me, it seems like you might have a short between output and negative input (feedback path). It could be a solder bridge or a defective diode. That would make the clipping stage a non-inverting buffer and the diodes and the gain control would not have any effect on the sound.

The boost (if really any) and a tiny amount of grit could come from the tone control stage.

What happens if you take the diodes away from the circuit?

No shorts. Pulled the germaniums, put the old diodes back in and everything worked as it should. The germaniums turned it into clean boost and a nice one at that. I guess I'm just surprised at the result. I guess that's what experimentation is all about. 

[Elektrojänis]

No shorts. Pulled the germaniums, put the old diodes back in and everything worked as it should. The germaniums turned it into clean boost and a nice one at that. I guess I'm just surprised at the result. I guess that's what experimentation is all about. 

Yes... Thats what it's all about.

Can't help myself but I still think that there might be something wrong with those diodes though. Have you measured them?

[yeahyeah]

hey scott,
I know what was going on.
The thing that makes the TS gain sound so smooth is that when you have diodes in the negative feedback loop of a non-inverting op-amp the clean signal will always have a gain of 1. The diodes clip the boosted signal then that signal is mixed with the clean signal at the output...which gives us nice rounded edges on the clipped signal. The diodes limit the output of the amplified signal, depending on thier turn on voltage... so the clipping stage of a stock TS with Si diodes can only output a .6v peak to peak signal. With Ge diodes you only have .2 v of signal swing before the diodes turn on and limit the signal, so the amplified signal is limited alot more and is much more quiet. The limited "clipped" signal was limited so much that the constant clean gain overpowered it, which is why you didn't hear any distortion.

-Nick 

[stm]

I'd say to obtain the same results as a pair of 1N4001's or 1N4148's you need 8 Ge diodes, 4 in each direction.  In practice, modern hot pickups like better two silicon diodes or even red LED's, so in order to obtain a similar voltage drop using Ge diodes you should consider anywhere from 4 to 12 Ge diodes in series in each direction !?!? That's a lot of diodes, but it's the only way to make an apples to apples comparison with respect to the clipping characteristics.

I don't know of anybody that has done this and reported back the results, probably because most people don't readily have such a large stock of Ge diodes, and others don't even consider using that many devices in a single pedal. Nevertheless, that area is still open for investigation.

Finally, the amount of Ge diodes can be reduced by building a full-wave diode bridge (using four diodes), and then placing 6 more diodes in series at the output of the bridge.  In this way, only 10 diodes are needed, however you get 8 diodes in series for each polarity, but only need 10 diodes instead of 16.

[Transmogrifox]

I'd say to obtain the same results as a pair of 1N4001's or 1N4148's you need 8 Ge diodes, 4 in each direction.  In practice, modern hot pickups like better two silicon diodes or even red LED's, so in order to obtain a similar voltage drop using Ge diodes you should consider anywhere from 4 to 12 Ge diodes in series in each direction !?!? That's a lot of diodes, but it's the only way to make an apples to apples comparison with respect to the clipping characteristics.

I don't know of anybody that has done this and reported back the results, probably because most people don't readily have such a large stock of Ge diodes, and others don't even consider using that many devices in a single pedal. Nevertheless, that area is still open for investigation.

Finally, the amount of Ge diodes can be reduced by building a full-wave diode bridge (using four diodes), and then placing 6 more diodes in series at the output of the bridge.  In this way, only 10 diodes are needed, however you get 8 diodes in series for each polarity, but only need 10 diodes instead of 16.

...Or you could put an Si diode in the middle of the bridge.  Not quite the same as 6 Ge diodes, but a lot less  of a mess inside the box.

[sjaltenb]

Ok im new to all this but what i do know is that im about to build a clone of the FullDrive2 using the TS808 board from BYOC.

Heres the diagram for the clone...



I own a real FullDrive2 and the CompCut mode is a beautiful clean tone with a tiny hint of drive when the drive pot is at 60% or more. The 2nd position, flat mids is just that and then the vintage is just a regular 808 specs. as you can see theres a 1N34A in there

I am about to build this, just ordered the 34A's from Mouser today. Im hoping this will replicate my FullDrive becuase i absolutely love it. Has anyone tried this mod out yet? Ya'll really seem to know your stuff, does it look accurate? If so, any suggestion as to how to tie all three of these diodes together properly and then attach the wire?

[$uperpuma]

it looks as though the "boost" actually cuts the signal then allows it to flow full when turned on... am I seeing this right?

[sjaltenb]

yup, that is how the real FD works