Clipping Diode Shootout

[AnalogPackrat]

Some Ge diodes are point contact types that have relatively high series resistance. You can see these in the curves. They turn "on" at a relatively low Vf but have a shallow slope. If you have a decent magnifier you might be able to see the construction through the glass encapsulation.

 

[CJ]

Yeah they kind of remind me of the first transistor,

 

[AnalogPackrat]

Here's a mystery unmarked Ge that might be a "1N34" from a Radio Shack package. You can see the small black square Ge chip at the bottom of the glass capsule with the tiny point contact wire that connects to the upper lead.


Here's another mystery Ge diode with a single black band. Here the Ge chip is the smaller silver circular object connected to the upper lead with the big solder blob and the point contact wire is below.


This is a Fairchild Ge diode I got from a former boss whose father worked in DoD/Aerospace in SoCal in the 60s-70s. This came in a package of 30 devices each with a tiny numbered sticky aluminum tag on one lead. Note the heavy bond "wire" which likely means this is a diffused or grown junction type. It's I-V curve has the sharper turn-on of Si, but at about 60% the Vf.

If I get time I'll post some scope pics of these and other Ge, Si, and Schottky types I tested today. I built a modded BMP back in 2020 that has a switch for diodes in each clipping stage. First stage Ge-none-Si, second stage Si-Red LED. I used my ears to pick the Ge. Don't recall what they are. I like Ge into LED best on that design.

 

[midwayfair]

I did a much less sophisticated diode shootout one time. I mentioned that I liked a particular unusual diode in my klone and then someone went and bought all of them off e-bay and started charging $5 a piece for them.

The results are interesting, but I will say that my opinion has always been, after trying a gazillion of them in a bunch of different circuits, that for distortion almost any changes you get from the diode choice are swamped by whatever your tone control is doing (which tone control, the guitar, pedal, or amp? yes). It can be fun to try stuff but it's hard to actually blind A/B stuff for listening, especially if a small twist of a dial can eliminate the difference anyway.

However, their characteristics might actually be important in other types of circuits.

I don't visit the lab often enough ...

 

[AnalogPackrat]

I did a much less sophisticated diode shootout one time. I mentioned that I liked a particular unusual diode in my klone and then someone went and bought all of them off e-bay and started charging $5 a piece for them.

I recently bought a used ProAnalog Manticore pedal that is supposed to be some kind of Klon-based thing with two extra controls. So far I haven't found that it does anything special with my limited skills. I'll keep playing with it as I reconstruct/expand my disassembled pedalboard.

The results are interesting, but I will say that my opinion has always been, after trying a gazillion of them in a bunch of different circuits, that for distortion almost any changes you get from the diode choice are swamped by whatever your tone control is doing (which tone control, the guitar, pedal, or amp? yes). It can be fun to try stuff but it's hard to actually blind A/B stuff for listening, especially if a small twist of a dial can eliminate the difference anyway.

In the BMP the clipping diode differences are stark even with changes of guitar and pedal tone controls. Beware that the various historical BMP models and more modern variants can suck you into analysis paralysis with all the build options and transistor and diode choices.

 

[midwayfair]

I recently bought a used ProAnalog Manticore pedal that is supposed to be some kind of Klon-based thing with two extra controls. So far I haven't found that it does anything special with my limited skills. I'll keep playing with it as I reconstruct/expand my disassembled pedalboard.

"Doesn't do anything special" is personally my favorite description of the Centaur. (Yes, I've played a real one.)

In the BMP the clipping diode differences are stark even with changes of guitar and pedal tone controls. Beware that the various historical BMP models and more modern variants can suck you into analysis paralysis with all the build options and transistor and diode choices.

I've built or breadboarded [edit: almost, for safety] every historical version of the BMP. And I agree that the diodes can make a substantial difference going between something like all germanium/schottky and LEDs, but I didn't find that different silicon diodes made a big enough difference to get excited over. I couldn't hear the difference between transistors in there at all.* I'm not saying these things don't change at all, I'm just suggesting that they get swamped by a lot of other stuff in a typical guitar setup.

* My preferred "big muff" is actually this: Whisker Biscuit -- which isn't really a big muff. You can get huge tonal changes messing with with the only diode in there! It's a monstrously good sounding circuit, and tiny. (The transistor choice is super important, too. It really won't work with anything except a Darlington.)

 

[CJ]

i gotta try that whisker bisquit, is there a board for that?

ok here is a graph of the 1n34 diodes i bought off evilbay, they do not say 1n34 on them so who knows, in fact, the do not look anything like the curves whoops posted for the 34, i got burned,

but there are probably a lot of 1n34 diodes that are not marked.

i like the way those metal phillips/mullard OA9's look, plus they say mullard so they have to sound good, right?

tried the 1n34 out in the klon and was not pleased. Mazda red diodes are going in there next,

AP, how did you get those closeups?

 

[AnalogPackrat]

i gotta try that whisker bisquit, is there a board for that?

ok here is a graph of the 1n34 diodes i bought off evilbay, they do not say 1n34 on them so who knows, in fact, the do not look anything like the curves whoops posted for the 34, i got burned,

but there are probably a lot of 1n34 diodes that are not marked.

i like the way those metal phillips/mullard OA9's look, plus they say mullard so they have to sound good, right?

tried the 1n34 out in the klon and was not pleased. Mazda red diodes are going in there next,

Your new 1n34 curve looks a lot like my 1n60s from Anchor in Santa Clara.

AP, how did you get those closeups?

I used a cheap plastic loupe held in front of my phone camera. Plastic optics FTW.

View attachment 103112


View attachment 103113

 

[AnalogPackrat]

Ok. Here's my "data." I used a cheap Instek signal generator at 1kHz on sine or triangle with a DC offset to produce a 0 to 2V peak wave. This fed directly into the test diode then a 100R to ground. Voltage (X-axis) measured across both diode and R (I may try again with floating signal setup), "current" (X-axis) measured across 100R.

All plots below taken with my old Tek 2235 in X-Y mode. X probe is a 10x Tek unit so 0.1V/div. Y probe is 1x HP unit so 50mV (I = 50mV/100R = 500uA/div). The curve shapes look right, but the Y scale seems off compared to quick comparison against data book plots.

Scope settings:


A few diodes--paper label on CRT for reference. Origin is (-1, -3) div. I should have shifted it left two more, but I'd already taken a bunch of data before I realized my mistake.

 

[midwayfair]

i gotta try that whisker bisquit, is there a board for that?

CultureJam (a different CJ!) made one years ago, but I have no idea if anyone is making one now. I just do them on perfboard anyway, it's not that many parts. But I understand life is short lol.

Highly recommend making it with a switch to get rid of the mid scoop.

Here's a couple tracks I used it on (with the mid scoop bypassed)

Guitar solo in this around 2:50


And it's on the guitar in this the whole song -- the clean parts I think were just with the guitar volume turned way down:


(Sorry to derail the thread a bit, I just really like that pedal.)

 

[CJ]

That scope setup is kind of like a B-H curve tracer, man that is so much faster than hardcopy I gotta give that a try,

Midwayfair those are some cool tracks, reminds me of early pink Floyd in a way. I gotta build one of those, sounds like your getting a bit of the octave thing at certain frequencies , did you use bc549 Instead of 2n5088?

 

[soapfoot]

"Doesn't do anything special" is personally my favorite description of the Centaur. (Yes, I've played a real one.)

It's definitely a polarizing device! And with all the guitar-culture hype surrounding the Centaur, there's almost no way it could ever live up to anyone's expectations.

But it very much does something special--it's just that the "special" thing it does isn't very exciting (or even immediately-obvious when trying it out).

The Centaur's superpower is to make balancing an overdriven guitar sound against a band very easy. It has just the right kind of midrange poke.

Many overdrive pedals have a tendency to create a feeling of disappearing into the ensemble blend unless too loud. The Centaur (probably just through dumb luck) seems to emphasize all the right midrange to just sing right over everything (while still being at a level that feels nice and supported by the band).

This was really driven home for me during an EU run last summer when Lufthansa lost all of my luggage and gear. I was using rental gear for the rest of the tour... a series of very nice overdrive pedals (including many of the usual suspects), but none of them really did "that thing" for which I've come to value the Centaur.

 

[Whoops]

That scope setup is kind of like a B-H curve tracer, man that is so much faster than hardcopy I gotta give that a try,

The fastest is to use the Peak DCA75, its the one used in the video and also the one I used.
It does the curve tracing and then you can export the measurements to an excel file. From there i just added those to the document supplied in the video and that I posted here and then it shows the graph for you comparing to the other diodes.
It's pretty easy and very fast to do

 

[Whoops]

ok here is a graph of the 1n34 diodes i bought off evilbay, they do not say 1n34 on them so who knows, in fact, the do not look anything like the curves whoops posted for the 34, i got burned,

but there are probably a lot of 1n34 diodes that are not marked.

That 1N34 diode graph it's pretty similar to two of the different 1N34A diodes I tested,
check the graph I posted. There's two 1N34A with the curves on the right of the KTR diodes, actually those were the ones that I said they measured more like silicon diodes than germanium

 

[AnalogPackrat]

That scope setup is kind of like a B-H curve tracer, man that is so much faster than hardcopy I gotta give that a try,

If you've still got your HP200CD you can use its floating output and move the scope probe ground point to between the resistor and diode. Then your voltage measurement is purely across the diode and won't include the resistor drop like mine.

Midwayfair those are some cool tracks, reminds me of early pink Floyd in a way. I gotta build one of those, sounds like your getting a bit of the octave thing at certain frequencies , did you use bc549 Instead of 2n5088?

Yeah, interesting break-up tone. Might have to breadboard one. I really like their Omega boost. Have had mine on my pedalboard since I built it. Also made one for a friend. Mine has a 3-way switch for bias and his has a pot.

 

[AnalogPackrat]

It's definitely a polarizing device! And with all the guitar-culture hype surrounding the Centaur, there's almost no way it could ever live up to anyone's expectations.

I'm old enough to ignore marketing hype.

But it very much does something special--it's just that the "special" thing it does isn't very exciting (or even immediately-obvious when trying it out).

Well, the Manticore is underwhelming. Gain has a very irritating gritty brightness floating over a dull tone. The "savage" control is just a low end boost/cut. My Omega boost does that and I can put it where I want in my signal chain. The first stage drive control is ok. I haven't found anything "special" after several hours fiddling around.

The Centaur's superpower is to make balancing an overdriven guitar sound against a band very easy. It has just the right kind of midrange poke.

I'm not hearing it with the Manticore.

Many overdrive pedals have a tendency to create a feeling of disappearing into the ensemble blend unless too loud. The Centaur (probably just through dumb luck) seems to emphasize all the right midrange to just sing right over everything (while still being at a level that feels nice and supported by the band).

This was really driven home for me during an EU run last summer when Lufthansa lost all of my luggage and gear. I was using rental gear for the rest of the tour... a series of very nice overdrive pedals (including many of the usual suspects), but none of them really did "that thing" for which I've come to value the Centaur.

I'm sure it works for some people. I'll keep futzing with it. In no hurry to dump it.

 

[AnalogPackrat]

The fastest is to use the Peak DCA75, its the one used in the video and also the one I used.
It does the curve tracing and then you can export the measurements to an excel file. From there i just added those to the document supplied in the video and that I posted here and then it shows the graph for you comparing to the other diodes.
It's pretty easy and very fast to do

A live trace has other advantages. You can heat or cool the device and watch the trace change. With a storage scope you can make that persist in one view. You can also control the frequency of the sweep and the peak voltage swing/offset. A good digital scope that can grab the X-Y data for export would be nice, but I hate menus and soft buttons.

 

[dmp]

i gotta try that whisker bisquit, is there a board for that?

There is a link on the page, PCB is $15

http://www.olcircuits.com/olc_whiskerbiscuit.html

 

[soapfoot]

I'm sure it works for some people. I'll keep futzing with it. In no hurry to dump it.

Overdrive pedals are a bit like control room monitors—someone will always hate the thing you love, and someone will always love the thing you hate.

Three things with the centaur:

It’s at its best when not being asked to do too much (I rarely have the drive control above about 1 o’clock)

It’s likewise at its best into a decent tube amp that’s already sweating just a little—no need for it to be full-on audibly distorting, but it’s not most flattered in front of a Twin Reverb on “2” (nor a solid-state amp)

Finally, and for me most importantly, I find it most rewarding in the context of a group, rather than playing alone. That’s where I find it really makes my life easier relative to similar devices. This revealed itself to me over time… maybe about a year of using it before it really became clear to me.

But not everything will work for everybody. Lots of people love the Ibanez TS-808 and/or the Boss BD-2, and I don’t really like either

 

[AnalogPackrat]

Overdrive pedals are a bit like control room monitors—someone will always hate the thing you love, and someone will always love the thing you hate.

There's no accounting for taste.

Three things with the centaur:

It’s at its best when not being asked to do too much (I rarely have the drive control above about 1 o’clock)

I've tried all combinations of settings. Best was moderate drive, gain below half, "savage" to taste. Still pretty bland.

It’s likewise at its best into a decent tube amp that’s already sweating just a little—no need for it to be full-on audibly distorting, but it’s not most flattered in front of a Twin Reverb on “2” (nor a solid-state amp)

I used my '77 Vibro Champ into a big 2x12 cabinet at various levels up to pushing hard. Haven't tried with my '77 Deluxe Reverb. I don't do squalid state amps for guitar.

Finally, and for me most importantly, I find it most rewarding in the context of a group, rather than playing alone. That’s where I find it really makes my life easier relative to similar devices. This revealed itself to me over time… maybe about a year of using it before it really became clear to me.

I'm not currently playing out.

But not everything will work for everybody. Lots of people love the Ibanez TS-808 and/or the Boss BD-2, and I don’t really like either

I've got a very early TS-9 that's ok for certain things, but I prefer my modded Dirty Little Secret clone with Omega boost in front and maybe a fuzz (DIY BMP or hybrid Tonebender II currently) for saturated stuff.

I prefer pedals that are more dynamic and sensitive to guitar volume and tone changes (BMP isn't and neither is TB). I've slowly morphed my board as I build and try new things. Want to expand it (physically) before the next reassembly. I need to build a good FF, BoR, and other dirt flavors first.