The dreaded diode debate

[HamishR]

For me it's not as much about sound as feel. And the differences can be very subtle. I can often tell if a pedal I haven't looked inside uses LEDs as clipping diodes because to me they feel harsh. And I have guessed (correctly) before if a pedal has used a single pair of diodes such as 1N4148 before just by the amount of compression. I didn't guess the kind of diode, just that it was a single pair.

But there is so much variation between circuits, both in design and parts tolerance that there are no hard and fast rules. What works in one circuit may not work in another. I have pedals where a pair of BAT41s is fantastic, and others where they feel harsh. And I often build multiples of pedals I like and none of them sound exactly the same. The amount of variation in just the pots can make pedals sounds radically different.

And then you actually get to play with a band and all of that becomes irrelevant!

 

[Fama]

Just goona leave this here for the thread . . . .

A 6-way rotary PCB to select from six different clipping elements. Anyone recognize any of them?

View attachment 44315

I think one is LM313H.

Always ready to help!

 

[jhaneyzz]

For me it's not as much about sound as feel. And the differences can be very subtle. I can often tell if a pedal I haven't looked inside uses LEDs as clipping diodes because to me they feel harsh. And I have guessed (correctly) before if a pedal has used a single pair of diodes such as 1N4148 before just by the amount of compression. I didn't guess the kind of diode, just that it was a single pair.

But there is so much variation between circuits, both in design and parts tolerance that there are no hard and fast rules. What works in one circuit may not work in another. I have pedals where a pair of BAT41s is fantastic, and others where they feel harsh. And I often build multiples of pedals I like and none of them sound exactly the same. The amount of variation in just the pots can make pedals sounds radically different.

And then you actually get to play with a band and all of that becomes irrelevant!

They all feel similarly tubular to me..

 

[jhaneyzz]

For me it's not as much about sound as feel. And the differences can be very subtle. I can often tell if a pedal I haven't looked inside uses LEDs as clipping diodes because to me they feel harsh. And I have guessed (correctly) before if a pedal has used a single pair of diodes such as 1N4148 before just by the amount of compression. I didn't guess the kind of diode, just that it was a single pair.

But there is so much variation between circuits, both in design and parts tolerance that there are no hard and fast rules. What works in one circuit may not work in another. I have pedals where a pair of BAT41s is fantastic, and others where they feel harsh. And I often build multiples of pedals I like and none of them sound exactly the same. The amount of variation in just the pots can make pedals sounds radically different.

And then you actually get to play with a band and all of that becomes irrelevant!

Joking aside, I’m not enough of a guitarist to have that kind of feel but I do a pretty good job with recognising 4148 and LED clipping.

 

[Aentons]

Honestly, I don't hear much difference in diode types when playing a single clipping stage thru a clean amp. But, I hear it more and more when there are multiple gain stages. So, to play devil's advocate a bit here...

As others here have mentioned, the knee is an important part of the equation. We often picture the knee in a theoretically ideal form but clipping causes it to be anythying but. The details of the complex waveform are made in that transition and are additive down thru the signal chain until it hits our ears. The different diode types have different non-ideal characteristics like fast and slow switching, and leakage, that cause different artifacts like ringing and aliasing that are perceived as harsh or crunchy or whatever. It all adds up.

Yes, FV matters in how much the signal is clipped relative to its strength, but I think forward voltage can be considered fairly interchangable between diode types in series, as long as the FV is about the same.

While not specifically about diodes, here is a relevant post from Blackstone about harmonic distortion waveforms:

"Most of what makes an overdriven component sound the way it does, then, is how it shapes the corners that it adds to the wave, and whether it treats the bottom and top half the same way."

Guitar Distortion 101

Set aside what you know about equipment for a minute, and focus on the actual physics.

blackstoneappliances.com

 

[Rixroxx]

For me, in most cases, Ge diodes sounds a bit warmer and rounded on the edges. Perfect for vintage sounds or tube-like tones. Led sound different, less noticeable, more transparent, but in the same tube emulation spirit, warmth and round edges.

Silicon like 1N4148 or 1N914 are cleaner, sharper, bringing a bit more precision.

Mosfet clippers are somewhere between led and silicon, with its own flavor.

I think it's a bit like phaser sweeps : If you listen to 3 or 4 different phasers, set on low speed and high depth, connected after a distortion or a fuzz. You can hear that they all have their own character, their own color. The sweeps textures aren't exactly the same. It's subtle and probably unoticeable for unexperimented people, but very noticeable and very important once you are experimented and pay close attention to the sound. It's probably the same with reverbs, compressors, and every kinds of effects I guess.

To my ears, yes the LED is transparent and rounder yet I noticed it less forgiving under my fingers especially on soloing since i use legato a lot. If i'm into a more picking on solo it should work fine (like YJM). Also, it can be bit ice picky on most amps too that can sound transparent. This is the reason why 1n4148 is my go to clipping diode most of the time. Because IMO, it is tamable and has this spongy feeling under my fingers. Maybe it works on drop tunings or 7 string set up. But most of the time I use standard tunings on all my guitars.

I play from clean, to moderate and high gain music.

 

[YourGuitarist]

There’s not nearly as much in it as ad-copy writers would have you believe, so wouldn’t really worry about being a diode snob.

Fv is, like, obviously the most important aspect of diode selection since it determines the clipping threshold. There are other factors, like diode capacitance and knee, that are measurable differences, but for the purposes of guitar pedals don’t matter nearly as much as Fv.

That’s why you hear very little difference between a BAT41 and a germanium diode, both are clipping at around 0.3V. Most of the difference you will end up hearing can manifest in a bunch of small-and-inconsistent ways between a BAT41 and a germanium (and also from one germanium to another for that matter) is going to be nonlinearities caused by germanium being kind of an old crappy material - it won’t clip as consistently, it’ll be more likely to have capacitance weirdnesses, drift with temperature, etc. - but the overall behavior is largely the same.

All that said, clipping threshold *is* a big deal, because it determines headroom and feel (e.g. how hard you can play before clipping), and because a guitar isn’t a perfectly uniform white noise generator, some frequencies your guitar generates are going to be louder than others (in general mids and highs), and therefore they will hit the clipping threshold and clip earlier and harder, changing the perceived EQ.

If you recorded a looped guitar part and put it through the Stockade with LED clipping, then 1n4148s, then BAT41s, and then equalized the levels on the three clips, you should notice a change both in the EQ and the character of the crunchiness (which can sort-of be compensated for by playing with the gain control and essentially changing the overall amount of compression, but it’s not going to have the same dynamic range if you go that route).

Sorry for reviving an old thread...

but I think knee would be the most important quality as forward voltage can be shifted either by current-limiting/voltage dividing into the diode or adding clean gain prior. The slope of the onset of clipping however, cannot be easily accounted for with other circuit blocks besides maybe expanding or compressing the signal prior. I can't speak for diode capacitance yet as I don't have enough experience with it.