All About Diodes...

A few questions...

1. What do they all sound like? What is the sound difference between silicon ones like 1N4148's + 1N4001's or germaniums like 1N34's or LED's?

2. Does LED color make a difference (red, white, yellow, black, infrared...). Do they glow when you drive them hard?

3. Symmetrical vs. Asymmetrical clipping. What is the sound difference?

A few questions...

1. What do they all sound like? What is the sound difference between silicon ones like 1N4148's + 1N4001's or germaniums like 1N34's or LED's?

2. Does LED color make a difference (red, white, yellow, black, infrared...). Do they glow when you drive them hard?

3. Symmetrical vs. Asymmetrical clipping. What is the sound difference?

1. this guy - http://www.thefret.net/archive/index.php/t-
Has some good sound samples

this guy explains it pretty well too
http://www.diystompboxes.com/pedals/diodes.html

2. Yes! Because they're creating different wavelengths of light they alter the sound a little differently. Listen to the sound samples and see what you think.
Yes they do! It looks pretty cool on a pedal too ;P

3. Ibanez uses Symmetrical and boss uses asymmetrical. Most of the time asymmetrical is considered more amplike though

Hope that helps!

Thanks for the help!

I have a vague memory that BOSS has managed to "patent" asymmetrical clipping.Am I right?

I have a vague memory that BOSS has managed to "patent" asymmetrical clipping.Am I right?

I heard that before too. I heard that's why Ibanez went with symmetrical clipping.
Not sure if there is any truth behind this or not. It seems a little far-fetched that they could patent something like that but I guess it is possible! I'd love to know if it's true or not if anyone has any insight about that

The different voltage drop of different devices clip the waveform at different amplitudes. 10Gauge posted this once, in part:
You chain diodes in series to increase the cumulative forward voltage, to increase (change) the clipping threshold. At the two ends of the spectrum, you have examples like the following:

- BAT42: ~280mV
- LED: ~1700mV (actually varies with the color)

Everything else falls in between:

- 1N34A (germanium): ~ 340mV
- BAT41: ~ 370mV
- 1N400X series: ~500mV
- 1N4148/1N419 ("standard" silicon): ~600mV
- many transistors (including mosFET's): ~720mV

Putting these devices back to back gives you different levels of asymmetry.
Also, I doubt Boss patented asymmetrical clipping since many others do it.

- BAT42: ~280mV
- LED: ~1700mV (actually varies with the color)

Everything else falls in between:

- 1N34A (germanium): ~ 340mV
- BAT41: ~ 370mV
- 1N400X series: ~500mV
- 1N4148/1N419 ("standard" silicon): ~600mV
- many transistors (including mosFET's): ~720mV
.

Is that a fact? I Always thought the 1N400's where just below LED's on the scale. Learn somethin' everyday!
Hey, how do you measure diodes? I have a ream of NOS telefunken black glass silicone diodes,and I dont know where they fall on this scale?http://www.ebay.co.uk/itm/100-OA128-vintage-silicon-diodes-TELEFUNKEN-/380372428790?pt=UK_BOI_Electrical_Components_Suppl ies_ET&hash=item588ff6a7f6
They sound great in my TS-9 but I'd like to know what I've got.

Also, I doubt Boss patented asymmetrical clipping since many others do it.

It is a phenomenon, not an engineered effect. Many amplified devices clip asymmetrically on their own, such as the long tailed phase inverter in tube amps, and even the Fuzz Face. Neither has diodes in the standard sense. The phase inverter mentioned actually has its circuitry "tweaked" to try and correct the EQ/clipping asymmetry. You will see a smaller cap on the "first input," and usually a smaller plate resistor, too. For example, the "first" input cap will typically be .001uF, .01uF, or .022uF, and the "second" input cap will typically be .1uF. The coinciding "first" plate resistor will usually be 82K, with the other being the standard 100K.

Most folks don't understand the transition between linear amplification and what happens when the gain exceeds what can be amplified without clipping - it is darn near impossible to have a device that stays symmetrical on both the positive and negative swing. The op amp is about as close as the technology allows without getting into very sophisticated circuits.

...Since the op amp provides such good gain figures prior to clipping, the clipping diodes are typically seen as a sort of "necessary evil" - you don't want to push the op amp into distortion on its own (in most cases), so you clip it with the diodes. That said, it is *slightly* out of fashion with many new pedal builders, since it is viewed as being kind of like crude technology. The "new clipping order" typically uses a combination of jFET or mosFET transistors that are supposed to give a more "tube-like" clipping response, because they aren't as "simple" in their operation as clipping diodes are. But even that is a bit of a fallacy, because either you have entered the proximity-to-clip range, or you haven't.

Is that a fact? I Always thought the 1N400's where just below LED's on the scale.

The 1N4001 to 1N4007 diodes all have smaller forward voltages than even common silicon 1N4148 diode types.

Think about what most diodes were designed to do - they are used for switching/power supply/d.c. rectifying purposes. For that reason, a forward voltage of something like a LED at 1700mV is going to be bad.

There is a quad of 1N4007 diodes used as a solid state rectifier in Marshall amps, and a ton of other amps. The Twin Reverb uses SIX.

Germaniums, Schottkys, epoxy rectifier types, silicons will all have typically small forward voltages, with germs and some schott's being at the low end of the scale. The GGG Bluesbreaker kit has what it calls a 1N4001 diode that is actually a Schottky that has a REALLY low forward voltage of around 250mV or so!

...The nice thing about the Schott's is that they fit better than the old germanium glass diode types, are cheaper, and have similar or same forward voltages. The newest D+'s have a Schottky that is a replacement for the original germ diodes. It was a necessity since a big company like Dunlop needs lead free components, and the old stuff is anything but environmentally compliant.

Hey, how do you measure diodes?

There are multimeters that specifically have a diode testing function. You basically put the black probe on the cathode (striped end) of the diode, and the red probe on the anode. It will give you a number in mV, unless the forward voltage is too high (like with some LED's).

...The BEST part about the meter is that you can chain diodes in series to get SPECIFIC forward voltages, and the meter will tell you when you've hit your target number.

Lots of folks get hung up on diode types when (IMO) the forward voltage is the only thing that matters. Take something like the Zendrive, and measure its series diode combinations, and you will realize that Hermida was not just slapping diodes together - the Vf's (and they are asymm, too) are very specific, and well thought out. The "larger" Vf is just slightly less than the lowest LED that you can get - it's about 1400mV or so (some red LED's are around 1600mV). And - two 1N4148's in series gives you around 1200mV. So Alf realized that neither of those diode types would work. But he found that by chaining together 2N7000's (mosFET's wired up to just use the internal body diodes) and BAT41's would give him the "right" numbers.

...You will find similar things going on in the FD2, the Gristle King, and other pedals. Since the forward voltage of diodes is "fixed," you have to use series combinations to get values that you target. The "problem" with LED's is that they typically have forward voltages that are too high, and there is no way to make it lower. You need different diodes for that.

11G- very thorough and complete answer. Thanks a ton.

One last thing, just to make sure I've got this: 2 diodes in series adds their forward voltages, and in parallel it...? Would I ever use them in parallel?

One last thing, just to make sure I've got this: 2 diodes in series adds their forward voltages, and in parallel it...? Would I ever use them in parallel?

You always need at least a pair of diodes in parallel with the polarities opposing each other.

But what you are referring to (if I understand you) is just altering the Vf, which yes - must be done in series.

Just think of the diodes like flashlight batteries. A pair of 1.5V C's in series will give you 3V.

The Timmy is a good example of what happens when you put a single diode parallel to a series pair, because Paul C. has switches to do just that - if you "switch in" a diode (or combination of them) with a smaller forward voltage, it is the one that is "active." The diode(s) with the bigger threshold aren't doing anything. It's kind of cool, because it means you can come up with switches that don't have to take some of the diodes out of the circuit.

Lots of people will just get a toggle and only do 2 variations of clipping diodes. I actually suggest getting a rotary and putting in 4 or more combinations. Actually, the best thing to do is to temporarily run 2 wires out of the pedal and temporarily sub in diode combinations. With a meter, you can write down the numbers of the combos, and then base your next experiments on your observations up to that point. Eventually, you could guesstimate what combination might work best for you.

10Gauge posted this once, in part.

....and I cannot BELIEVE that I called you 10 Gauge.... I know better - my apologies!

Good to see you back!

Yeah, I get the whole opposing-polarity deal. I was just wondering what it would do to the forward voltage if I were to attach three diodes in parallel: two with the same polarity, and one with reversed polarity. But it seems there is no real use for this other than to possibly aid in acquiring precise forward voltages (as you elaborated upon earlier).

....and I cannot BELIEVE that I called you 10 Gauge

Don't sweat it - it's a goofy handle!

I chose it because of guitar string gauge relevance, and because it's kind of like a "wrong shotgun," if that makes sense, and kind of is a suitable description to those who know me (i.e I come across a little strong at times, and I'm a "little off"). :lol:

Teleworshipkid -
Diodes in parallel, same polarity, gives the same voltage drop as with one diode. You're correct, as you don't gain anything by doing that. If the diodes are the same type, you have increased the current capability, useful in power supply applications. With dissimilar diodes, the one with the lowest forward voltage conducts first and the other is essentially open circuit, doing nothing (until the first one fails).

Thanks .011 for the reply!
One final question, can this be correct? These silicone diodes http://www.ebay.co.uk/itm/100-OA128-vintage-silicon-diodes-TELEFUNKEN-/380372428790?pt=UK_BOI_Electrical_Components_Suppl ies_ET&hash=item588ff6a7f6 say 150mA in the listing,but you guys measure them in mV.Is this a typo where he meant to say mV?And if so isnt that way low for a silicone diode?

The listing says 35V, 150mA. That would be be peak reverse voltage, and forward current, respectively. The forward voltage is not listed, other than it is silicon.
Similarily, the 1N4001 is 50V, 1A; the 1N4007 is 1000V, 1A.
For signal diodes, the 1N4148 is 100V, 200mA.

^ Thanks Fenderlover!

The listing says 35V, 150mA. That would be be peak reverse voltage, and forward current, respectively. The forward voltage is not listed, other than it is silicon.
Similarily, the 1N4001 is 50V, 1A; the 1N4007 is 1000V, 1A.
For signal diodes, the 1N4148 is 100V, 200mA.

Yes - all numbers that are useless in regards to using diodes to clip an analog audio signal.

Is the selling point with those diodes that they have some kind of NOS magic or something? Because again, it's just a simple passive component, at least in regards to how it is used in a dirt pedal.

Some will argue that different diode types have a different "knee," and I won't particularly dispute that. There are certainly bound to be things that there is no data on, because it is irrelevant to the intended design of the component (they probably have no clue that anyone would want to use them for such a purpose!).

...But I still say that the forward voltage is really the only dominating characteristic of what gives the different types of clipping effects. While my testing has not been scientific, when I replicated identical forward voltages with (often vastly) different diode types, the audible results were consistent. IOW, no magic from NOS diodes (I've got a bunch of old ones from different gizmos that have been scrapped over the years, and diodes are typically an easy thing to extract), not even really any magic from mosFET transistors used as clipping diodes (in the "traditional" sense, like they are used by Fulltone and others) as long as the forward voltage is the same.

If you're going to concentrate on grabbing vintage/NOS/obsolete stuff, I'd personally recommend things like the LM308N chip. The design has been horrendously obsolete for decades, and Texas Instruments just acquired National Semiconductor. The big issue right now is also obviously RoHS, and many of these components either introduce lead and such in the tooling process, or they are simply within the components themselves.

Now, if the diodes are cheap and you are just having fun collecting them, that's fine. And I do understand that some folks just love the aesthetic of putting old components in a pedal, or using axial caps and such.

...I just personally only care about how it sounds. There are so many great modern diodes like Schottky types that work great in combination with common types such as silicons or even a simple BJT used like a clipping diode (its forward voltage tends to be just a little higher than a 1N4148).

There are also some Vf's that lots of folks haven't "experienced," because they are either higher or lower than it. For instance, 850mV - 950mV tends to work fairly well in some pedals, IMO.

^ Hey .011, I just got em' cause they were fairly cheap,and it turns out they do sound good in my TS9 :-)
Definately not for their NOS mojo or any such nonsense :lol:

Definately not for their NOS mojo or any such nonsense :lol:

I hear ya! :smile:

To make something clear, LEDs run on current, not voltage. The forward voltage varies with current through them, age, temperature etc. To correctly bias an LED, you can put a resistor in series (simplest way) or use a driver to PWM the diode to change it's light output or change it's current (analog dimming), or combinations of these.

The datasheet will give the forward voltage for a place to start. Decide the current you want, and given your input voltage, you can subtract the forward diode voltage and divide the voltage remaining by the current you're going to run and you have the resistor value to put in series.

Those are good directions for turning on an LED, but the discussion was primarily, or exclusively, about using various diodes as clipping diodes.
Diodes used in the feedback loop of an op amp often have no series resistor at all.
I guess the thread is now closer to the all-encompassing "all-about diodes" title, though there are far more applications than people that would be interested. (I looked up varactor diodes yesterday for RF tuning...)

To make something clear, LEDs run on current, not voltage. The forward voltage varies with current through them, age, temperature etc. To correctly bias an LED, you can put a resistor in series (simplest way) or use a driver to PWM the diode to change it's light output or change it's current (analog dimming), or combinations of these.

The datasheet will give the forward voltage for a place to start. Decide the current you want, and given your input voltage, you can subtract the forward diode voltage and divide the voltage remaining by the current you're going to run and you have the resistor value to put in series.

This info is only relevant to make an LED function as a status indicator, and not for clipping an analog audio waveform. Current is (mostly) irrelevant in this situation, even if they operate off of them for "normal operation."

IOW, there is no biasing or any of that stuff to contend with.

The data sheet is all but worthless (IMO) in regards to LED's as clippers. The reason is simple - a forward voltage is just one small component regarding what is going on. It's like knowing how your car will do in a slalom course based on the tech sheets for your tires' ratings.

Along with that, there are variations despite the published numbers. The only way to know for certain what any given forward voltage for a diode is, is to hook them up to a meter that tests diodes. Since LED's are at the "upper limit" of what would be used as a forward voltage figure in a typical clipping diode circuit, it is a bit extra important to actually measure them.

...For instance, I have red LED's that vary from 1600mV up to about 1800mV, and that upper end is supposed to be what you will find on a data sheet for a yellow or green LED (since their physical composition is different). A pair of 1N4148's have a pretty "hefty" forward voltage figure of around 1200mV when measured in series. Every 50mV or so will typically produce an audible difference, so you can imagine what 200mV or much more might do. The difference between two LED's could be as large as the forward voltage for something like a 1N34A germanium diode!

As much as I love data sheets, they aren't very helpful in this situation. A meter and your ears will be much better tools. IMO.

To make something clear, LEDs run on current, not voltage. The forward voltage varies with current through them, age, temperature etc. To correctly bias an LED, you can put a resistor in series (simplest way) or use a driver to PWM the diode to change it's light output or change it's current (analog dimming), or combinations of these.

The datasheet will give the forward voltage for a place to start. Decide the current you want, and given your input voltage, you can subtract the forward diode voltage and divide the voltage remaining by the current you're going to run and you have the resistor value to put in series.

Placing a resistor in series with an LED does not bias it. The LED will light (briefly, depending on the source voltage) without it. What a series resistor does is protect the LED by limiting the amount of current that it sinks.