Alternatives to diode clipping???

[Boy Howdy]

I have a Zendrive and a Barber Burn Unit Small Fry that I borrowed from a friend at my house right now. I've been taste testing these against my Nady TO-2 (a high voltage tube pedal (1 tube) with removed clipping diodes), a DOD 250 overdrive (with greatly emasculated diodes and the vintage spec. .01 input cap for max fatness), and a stock Dano Drive.

I think the Zendrive is extremely overrated. I like the Small Fry somewhat better. The Dano actually sounds purty great with the distortion below halfway. The DOD sounds really great if I have just a very little clipping on it. The Nady doesn't distort much since I removed the diodes but what distortion it does make absolutely distroys the others. The difference is obviously the clipping diodes. I hate the damn things. They sound fuzzy and buzzy, have unnatural decay characteristics, they kill your dynamics, they have a too easy sustain that just doesn't sound real. Shall I go on? Great strides seem to have been made of late with the EQ aspects of pedal design, but always there's that attack of the killer bees thing going on underneath.

Is there an alternative to diode clipping in pedals? Surely we can do better than this.

I've always just used clean boosters to overdrive the amp (which is essentially how I'm using the DOD and the Nady) but the volume, my god, the volume! Sometimes it's just not practical.

What say ye?

[Jazz P Bass]

clipping diodes

how about a blue LED?

[Chuck H]

I've always just used clean boosters to overdrive the amp (which is essentially how I'm using the DOD and the Nady) but the volume, my god, the volume! Sometimes it's just not practical.

First, diode clipping doesn't have to be buzzy and fizzy, or have an unnatural sustain. And tube clipping isn't always round and warm. It too can be buzzy and fizzy. Don't give up on pedal devices. There are many good ones out there. Some do employ alternatives to plain diode clipping like transistor clipping and LED's which both produce harder clipping than tube saturation (but not tube cutoff) but have a more organic feel than common diode clipping devices.

Speakers can have a profound effect on distortion character. A speaker that is too bright will characterize all the bad stuff happening in the higher order harmonics of diode clipping circuits. But a dark(ish) speaker seems to even things out and hide the uglies. Once you get there you may find some good things about diode clipping. It seems to stay tighter on the bottom end even when a fatter kind of distortion is used. small amounts of diode clipping are very good at getting that "still sounds like the clean tone but has an edgy sustain" sound. Marshall has built it's sound on internal diode clippers for over a decade.

Having said that...

I also prefere tube distortion, even if I do recognize some merits of diode clipping. As far as things getting too loud, well, thats phase inverter and power tube clipping. You can't get exactly that sound without overdriving a power amp circuit. I used to get really good results with a master volume on my Marshall 1959RI. I installed the MV and ran a BOSS GE7 EQ pedal set to roll off a little highs and lows as a clean mid boost to overdrive the front end (channels jumpered of course). It was a very solid and "classic" rock sound with generous sustain. Preamp tubes do clip somewhat like power tubes. But with the exception of crossover distortion (where the AB waveforms meet) there is very little cutoff clipping with power tube distortion. Overdriven preamp tubes do clip harder because of the "cutoff" portion of their class A operation. You might try a master volume and a clean boost pedal. If you find the cutoff clipping of preamp tube distortion unpleasant, like diode clipping, then your only other alternative is to buy an attenuator and enjoy your power amp distortion.

Chuck

[pdf64]

Or fit a dual pot type mater vol between the phase splitter and power tubes. They can work really well.

[Mark Hammer]

One of the reasons some folks do not like diode clipping is because one pair or locus of diodes is doing all the heavy lifting. When distributed over multiple loci,the sound can become much more palatable.

here's a circuit from Bernie Hutchins that presentsd an interesting approach to diode clipping:

And another from the late Fred Nachbauer:

[Boy Howdy]

Actually, I couldn't find blue LEDs locally when I worked on my DOD, so I went with yellow in series with the silicon diodes already in the unit. Still distorts plenty and/but greatly increased its boost capabilities.

[Boy Howdy]

Yeah, Chuck, thanks. I've tried all of that at one point or another over the years. But with old amps some options work better than others. I modified my Traynor, put in a slightly tweaked JTM 50 circuit, and I converted two of the input jacks into a line out/in loop so I could use an outboard master volume. That gets me there for certain sounds, but then trying to get back to clean sounds ... you need channel switching.

And then with my Sunns, they don't really distort very well (I swear I think Leslie West used something other than Sunn to record 'Mississippi Queen' - anybody know that story?). They only have one preamp tube, so not a lot of preamp gain to be had. They sound okay for distortion, nothing to write home about though. But I do love those amps for clean stuff. (note: the 1st stage cathode bypass cap needs to be changed to get a REALLY great guitar sound - I went with a 2uf.). If I could only come up with a great sounding and versatile distortion unit, I'd be in hog heaven with those amps.

[Boy Howdy]

Hey Mark, have built either of those? I'm a rank (yes, I stink) amateur when it comes to electronics. I'm not real conversant with some of this stuff. Looks like that's only part of the circuit in the top example, am I wrong?

I spent more time over the weekend with these borrowed (and my own) pedals. That Dano Drive is an amazing little pedal, makes me want to get the OCD it's supposedly a clone of. I especially like it set with the gain set all the way to min. Sounds like a small tube amp at that settting. But of course I really need something to do the high gain, singing sustain thing. Something like the Zendrive. Which sings very nicely, but it's tone is just so thin and unnatural. The Barber sounds much better (tonewise) for that kind of thing, but my god, it has so much of that diode "aftertaste" as I call it, as to be almost unusable.

Does anybody currently make these multiple loci pedal circuits?

[Mark Hammer]

No, I haven't built either, but I have played with "dual clipper" units that insert diodes at more than one place.

The Big Muff Pi is the "classic" dual clipper, with two stages using a diode pair for clipping. But there are other ways. For example the softness of the clipping can be varied by means of a small resistance in series with the diodes. So, imagine you have two op-amp gain stages, each with a diode clipping pair and a resistor in series. The first stage might use 2 germanium diodes to set a low clipping threshold, but make the clipping softer by means of the resistor and the use of modest gain. Now, you follow this with a second stage that applies more gain - though not very much is really needed - and uses a pair of silicon diodes or even a 2+2 diode arangement, again with a series resistor to "soften the blow".

Because the 2nd stage has a higher threshold for clipping, what happens is that the first stage clips a bit, and when you really dig into the note, the 2nd stage then adds more clipping to an already slightly clipped signal.

Important in all of this is proper toneshaping and filtering of the signal. The first stage should have much of the upper treble rolled off in order for this arrangement to sound good.

here are some examples of my own experiments: http://hammer.ampage.org/files/The_Crank.gif
http://hammer.ampage.org/files/Chaos.gif

[teemuk]

I personally like diode clipping. What is not often realised is how many different variations it can have:

- Diodes can be referenced to DC voltages to increase the clipping threshold voltage. I have also experimented with an idea of dynamically varying this DC reference voltage based on signal's RMS amplitude. Sort of a "sagging" simulation.
- Diodes can be connected in series, in series with resistors, in series with capacitors and in series with different types of resistive and capacitive "ladders" (and combinations of all those) to shape the transfer curve.
- Diodes can be replaced with FETs, Zeners, LEDs, gate-source junctions etc. (Although I have found that this has quite minor effects unless change in forward voltage is a drastic one)
-Diode clippers used in a feedback loop can be taken after a resistive divider located in feedback loop to increase the clipping threshold. They can be even fitted to active tone controls.
- You can phase split the signal, clip each of the two signals individually and then combine the signals. When the signals are summed the "clean" signal half will correct the "distorted" half somewhat, resulting into softer clipping effect similar to that taking place in push-pull tube amps.
- Asymmetric clipping arrangements can be used and they can be divided to multiple capacitively coupled stages to make the signal interact with DC offsets accentuated by asymmetric clipping process. Combine this with the previous scheme and you can even introdude crossover distortion caused by DC shifts. Just like in push-pull tube amps.
- Pre and post clipping equalizing can be used to create drastic effects on tone. Using effective and appropriate equalizing scheme is pretty much the heart of any distortion-based circuit. Even the cunniest distortion stage will most likely sound extremely pathetic if it's combined with a high-bandwidth equalization producing farting low end and a screeching ice pick top end.
- etc.

It's all about using these types of diode clipping techniques (and other I likely forgot) effectively. There is no single tone in diode clipping. The tone comes from the whole circuit.

[Mark Hammer]

Well-stated. Thanks.

[Boy Howdy]

The tone comes from the whole circuit.

So it seems. I took the diodes out of my DOD250 and that awful fizzy decay sound was still there. In this case, at least, it appears I have wrongly accused. The distortion in the IC (presumably) is not very pleasant (anybody got any suggestions for improvements?). I turn the gain down until I don't hear much, if any, distortion. Makes a really nice booster though.

That said, I hear the same sound (that fizzy, crackly decay), but less pronounced with other pedals, of course. The Small Fry is particularly bad. The Zendrive not so much, but then they roll off the high end somewhere other than with the tone control. BTW, Mark, I hate that sound. Always have. I like total clarity in the note. I have another old Nady unit, the original TD-1 (I have the recent piece of crap Tonebone copy too, hate that thing). The original TD-1 has a 470pF cap going to ground just before V2 and before the clipping diodes. Totally ruined the sound for me. I cursed that thing for years before someone finally pointed out to me that that cap, which passes for a tone control, the only tone control in the unit, was responsible. I ripped that sucker out and am much happier with it now (despite the poor sounding diode arrangement). I've heard this phenom in many many pedals, this distortion -> treble rolloff -> distortion thing (often with no additional final tone control). I call it crunchy mud. Makes the skin crawl. I realize that in the hands of a designer with ears this doesn't necessarily have to be the case though. No offense.

Thanks for what you wrote above teemuk. Lots to study there.

[Boy Howdy]

Hey Mark, I have some questions about this Dogzilla circuit, if you don't mind.

Is that just an EQ unit? Why is the diode configuration called a limiter? Is it intended, as with a limiter proper, to prevent distortion, in which case it might be of little use for the purposes being discussed here? (is that a question?) This multiple loci diode arrangement, is the theory to distribute the distortion over multiple frequency bands? It looks like a simple enough circuit (the diode configuration, that is), any reason I couldn't just hack that into a pedal? Why isn't this or something similar done by pedal makers? If it deals with fizz it would be well worth it, so it would seem.

Oh, another one: I noticed on the OCD and the Dano Drive pedal they have caps bypassing the diodes (.01 for the OCD and .001 for the Dano), is the idea here to pass some unaffected signal, to give a cleaner, more dynamic edge to the tone/signal/note?

Thanks in advance.

[Mark Hammer]

Sorry for the slow response. Out of town.

In theory, a simple diode pair, as used in so many distortion units, IS a limiter, in that it clamps the amplitude at a fixed maximum, whilst still permitting lower amplitudes. However, they become distortion units because that simple diode action has a "hard knee" and introduces a lot of additional harmonic content in the process. The multiple diode arrangement for the Dogzilla is to provide a much softer knee, by havng the overall conduction dependant on the joint action of many steps of diode conduction. Ideally, this should provide a cleaner limiting than that provided by a single pair of diodes.

Running a cap in parallel with clipping diodes simply reduces top end. Nobody likes the fizzies.

[RudeBoy]

Why is the diode configuration called a limiter? Is it intended, as with a limiter proper, to prevent distortion, in which case it might be of little use for the purposes being discussed here? (is that a question?) This multiple loci diode arrangement, is the theory to distribute the distortion over multiple frequency bands? It looks like a simple enough circuit (the diode configuration, that is), any reason I couldn't just hack that into a pedal? Why isn't this or something similar done by pedal makers? If it deals with fizz it would be well worth it, so it would seem.

Take a look at the original page describing this. There is a description of why he calls it a limiter. There are also some simulation curves there that show the overdrive effect of this circuit. I modeled it in spice, tried some variations etc. By biasing appropriately it is possible to get a nice asymmetrical limiting effect and distortion (for those that want that sort of thing).