Failing to get extreme distortion from op amps -

[Electric_Death]

I've tried soft clipping, hard clipping and combining the two.
I still can't get much more than some chunky grit or decent fuzz, do I really need to add a transistor??
I even added more clipping diodes on the output which yielded a tiny bit more grit but not the metal grinding crunch I'd hoped for.

The circuits are really straight forward. Gain of 1000, popular clipping diodes on the feedback path with a cap and 2 more clipping diodes shunt to ground on the output. With that much gain and two clipping stages, I expected a nice chainsaw grind but it's not even close. The circuits make excellent overdrives and the hard clipping really brings out pinch harmonics but, I was hoping to evolve these simple circuits into saw tooth monsters and didn't even come close

[John Lyons]

How about the Marshall Shredmaster?
ICs are nothing more than groups of transistors inside a small plastic case.
Whatever you can do with a transistor you can do with an IC. They sometimes sound different but....

JOhn

[soulsonic]

You can make a simple textbook Comparator circuit. It turns the input into a straight-up square wave. That's about as distorted as you can get. And once it's become square, it's waveshape can be manipulated with other textbook circuits.

Both the Metal Zone and Metal Muff are further examples of the usual opamp/clipper distortions that are pretty heavy. One difference with those circuits is the fact that they tend to not have straight caps to ground in the feedback loops - they use gyrator circuits to create a distinctly filtered sound.

[brett]

Hi
with al those diodes, I'm not surprised that it's soft.
Harsher clipping comes from diode-to-ground after an op-amp with no feedabck diodes and a minimal feedback cap.
Try something like the Distortion +, substituting LEDs for the usual diodes.  LEDs sound harder, clip harder, because they have a higher ratio of Vf to V"knee".
cheers

[Steben]

Hi, electric d,

I think you are disappointed by the nature of the distortion because of its frequency response. Real metal distortion requires a nice share of tone shaping. You describe a "decent fuzz". Well, "fuzz" means true "flat-in" distortion, often combined with the fact it is transistor clipping. Try some highpass filtering (low values of caps at the input).

[petemoore]

  Hit the diodes 'harder' w/ higher voltage at the input, put a booster in front of the clipping circuit.
  DiST+?.

[R.G.]

The circuits are really straight forward. Gain of 1000, popular clipping diodes on the feedback path with a cap and 2 more clipping diodes shunt to ground on the output. With that much gain and two clipping stages, I expected a nice chainsaw grind but it's not even close. The circuits make excellent overdrives and the hard clipping really brings out pinch harmonics but, I was hoping to evolve these simple circuits into saw tooth monsters and didn't even come close

It's the diodes. The diodes are slowing down and rounding off the clipping.

First - you can't use the (+) input. This forces unity times the input signal to appear at the output. You must use an inverting stage, input through a resistor to the (-) input and the (+) input tied to a bias voltage.
Second - don't use feedback diodes. Let the opamp bang against the power supplies. This requires using an opamp that recovers from overload well, so use something like the TL07x series.
Third - use a pad down on the output, because the signal will be very big.

The output will be razor sharp and flat when the opamp exceeds its output swing capabilities.

See if that helps.

[Mark Hammer]

Steben is absolutely right.  The flavour of a fuzz is not JUST a function of how much gain is applied, but where it is applied.  It is the relative emphasis on certain frequency bands that draws our attention and gives a tone an "aggressive" or "smooth" or "scary" tone. 

I think a perfect case in point is the venerable Proco Rat.  There, a simple single noninverting op-amp stage is used but with two parallel "ground legs" from the inverting input pin.  Use of these two ground legs means that as the gain pot is turned up, there is considerable more gain applied to the highs (content above roughly 1.5khz) than to the rest of the signal.  According to one schematic, at max gain setting, there is an amplification of 2143x applied to the highs, while there is "only" a gain of 268x applied to the lower-frequency content.  This gives the naturally lower-amplitude highs a chance to "compete" with the lows for clipping.

In other cases, the use of a midscoop helps to bring the frequency bands to prominence that serve to create the character of the tone.  Here, a good case study is the Univox Superfuzz and all of its cousins and siblings (Shin-Ei, etc.).  Removing certain midbands shifts attention towards the top end and rumbling bottom, which gives an ominous sense to its tone.  Guitarist Poison Ivy from the Cramps once described it as sounding like death. 

A circuit such as that of the Rat, in tandem with something like a well-situated (i.e., the notch location that does what you need/want) variable mid-scoop might be just the ticket.

[frankclarke]

Multiple distortion elements in series would work, if one by itself doesn't. The Rat has enough distortion for me, but maybe not for you.

[Electric_Death]

Thanks guys, that helps a lot. Someone had told me prior to pull my grounding cap if I really wanted beef up the distortion when I showed them the schematic and then filter out the unwanted frequencies in the output stage. Currently I'm running a clean gain of 1000 into a gain of 100 with both soft and heavy clipping on the 100 gain stage. Diodes with a filtering cap on the feedback path and diodes as well as LED's shunt to ground on the input.
Makes a great overdrive but I want to evolve it into a distortion suitable for using as part of an amplifier preamp. I think this bas circuit will serve well as a "hot" switch to kick in when I want to overdrive the main distortion.

R.G. I'd been planning on converting it to an inverting input for high gain distortion because I've heard this is a good start for designing an op amp based distortion that is input sensitive so I can go from chimey to distorted with the volume of my guitar. My reluctance has been due to reading that inverting input gain stages require high value resistors to achieve the desired gain and higher values mean noise issues when you crank the distortion level.

John Lyons I'm wanting to get this distortion so it's input sensitive meaning what I just referenced, the ability to go from clean/chimey to heavily distorted with the volume knob of my guitar. To me this says a ridiculous amount of gain in the clean first stage, say 10,000?, and second stage an inverting medium to high gain with hard clipping using LED's and no caps until the output stage. If that doesn't quite do the trick, maybe a photo transistor sneaked in for some compression will do the job?

Frequency wise I'm happy, feeding one gain stage into the next allows for easy manipulation of what gets boosted and distorted more by simple input filtering on the second stage. The fuzz I spoke of comes from toying with some passive filters I've come up with loosely based on the big muff tone control. Since this circuit is to evolve into a high gain distortion however, I can do away with about half the design characteristics and incorporate some of the ideas thrown at me here.

Much obliged 

[R.G.]

R.G. I'd been planning on converting it to an inverting input for high gain distortion because I've heard this is a good start for designing an op amp based distortion that is input sensitive so I can go from chimey to distorted with the volume of my guitar. My reluctance has been due to reading that inverting input gain stages require high value resistors to achieve the desired gain and higher values mean noise issues when you crank the distortion level.

Darned unfriendly of them to let you hear that and not tell you why, don't you think?

"Cleans up" is a characteristic of going into/out of distortion at a signal level and source impedance level that's inside the range of signal from the guitar volume control. In effect, the Thevenin resistance of the volume control becomes an additional series impedance into the circuit. It's that aspect that lowers the gain at the same time it lowers the signal voltage. That's why some Fuzz Faces "clean up" well - the series resistance of the volume pot acts with the feedback of the circuit. If the range of impedances or signal levels is outside the transition zone of the circuit from clean to dirty, then you don't get good cleanup.

First-glance inverting stages need high value feedback resistors to get really high gains. Dig a little deeper and you'll realize you can use a Y-form feedback network to give high gain without high value resistors. The output of the opamp feeds a voltage divider to ground. The junction of the voltage divider feeds a resistor to (-). The output is attenuated a lot without using high value resistors, thus you get lower thermal noise. Pretty straightforward.

[Electric_Death]

I had assumed the volume pot played a role with the gain set resistors with inverting inputs but unless you're simply building from kits, I'd think that's an obvious. I'm also the guy that insists on installing 2 meg linear on all my guitars but, I'll also insist the the volume increase and roll off is apparent with every single digit increment you turn it. I'll go on record for saying this numerous times now, I'll start out with the stock 250 or 500 k but generally after a month or maybe 2 at the longest, I always end up switching to a 2 meg linear or the rare occasion a 1 meg if the pickups don't seem to taper properly.
I've yet to have noise issues and the pickups sound looser and more open which is what I like.

Anyway I assume with these much higher value variable resistors, I'm going to experience a broader range of "clean up" manipulation depending on where I set it. I'd actually diagrammed a Y type formation just in theory without having known just quite how it'd work so I'm definitely onto the right track.
Clearly the next step is building the circuit and seeing if it will do what I want

[R.G.]

but unless you're simply building from kits, I'd think that's an obvious.

It's not obvious even to 99% of the people here. Never bet on how observant humans are - or on how observant they're not. You'll get creamed both ways.

I'm also the guy that insists on installing 2 meg linear on all my guitars but, I'll also insist the the volume increase and roll off is apparent with every single digit increment you turn it.

That's more a function of the pot taper than anything else.

I'll go on record for saying this numerous times now, I'll start out with the stock 250 or 500 k but generally after a month or maybe 2 at the longest, I always end up switching to a 2 meg linear or the rare occasion a 1 meg if the pickups don't seem to taper properly. I've yet to have noise issues and the pickups sound looser and more open which is what I like.

That's because the pickup is in parallel with the pot in most instances, and the net series resistance is 1/4 of the pot value; that makes it 500K for a 2M pot. At exactly half the pot resistance on both sides, the signal is cut by half and the top and bottom halves of the resistance are in parallel.

The pickups sound more open because they are less loaded.

Anyway I assume with these much higher value variable resistors, I'm going to experience a broader range of "clean up" manipulation depending on where I set it. I'd actually diagrammed a Y type formation just in theory without having known just quite how it'd work so I'm definitely onto the right track.

Whether there's a broader range of cleanup depends on the rest of the circuit. First the clipping in the circuit has to HAVE a range of cleanup. Then the other impedances have to be such that you have some range on the pot going across the clipping knee.

Clearly the next step is building the circuit and seeing if it will do what I want

Yep. As we used to say, Mother Nature waits for you at the end of the assembly line.