Phylosophy in Designing a High Gain pedal (and his EQ)

[richon]

I was thinking on discusing over the PHYLOSOPHY in designing a High Gain pedal (and the way the EQ is used):

I've been working on a HIGH GAIN pedal (my first time with High Gain Jfets emulating pedals). I know a lot of people would inmediatly think of the Dr. Boogey but in this ocation I started from cero without using other pedals o what I allready know from tube amps.

Well, so I started with one Jfet stage, then adding another stage and always checking the actual stage and the ones before it. I was always using the Jfet at his max ouput voltage (by adjusting the trimer) 'cause when using it a little up or down from his "point" it sounded not clear in all frec (sometimes a little fuzzy in the lows or not clear in highs)

So this induced some Hizz over the 6 stages.... and to take it out I added better filtering and some parallel caps to the trimers...   this way kills almoust all hizz but also builds a ceiling for the HF (loose some highs)... although for HIGH GAIN it's not needed so much highs as I usually do when playing blues.

So now comes the two questions:

Is it a good way to get a good sounding High Gain pedal to set the sound (with a right HF ceiling ) right before the EQ and use an active EQ to boost or cut and control all the frecuencys?

or

Set the pedal almoust without HF ceiling an let the EQ do all the work (although sometimes it seems the EQ looses his figth to control well the highs??



PS: what would be a good "ceiling" of HF for High Gain pedals? 4KHz?? more, less?

[R.G.]

Good thinking. Many do not spend the time thinking about what affects what in an effect.

As a suggestion, you may want to think about the following.

- Clipping a sine wave, the simplest possible signal, produces a series of harmonics of the original sine wave. So for a sine wave of frequency F, you get resulting signals of F, 2F, 3F, 4F... and so on to infinity. Exactly how it's clipped, determines the amount of each harmonic. Symmetrical clipping the top and bottom gives you only the odd harmonics, 3F, 5F, 7F...  This kind of distortion is called "harmonic distortion". Pure harmonic distortion caused by the lower order (2 through perhaps 5 or 6) is not unpleasant, and adds character to the sound. Harmonics over the 6th or 7th are increasingly unpleasant, according to musicologists.

- Distorting in any way (clipping included) a signal which is a mixture of more than one sine wave (that is, all real world signals) of frequencies F1 and F2 produces both a series of harmonics of both F1 and F2, producing 2*F1, 3*F1, 4*F1... and 2*F2, 3*F2, 4*F2...   In addition, it produces the sum and difference of the two frequencies (F1+F2 and F1-F2), the harmonics of both the sum and difference (2*(F1+F2),  2*(F1-F2), 3*(F1+F2), 3*(F1-F2), ... ), and the harmonics of (N*F1+/-N*F2) for N = 2, 3, 4...  This is called "intermodulation distortion", as the original frequencies are inter-modulated to make frequencies that are not harmonically related to the original frequencies. Musicologists say that low order intermodulation is "interesting", as in something like a ring modulator. Higher order intermodulation sounds "muddy" for the difference modulations, and "harsh" for the multiples-of-sums modulations. The frequencies made by intermodulation are not musically related to the original signal.

The trick in getting sweet/smooth sounding distortion for a guitar seems to be limiting the amount of harmonics generated as harmonic distortion and severely limiting intermodulation distortion. This is done in two ways.

1. Pre-filter the input to the distortion process. This removes some frequencies from ever being distorted, and therefore removes all of their distortion products from ever being generated.
2. Post-filter the output of the distortion process. This removes some of the distortion-generated sound signals, and usually makes the result be less harsh or jarring - smoother sounding.

In the extreme, there are distortion approaches that break up the incoming signal into small bands of frequencies and individually distort each band of frequencies, then mix the resulting signals. This helps in removing lots of the intermodulation products. Anderton's "Quadra-Fuzz" is the prototypical example.

Playing style matters as well. By playing only single notes or two-note intervals, the frequency content to the distortion process is limited, so the result is both less harsh and less muddy. "Power chords", the playing of only the bass two notes of a chord, is an example, as are single-note leads.

For a bit of insight about how this works in a production pedal, read my article "The Technology of the Tube Screamer" at GEOFEX, where I outline some of how this approach was applied to the tube screamer design.

Now, back to your questions.

Is it a good way to get a good sounding High Gain pedal to set the sound (with a right HF ceiling ) right before the EQ and use an active EQ to boost or cut and control all the frecuencys?

As a side note, "high gain" is a misnomer in almost all references to it in guitar pedals. What that really means is "heavily distorted".

But to answer your question, yes, both. The more sophisticated approaches to distortion are to both (1) pre-filter the input signal, so the resulting signal frequencies with the biggest levels are distorted more for those frequencies and to (2) post-filter the distorted result to remove any objectionable harsh sounding distortion products.

Set the pedal almoust without HF ceiling an let the EQ do all the work (although sometimes it seems the EQ looses his figth to control well the highs??

The reason the EQ loses this fight is that so much intermodulation hash is generated. Pre-filter, then post filter.

[Speeddemon]

I once fooled around with an Ibanez SM9 (Super Metal) circuit; apparently they really had thinked things through...because I wanted more range from the EQ section and things started to oscillate in the min. and max. settings.

I recommend filtering the signal pre-EQ with a bass-cut, so in case you use and active EQ (post-gain stage), the pedal won't sound as muddy. It cracks me up everytime when someone uses a pedal like a Metalzone with the bass on +6 into a Dual Rectifier, with a Les Paul and complains that his sound isn't tight enough. 

[richon]

when looking it very simple , it is:

keep the good frecs in, let the bad out?

(this is by reducing enough highs and bass to let the guitar frecs intouchable, but killing all undesired?)

and this is by doing it stage by stage?

[raulgrell]

keep the good frecs in, let the bad out?

in, out, in, out, shake 'em all about!

A lot can be learned from tube amp designs... if you look at the 'best' metal distortions out there - the 5150, Dual Rect, ect - they're very simple - lots of gain stages and filtering in between... The stompboxes based on them are the same, shaving off the higher frequencies so that the sound is not too harsh, and taking some of the lower frequencies too so that the distortion is not farty/muddy...

The intermudolation thing RG was talking about is (in my opinion) the big problem with most distortions - Try playing a full chord (as in not power chord) with most distortions and you get an unrecognizable mess of noise - tune it right and you can get some great sounds.

If you take too many high frequencies too soon, you lose out on a lot of harmonics that are fundamental to the sound (take a distortion you like, then turn down the tone and use the neck pickup on your guitar to see its effect)... take too many low ones, you end up with a very thin sound.

I tried a few things to make 5/6 gain stage distortions simply putting a bunch of stages in series and putting filters between them.

[Paul Marossy]

Don't try to squeeze too much gain out of each stage. If you do, you'll also amplify all the noise from the previous stages and it will become very noisy really quickly. Filtering may or may not help with that, depending on the circuit topology and the devices used.

For example, MOSFETs sound nice, but they are rather noisy. Build something with six gain stage using MOSFETs and it can become very noisy. Too noisy to really be practically usable. Adding a bunch of filtering would kill all the dynamics.

[richon]

I know a few ways to reduce the gain (please coment if you know any new ones):

taking the source bypass cap out of a stage (or a few)
using more voltage divider between stages
using high pass filters after the RC formed from the coupling cap and the resistor to ground

besides, taking some highs:
paralel caps to the trimmers at drain
once again high pass filters between stage
increasing the Miller caps (and the RC formed with the  resistor before the Stage)


any more ideas?

is there a way to look or compare a few tube stages to a few Jfets stages to compare how the sound gets changed?

[petemoore]

  Thank you for asking !
  Bass can get loose if let in, a harder HP filter at the input .or. Sprinkling HP filters lightly across the circuit may offer smoother response and tighter differences in attenuation @ certain frequencies. [2 different approaches].
  Treble can be 'let go' and won't tend to 'current swamp' the transistors, controlled 'later' for fuzzytone .or. same as the bass, sprinkling of LP filters more lightly across the path.
  A clone-comparator circuit is something I found handy, similar but modded, then bring the cloned one up to the new clone specs if desirable. The clone and original can be raced around the track to see if the cornering or acceleration is more winning.
  A lot can be done with Jfet's, supply V alterations.
  http://www.muzique.com/lab/buffers.htm
  http://www.muzique.com/lab/lowvolt.htm
  Good reads...

[WGTP]

Some of the "metal" pedals seem to have a midrange boost pre-distortion in the circuit which effectively reduces the highs and lows, and then a notch in the midrange EQ after the distortion circuit.   Some also have GE diodes in the circuit path for crossover distortion. 

[richon]

I don't remember where , but I heard (or read) that "always" the third stage Jfet is where the distortion happens...

Is this real?

what would happend in the next stages? more distortion to? or just amplification of the previous stages? (I suspect more distortion)


PS:(if more distortion is added in the latter stages) would it be good to add very little distortion or just the same as is push into Q3?

[philbinator1]

Side-question for R.G. ..do you have a book out?  Other than the info on your website??

[oliphaunt]

I have been building multi-stage JFET and MOSFET ODs for low to medium gain, also my own design and not based on any specific amp or pedal.  I have used small caps from drain to gate to filter out unwanted high end on multple stages , and have made several designs where a tone control is not even necessary as I have achieved a fairly flat response without losing the detail in the high end.  I also tend to use a near max gain stage for stage one, then less gain in the following stages to keep things from getting 'farty" sounding.  I also use small coupling caps between some stages to filter out lows.  

My theory (it's just a theory, I am a relative beginner) is that less distortion in each stage produces a smoother, more pleasing result with less need for extreme filtering and better dynamics overall.

Also try lower gain JFETs like the MPF102.  They can distort a little more smoothly under some circumstances.

[richon]

so , you made me INSTALL WINDOWS ON MY IMAC!!! to also install LTSPICE  


here are some results:

6 stages and his frecs


5 stages and his frecs


6 stages and his wave


5 stages and his wave



by this I can guess:
-the signal gets "square" after the buffer almoust the same after 5 or 6 stages (always last stage was a "low impedance out" BUFFER )
-with the 6 stages, the signal keeps square a "little" more time.
-the 10K frec is at 8dB in 5stages and at 15dB at 6 stages....  so this maybe means that 6 stages add more highs (and more HF noise, right?)

I'll expect some comments  

[richon]

[richon]

no comments?

[philbinator1]

I would but I have no idea what all that stuff means.     Enlighten me.

[goulashnakov]

I do notice the similarity in the buffer's waveform's on the 5- and 6- stage prototypes.  I also notice that the later common-source stages seem to have a more "angular" clip, versus the common-drain's square shape.  Are you noticing a difference in sound of the two waveforms?  Which clip-shape sounds better to your ears?