The Goat: a brutal high gain fuzzstortion

Started by Fancy Lime, March 26, 2020, 06:50:54 AM

Previous topic - Next topic

Fancy Lime

Is your Big Muff too polite sounding? Does your Rat lack chest hair? Is your HM-2 too inflexible and not punchy enough? Try

The Goat




Please excuse the hastily drawn schematic. I checked the values by eye but may have missed some errors. Please let me know if something looks fishy to you.

As you can see, this takes hints from a lot of classic designs, nothing too special there. The "trick" that makes it tick is the deliberate and in some parts somewhat unusual frequency shaping.

The input and output buffers were included for easy addition of soft switching but can be left out when using true bypass switching. If the input buffer is left out, R5 should be increased to 1M or 2M2. If the output buffer is left out, you might want to increase C17 to 470n or 1u. The input buffer can be almost any (low noise) JFET. "Sk30, 2SK170, J201, J112, J113, or 2N5457 all ought to work but were not tested.

I actually used NJM2068 opamps but almost any low-noise BJT opamp will do, 5532, 4580, even the old 4558 (although that one will be noisier).

The Z-switch should be wired such that both legs are connected to ground when the switch is in the middle position and either one is connected in the outer positions. You can use two separate SPST switches instead.

Diodes D3-D9 can be any Silicon diodes but I found Zeners to sound the most interesting. Somehow more brutal than 1N4148's. Different voltage Zeners give different sounds due mostly due to their different capacities. C9 may need to be lowered when using lower voltage Zeners. The clipping threshold is not affected by the Zener voltage because they are only used in forward configuration.

The Mids control is a pure cut and is neutral when in the max position. It cuts an increasingly deeper notch at around 250Hz as you turn it back from max to about 12 o'clock. Below 12 o'clock, the notch starts shifting to higher frequencies while a boosted hump appears between 100 and 300Hz. When using This thing with guitar, the upper half of the range is most useful. Cleans out the lower mid mushyness without loosing the important upper mids. With bass, the lower half of the range gives a solid low mid punch while getting the bass out of the guitars hair in the upper mids.

D10 and D11 can be any Schottkys but I found BAT41's to have the ideal threshold to allow near-maximum gain even with single coil pickups while adding some nice crossover distortion. The gate switch is an SPDT.

Treble allows boost and cut and is somewhat interactive with the mid control, when the latter is between 12 and minimum position.

I am excited enough about how this thing sounds to be planning to have professionally made PCBs. Preferably after some optimization based on feedback from the community, wink, wink, nudge, nudge (I am talking to you, iain!). This is probably a good opportunity for me to finally jump into SMD. Is there any good primer on audio SMD design? Things like available form factors, cap and resistor types and how to identify them (sellers often don't tell you anything other than "it's a resistor" or "its a ceramic cap", which is only marginally more useful than saying "it's some stuff" unless there is implicit information which I am not aware of), soldering methods and pros and cons thereof, and so on. It may end up being a combination if SMD and TH to allow socketing for the diodes and maybe some caps for some customization by the user.

Cheers and enjoy,
Andy

My dry, sweaty foot had become the source of one of the most disturbing cases of chemical-based crime within my home country.

A cider a day keeps the lobster away, bucko!

antonis

Are you sure U1B & U2A are biased at 0V..??
"I'm getting older while being taught all the time" Solon the Athenian..
"I don't mind  being taught all the time but I do mind a lot getting old" Antonis the Thessalonian..

Fancy Lime

Thanks Antonis, good catch! Now corrected.

Also: I forgot to mention that the output buffer can of course be pretty much any NPN BJT. 2N2222, 2N5088, 2N3904, whatever you have lying around. High gain, low noise types like the BC550C, 2N5089, MPSA18... will theoretically give the closest-to-1 gain and best noise performance but the difference will matter fairly little in actual practice in most rigs.

Cheers,
Andy
My dry, sweaty foot had become the source of one of the most disturbing cases of chemical-based crime within my home country.

A cider a day keeps the lobster away, bucko!

antonis

#3
At a first glance, both Input & Output buffers could be omited..
(470k input impedance & 25k - for Vol pot set to 50% - ouput should be considered just fine..)

Of course, buffers almost never do no harm..  :icon_wink:

I'm not aware of NE5532 input bias current, so I can't tell for 470k bias resistor value voltage drop..
(if any, due to bipolar inputs..)
edit: for a typical value of 200nA, voltage drop should be almost 100mV so minimum input bias voltage offset should be about 80mV [200nA X (470k - 100k pot full range)]
For maximum value of 800nA, that offset should be up to 300mV or 400mV for Gain pot fully CCW(*)..

(*) You should interchange Gain pot lugs 3 & 1 for a "conventional" CW gain raise.. :icon_wink:
"I'm getting older while being taught all the time" Solon the Athenian..
"I don't mind  being taught all the time but I do mind a lot getting old" Antonis the Thessalonian..

Fancy Lime

Hi Antonis,

you are right again about the gain pot, thanks!

But I am afraid I do not follow the rest of your post. What voltage drop? What is "input bias voltage offset".

Cheers,
Andy
My dry, sweaty foot had become the source of one of the most disturbing cases of chemical-based crime within my home country.

A cider a day keeps the lobster away, bucko!

iainpunk

#5
hey,

looks nice, i guess that this was the pedal you were talking about in the gorn overdrive thread?

i have 3 questions/suggestions

1] what is the frequency response of the first gain stage and the switches?
i gather that the active bridged "lower case pi" filter (from the Topic: What's this notch filter called that I just "invented"?) is there for the same reason as the gap in the HM2 eq response, to take out the mud? (also, i think i forgot to click post on my name suggestion(the capital pi looks like a normal pi filter, while this filter looks more like a lowercase pi, but bridged.))

2] i really like the gate switch, but i'm not a fan of germaniums for this purpose, Si diodes are more to my taste when it comes to crossover distortion style gates, or even LED's. My current project is a boost/filter/gate where the crossover gate has a 3way (on off on) switch choosing between (on left)none, (off)leds or (on right)1n4148's for more flexible gating. the leds are fine for humbucker guitars, while the preceding circuit has way less gain than your goat has, i dont think the germanium transistors will have much effect on the noise floor, especially with the gain set higher.

3] maybe a little bit of mid rage boost somewhere between 1 and 2 kHz, something like 6db (or 30db if you like adventure) might help the pedal cut through even more. especially because it will lose some of that in the broad scoop in the lower mids, even if the presence is not in the middle of  the scoop, this might add a bit of "presence" to your pedal. i suggest changing the treble control opamp to a sallen key resonant lpf with a q control from 0 to 6, ill crunch the numbers for you if you'd like
friendly reminder: all holes are positive and have negative weight, despite not being there.

cheers

iainpunk

i came up with this:



when the pot is at 100 ohm, the -3dB is at 9kHz and no resonance
when the pot is at 500 ohm, there is a resonant peak at 5kHz and it boosts 2dB
when the pot is at 1k ohm, there is a resonant peak at 4kHz and it boosts 4dB
when the pot is at 5k ohm, there is a resonant peak at 2kHz and it boosts 8dB
when the pot is at 10k ohm, there is a resonant peak at 1.4kHz and it boosts 9dB

this shows me that the pot should be log. and that the control is somewhat wah-like, but at higher frequencies. this could be used to precisely dial in the presence. i have used something similar in an overdrive/distortion i build. it was loosely based on the idea of a kl0n cƐntaur, where the clean bass and treble are mixed with distorted mids. if i were to build a bass pedal, id reverse those, clean mids and compressed bass and treble
friendly reminder: all holes are positive and have negative weight, despite not being there.

cheers

Fancy Lime

#7
Quote from: iainpunk on March 27, 2020, 12:36:35 PM
hey,

looks nice, i guess that this was the pedal you were talking about in the gorn overdrive thread?
Exactly.

Quote
i have 3 questions/suggestions

1] what is the frequency response of the first gain stage and the switches?
i gather that the active bridged "lower case pi" filter (from the Topic: What's this notch filter called that I just "invented"?) is there for the same reason as the gap in the HM2 eq response, to take out the mud? (also, i think i forgot to click post on my name suggestion(the capital pi looks like a normal pi filter, while this filter looks more like a lowercase pi, but bridged.))
The "tight" side has a corner frequency of 720Hz, the "fat" side is flat down to 21Hz. With both engaged you have about 13db more boost over 720
than at under 150Hz. Depending on the gain and tone settings this switch provides vastly different characters.

Yes, this thing is where the bootstrapped bridged pi filter comes from and yes, it is designed after the frequency response of the HM-2 and for exactly the same reasons.

Quote
2] i really like the gate switch, but i'm not a fan of germaniums for this purpose, Si diodes are more to my taste when it comes to crossover distortion style gates, or even LED's. My current project is a boost/filter/gate where the crossover gate has a 3way (on off on) switch choosing between (on left)none, (off)leds or (on right)1n4148's for more flexible gating. the leds are fine for humbucker guitars, while the preceding circuit has way less gain than your goat has, i dont think the germanium transistors will have much effect on the noise floor, especially with the gain set higher.
I'm also not a big fan of germanium diodes as crossover clippers/gates, that's why I used BAT41's  ;). These are Schottky diodes and are on the high side of forward drops for Schottkys we tend to use in effect design. I liked them better than the 1N4148's I also tested but you may like the Si better. I found the BAT41's to be somehow rougher, earthier, less fizzy. It would also be no problem to include both options, which I will probably do in the next version. If you use Si diodes, you may want to lower R15 from 2k2 to 1k5 or even 1k to maintain a constant volume between the gated and non-gated positions.


Quote
3] maybe a little bit of mid rage boost somewhere between 1 and 2 kHz, something like 6db (or 30db if you like adventure) might help the pedal cut through even more. especially because it will lose some of that in the broad scoop in the lower mids, even if the presence is not in the middle of  the scoop, this might add a bit of "presence" to your pedal. i suggest changing the treble control opamp to a sallen key resonant lpf with a q control from 0 to 6, ill crunch the numbers for you if you'd like
That is exactly what C14 does. I haven"t Spiced the hi control, just tuned it by ear but it essentially boosts uppermost mids and highs and then allows to cut that back with the knob. I hope I'll find the time to Spice it and drop some frequency response plots in here. I used an inverting arrangement instead of a Sallen-Key here to allow the R15/R16 trick, which lets us have a switchable gate without a volume drop.

Cheers,
Andy
My dry, sweaty foot had become the source of one of the most disturbing cases of chemical-based crime within my home country.

A cider a day keeps the lobster away, bucko!

PRR

> the pot should be log

Reverse Audio.

(Assuming low frequency is counterclockwise.)
  • SUPPORTER

iainpunk

#9
Quote from: PRR on March 27, 2020, 05:30:43 PM
> the pot should be log

Reverse Audio.

(Assuming low frequency is counterclockwise.)

meh, its not important if you call it filter instead of tone, haha

im not really anywhere near sober, but i really wanted to reply to this thread because your project excites me, but just know that this was written with a foggy mind... take it with a grain of salt.


Quote from: Fancy Lime on March 27, 2020, 01:25:09 PM
Quote from: iainpunk on March 27, 2020, 12:36:35 PM
hey,

looks nice, i guess that this was the pedal you were talking about in the gorn overdrive thread?
Exactly.

Quote
i have 3 questions/suggestions

1] what is the frequency response of the first gain stage and the switches?
i gather that the active bridged "lower case pi" filter (from the Topic: What's this notch filter called that I just "invented"?) is there for the same reason as the gap in the HM2 eq response, to take out the mud? (also, i think i forgot to click post on my name suggestion(the capital pi looks like a normal pi filter, while this filter looks more like a lowercase pi, but bridged.))
The "tight" side has a corner frequency of 720Hz, the "fat" side is flat down to 21Hz. With both engaged you have about 13db more boost over 720
than at under 150Hz. Depending on the gain and tone settings this switch provides vastly different characters.

Yes, this thing is where the bootstrapped bridged pi filter comes from and yes, it is designed after the frequency response of the HM-2 and for exactly the same reasons.
nice
Quote
Quote
2] i really like the gate switch, but i'm not a fan of germaniums for this purpose, Si diodes are more to my taste when it comes to crossover distortion style gates, or even LED's. My current project is a boost/filter/gate where the crossover gate has a 3way (on off on) switch choosing between (on left)none, (off)leds or (on right)1n4148's for more flexible gating. the leds are fine for humbucker guitars, while the preceding circuit has way less gain than your goat has, i dont think the germanium transistors will have much effect on the noise floor, especially with the gain set higher.
I'm also not a big fan of germanium diodes as crossover clippers/gates, that's why I used BAT41's  ;). These are Schottky diodes and are on the high side of forward drops for Schottkys we tend to use in effect design. I liked them better than the 1N4148's I also tested but you may like the Si better. I found the BAT41's to be somehow rougher, earthier, less fizzy. It would also be no problem to include both options, which I will probably do in the next version. If you use Si diodes, you may want to lower R15 from 2k2 to 1k5 or even 1k to maintain a constant volume between the gated and non-gated positions.

i didn't realise that bat41's were Schottky's, for some reason i always assumed that they were Ge's. are those diodes a high enough threshold to cut down on most noise? i mean, with the gain pot on minimum, you are still boosting 45 times, when using any guitar, which might have at least 10mv of noise (which i think is a low ball) you would have 450mv of noise going in to the dirty gate, that is just above the diode threshold, with the gain at min and no long leads, vintage guitars or other noisy pedals (like my small stone, which is almost always on at the start of my pedal chain), i think it might be more logical to use different diode sets
i also forgot to mention in my first post that i always like dirty gates in between gain stages, but always before clipping, it seems to sound more natural that way, and you aren't amplifying the stray noise. you might want to experiment with that.

Quote
Quote
3] maybe a little bit of mid rage boost somewhere between 1 and 2 kHz, something like 6db (or 30db if you like adventure) might help the pedal cut through even more. especially because it will lose some of that in the broad scoop in the lower mids, even if the presence is not in the middle of  the scoop, this might add a bit of "presence" to your pedal. i suggest changing the treble control opamp to a sallen key resonant lpf with a q control from 0 to 6, ill crunch the numbers for you if you'd like
That is exactly what C14 does. I haven"t Spiced the hi control, just tuned it by ear but it essentially boosts uppermost mids and highs and then allows to cut that back with the knob. I hope I'll find the time to Spice it and drop some frequency response plots in here. I used an inverting arrangement instead of a Sallen-Key here to allow the R15/R16 trick, which lets us have a switchable gate without a volume drop.

Cheers,
Andy
[/quote]
yeah, i didn't consider c14 to have an effect on the high mids since i am not used to working with inverting opamps. but that would indeed negate the need for some high mid boost.

the only thing i would strongly suggest to experiment further with is the placement of the dirty gate.

to be honest, im not a fan of having one pedal that does a lot, i rather have a lot of pedals that do little. i love my grindy overdrive, not because it sound great on its own, but because it can change the character of the pedal or amp that comes after it. i also get a bit of a hard on for minimalistic circuit design. i certainly understand the appeal of this pedal and why people would want this pedal, but i'd like to do it with 3 pedals instead of one, because that gives a bit more flexibility.
i wish you the best of luck further perfecting this pedal.

iain

ps. never forget to learn from the mistakes of the future
friendly reminder: all holes are positive and have negative weight, despite not being there.

cheers

Fancy Lime

Actually, the placement of the gate is the result of quite a bit of experimentation. I find it sounds best near the end of the circuit and just before the treble control but integrated in such a way that the diodes form part of the resistance of the RC filter that cuts treble. This makes the gate vastly more efficient for hiss frequencies, which I usually find the most annoying.

The forward voltage was chosen by the following test: My desktop computer sits on top of my desk, about one and a half arms lengths away from the guitar in playing position. There is an LED lamp on the desk with the transformer and AC-DC converterin the bulb. There is a small 2x30W switch-mode amp on the desk just in front of the monitor that is about an arms length from the guitar. A Focusrite Scarlett 2i4 was used as a headphone amp for the test. All of the mentioned equipment was turned on. Pretty noisy environment. The guitar was a Yamaha Pacifica 311, which has a P90 in the neck and a humbucker in the bridge position, the latter can be split to single coil operation. The electronics compartment is not shielded. With the gate off there is considerable hum even with the humbucker. With the gate on, the humbucker is dead silent even at maximum gain but if you so much as think about breathing on the strings, the gate opens. To get a noticeably gated sound, you need to dial the gain back below 12 o'clock. With the P90 or split humbucker, you get a LOT of hum without the gate at max gain. With the gate in,it depends on the orientation of the guitar, whether you hear the hum come through or not. If you take back the gain a smidge, all is silent again.
Since i wasn't going for a "gated fuzz" type of effect, that was the best compromise for me.

One ultra flexible vs many specialized pedals: I actually agree. I also like to have pedals that do one thing and do that well. The Unix Philosophy works well in pedals building for me. The reason why I (over)stuffed this particular one with features is that I intend to develop a proper production PCB for it. I want this to be DIY friendly and therefore allow a range of different pedals to be build on the same PCB. The schematic I posted here is kind of the maximum version. In actual practice, some people may want to put the the mid control to a fixed value or have a switch to allow a pre-determined mid cut, instead of a pot. The tight/fat switch will not be relevant for everyone, so you can fix that. Or you may want more control here, in which case you can put a 5k pot in place of the switch. I fancy most people will either always have the gate in or always out, so they can omit that switch. Some may think gain pots are for sissies and nail gain to the maximum with fixed resistors. The clipping diode arrangement with seven diodes (there will be eight diode slots in the PCB) was chosen to provide a lot of flexibility for playing with different diode types, possibly series resistors and whatnot. By having one very flexible PCB, that can be made into many related but distinct things, it is possible to produce larger batches, which brings the price down. So you can order three PCBs for the price you would normally get one for and build three different specialized effects.Building several different effects on the same PCB is not exactly new. DOD, for example, made an art of multipurposing PCBs for *very* different effects, particularly with their 90's high gain distortion boxes. That is the type of thing I'm going for here. You can build the full smörgåsbord or pick and choose what you like. If I were to predict the most popular options, I would expect them to have four pots and no switches or three pots and one switch.

Cheers,
Andy
My dry, sweaty foot had become the source of one of the most disturbing cases of chemical-based crime within my home country.

A cider a day keeps the lobster away, bucko!

iainpunk

i believe that if you found the placement of the diodes in this circuit optimal, you should go with it, but i'd have to call into question your method of testing. i have experienced that most switching psu's, led lights and computers aren't as noisy as incandescent stage lights, big tube amps (which generally have suboptimal transformers, almost always too small even high end amps, basically clipping the incoming power (transformer saturation), thus creating a lot of harmonics on the grid which get antenna'd out by all power cables that are on the same group) and worst of them all, TL lights (even the small ones above emergency exits are noisy as heck). i have a lot of experience with noisy stages, i worked at one as a light guy (even guitars with shielded humbuckers i tested would give out up to 28mV of hiss and noise at our stage) .id advise you to try it out on an actual stage, instead of your workplace.

i really like the idea of a dod style "lets reuse our universal pcb for every pedal we put out" approach to pcb's, never thought of it that way, but maybe you could give the people the power to change the order of gain stages using a jumper-like system, to make it even more flexible for diy use.

i have dabbled in the concept of smaller "single stage" pcb's that can be soldered together in any order desirable. all that connects between the sub assemblies are the gnd, vcc and the signal, which is always supposed to be biassed to vcc/2 on the output side, so most stages won't have to create their own bias point. this allows for simple but infinite expansion and flexibility. there would be an input stage and an output stage which also incorporates foot switching and biasing so that everything could be as simple and fast as possible for the builder.

good luck building and experimenting,
iain
friendly reminder: all holes are positive and have negative weight, despite not being there.

cheers

antonis

#12
Quote from: Fancy Lime on March 27, 2020, 11:23:11 AM
I am afraid I do not follow the rest of your post. What voltage drop? What is "input bias voltage offset".

Nothing to worry about for FET input op-amps, but might be important for bipolar input ones..
('cause the second draw significant input bias current..)
NE5532 datasheet quotes 200nA (typ) and 800nA (max), so these currents would give a voltage drop across R5 of 100 and 400mV respectively (with input offset voltage of 4mV added, simply to make numbers rounded.. :icon_wink:)
This offset voltage is "reproduced" at the output of the op-amp (whatever its polarity) and "goes" to pin 2 as it is (due to unity DC gain of NFB loop - R7 & R8 are consider "open" due to C3 & C4 caps..)

The 5532 has n-p-n input transistors so the bias current flows into the input pins, hence the voltage at pin 3 will be negative with respect to VR at anything up to 400mV..
(it will actually dominate 4.5V down to 4.1V..) :icon_wink:

So you'll result in asymmetrical clipping where this was not your intention.. :icon_wink:

You could easily overcome it by making R25 value 27k to raise VR up to 4.9 so VR at pin 3 stays at 4.5V, for equal positive/negative headroom but you should also bias U1B & U2A with 470k resistors..
IMHO, lowering R5 value about ten times (down to 47k) should be OK, as far as you raise C2 value up to 1μF, due to Q1 low output impedance..
"I'm getting older while being taught all the time" Solon the Athenian..
"I don't mind  being taught all the time but I do mind a lot getting old" Antonis the Thessalonian..

lborl

This clearly needs to be called The Black Phillip.