Filter Design for Specific Tones

psb962

Active member
I've built a few pedals, built a bunch a pedal models in LTspice, and have some electronics knowledge from university (ok, in 1982..).

I've been looking specifically at the design of the filters in some of the popular pedals to try and figure out what the designer wanted. There's been a lot of discussion on clipping on various boards, but not much on filters, and to me, filters are what gives a pedal its mojo.

One major problem with this is when a pedal is described by the designer as 'transparent' - which to me means no tone coloration ie flat EQ across audio band - but the frequency response of some of these 'transparent' pedals in LTspice shows that the EQ is nowhere near flat across the audio band. However, I don't want to get sucked into a discussion on transparency as I prefer to look for distinctive, colored tones in any pedals I might design as it's good to stomp on a pedal and get a whole new sound.

One design issue that the EQ of some pedals changes a lot not only with 'tone' controls but often with 'gain' controls too - this is easy to understand if you are conversant with active filter design using op amps - changing the value of any resistors in the negative feedback loop will change the EQ if there are capacitors in the loop as well. For example, the last pedal I built was a Clark Gainster clone, and adjusting either the gain or the presence control has a marked effect on EQ as both are variable resistors in the loop (as shown below). Are there any successful designs out there where the designer has managed to separate gain and EQ?

Another issue is that clipping adds higher harmonics that weren't there before, so only filters after the clip can affect the new harmonic content.

Gainster Spice.png

I wonder if the more experienced members of the forum have heard of anything like a pedal design cookbook which has a variety of modules (ideally in LTspice) with guidance on how to adjust the parameters to achieve the end goal?

One other key piece of info would be a comprehensive understanding on what EQ gives what tone - the Duncan Tonestack tool is a start, but that doesn't cover pedals. I'm sure someone has done this before so if anyone knows of a resource please post.

Lots of topics there. I'm hoping this thread might become a useful resource for those seeking a better understanding of the design of the filters in guitar pedals so we know what to do to get them to sound just the way we want.
 
About your points in the post:
- Transparent doesn't mean it IS flat, but it feels/sounds flat. a completely flat overdrive/distortion would be shrill and fizzy as hell on the top end, and excessive low end gives you mud and in the worst case oscillation. It is of course often just a marketing buzzword or completely misunderstood (such as when people call the Klon transparent) and of course is open to personal interpretation.
- Gain controls change the tone with those "smoothing" caps in the feedback loop. This is often done to counteract harshness that occurs from clipping, to keep the overall tone more even and reduce noise. Again, simulations may look like it's muting the top end quite a bit, but in reality it's way less noticable if the chosen values are good.
- The presence knob in the Gainster sets the cutoff frequency and is MEANT to change EQ. Usually designers use a fixed value here or a combination of more filters to set the frequency response. Unless it's something like the DOD250 or MXR Distortion+ which have their gain control here with the feedback loop resistor being a fixed value. Which also means that low gain tones will be darker, while higher gain is darker. Having the main gain control be the feedback loop resistor is a way to counteract that.
- Post-clipping filters are basically almost every pedal tonestack ever, since as you say, you want to filter also AFTER the clipping, especially in hard clipping distortions.
- There's Wampler's book that has been suggested above could be a place to start. Otherwise I'd say just look at a bunch of schematics, build a bunch of pedals and find out what type of EQ/filters you like specifically.
- All of the tonestacks in the TSC can be used for guitar pedals. The Big Muff tone stack specifically is a guitar pedal one and the amp-style and James tonestacks are also used a lot in pedals. Otherwise as I said, check out pedal schematics, Electrosmash.com is a good resource for explanations of some classic designs as well.
 
Really good info, thanks. I will check out the Wampler book(s).

Are any of this community into LTspice modelling at all? I'm finding it a great tool for exploring circuits before breadboarding them.
 
Really good info, thanks. I will check out the Wampler book(s).

Are any of this community into LTspice modelling at all? I'm finding it a great tool for exploring circuits before breadboarding them.
I've been using LTspice for many years and rarely find a need to breadboard anything - I just go straight to PCBs. But I typically only build clean-ish widgets for bass or acoustic instruments, and have a lot of well vetted building blocks, so tailoring circuit interactivity and EQ curves is often my primary focus.

OTOH, one of the first PPCB boards I bought was waiting on documentation at that time, so I did a board trace and put it into LTspice just to see what mods I might want to try first. Might finally get around to doing that build this morning, actually.

Heve you ever tried sending WAV files through your models? I used to do that a lot, and hopefully when my new computer arrives next week it'll go a bit more quickly, as renders can be painfully slow on my current ancient one, which originally was a Win7 box.

Oh, and finally: I really like resonant low pass filter EQs a la Alembic and Wal, and have a long running thread on Talkbass with plenty of Spice graphs and participation from a few commercial makers: https://www.talkbass.com/threads/the-passinwind-open-source-preamp.1259692/
 
Last edited:
Thinking more about this topic, I recently acquired a Fargen Mini Plex amp, and one of the key features of this amazing little tone machine is the "decade switch" which changes the voice of the amp from 60s (JTM) to 70s (JMP) or 80s. (JCM). The decade switch is a very useful feature and the voices are certainly very different. Lately I learned that the circuitry behind the decade switch is very simple - just a change of input cap - so that is making me think hard about how same approach to filtering might help create three pedals in one for pedals where that would be appropriate.

Have any of you tried breadboarding or building pedals with this sort of feature?
 
... this sort of feature?

I built a Rat with 3 different bass-responses:
More Bass​
Stock​
LOTS More Bass​

You can add a cap in parallel to the input cap and blend it in with a pot instead of a switch.


The output cap is also fertile ground for experimentation; just be aware that messing with the input/output caps might affect impedance and take that into account.


Also, some of the circuits already have that "feature" —
Dragon's Breath Boost, for instance (Naga Viper which itself is basically a Treble-booster).

You'll find that sort of "feature" thingamabobbee in boutique YAFFs, YATS, JAMs etc.
 
Thinking more about this topic, I recently acquired a Fargen Mini Plex amp, and one of the key features of this amazing little tone machine is the "decade switch" which changes the voice of the amp from 60s (JTM) to 70s (JMP) or 80s. (JCM). The decade switch is a very useful feature and the voices are certainly very different. Lately I learned that the circuitry behind the decade switch is very simple - just a change of input cap - so that is making me think hard about how same approach to filtering might help create three pedals in one for pedals where that would be appropriate.

Have any of you tried breadboarding or building pedals with this sort of feature?
I've played with this idea with a bassified BMP build but haven't found values I loved yet. Was taking the approach of blending with a pot but tabled out of frustration of tayda solid core wire continually breaking.
I've since purchased more durable mil spec solid core and need to breakout the desoldering iron and get back to it.

IIRC, I was using a 5k pot but the problem was I couldn't get enough sweep. Also, readings with my DMM looked odd. As in. The capacitance measurement would increase over time with no variables being changed. Wondering if it is parasidic capacitance on the stripboard.
Anyway, it's a great idea. I just can't tell you a proven way to do it with a pot. I may just change it to a rotary switch and call it a day.
 
I've played with this idea with a bassified BMP build but haven't found values I loved yet. Was taking the approach of blending with a pot but tabled out of frustration of tayda solid core wire continually breaking.
I've since purchased more durable mil spec solid core and need to breakout the desoldering iron and get back to it.

IIRC, I was using a 5k pot but the problem was I couldn't get enough sweep. Also, readings with my DMM looked odd. As in. The capacitance measurement would increase over time with no variables being changed. Wondering if it is parasidic capacitance on the stripboard.
Anyway, it's a great idea. I just can't tell you a proven way to do it with a pot. I may just change it to a rotary switch and call it a day.
One design that works well with a 3PDT switch is the Radian Ge Treble Booster by AionFX. With the switch open you get the 4n7 and with it closed you get the 4n7 in parallel with either the 6n8 or the 22n for treble, mid, or full range boost.

Radian.png

In other news I've just breadboarded a Muffin Fuzz using the Violet Rams Head values and will be playing with the caps in the feedback paths to see how they impact the sound.
 
One design that works well with a 3PDT switch is the Radian Ge Treble Booster by AionFX. With the switch open you get the 4n7 and with it closed you get the 4n7 in parallel with either the 6n8 or the 22n for treble, mid, or full range boost.

View attachment 59899

In other news I've just breadboarded a Muffin Fuzz using the Violet Rams Head values and will be playing with the caps in the feedback paths to see how they impact the sound.
I am, or was at least, trying to get the pot to sweep from slightly less bass than the normal muff to nearly no bass roll off. I doubled (220) a lot of the other filtering caps to allow more bass throughout but wanted the option to tune the input.
Hence the desire to filter a little more upfront.
Also modified the tone stack to sweep from bassy to almost flat to mid lift and added 3 clipping options to each stage, si. GE(2 diodes back to back to raise Vf) and either IR or UV LEDs, can't remember. I think IR. Those will be basically no clipping except on hot signals from an active bass, in theory. But no clipping in the first stage for a bass on a BMP is said to be desirable.
 
I am, or was at least, trying to get the pot to sweep from slightly less bass than the normal muff to nearly no bass roll off. I doubled (220) a lot of the other filtering caps to allow more bass throughout but wanted the option to tune the input.
Hence the desire to filter a little more upfront.
Also modified the tone stack to sweep from bassy to almost flat to mid lift and added 3 clipping options to each stage, si. GE(2 diodes back to back to raise Vf) and either IR or UV LEDs, can't remember. I think IR. Those will be basically no clipping except on hot signals from an active bass, in theory. But no clipping in the first stage for a bass on a BMP is said to be desirable.
The Duncan tonestack app is a good way to experiment with the BMP tonestack. I'm having a problem with my breadboarded BMP at the moment as if I take the output off the collector of Q3 it sounds great, but if i take it off the cap after Q4 it sounds thinner. There must be something wrong with my tonestack stage but I cant find it. Does yours sound better at Q3 than at output?

However, it did teach me how to use a cap and an extra patch cable to probe around a breadboarded circuit to see what's up.
 
The Duncan tonestack app is a good way to experiment with the BMP tonestack. I'm having a problem with my breadboarded BMP at the moment as if I take the output off the collector of Q3 it sounds great, but if i take it off the cap after Q4 it sounds thinner. There must be something wrong with my tonestack stage but I cant find it. Does yours sound better at Q3 than at output?

However, it did teach me how to use a cap and an extra patch cable to probe around a breadboarded circuit to see what's up.
The BMP filter stage is passive.
So it causes signal losses.
It's actually, if you didn't know, a blended passive HPF and LPF that overlap. Hence the mid scoop.
There are tons bypass mods, and also PCBs.
I like madbean's mudbunny, personally.
If you're going down the rabbit hole a big muffs and you haven't read over his bill docs, I'd recommend.
The schematic shows the tone bypass really well. Though there's still some high and low filtering happening to keep the circuit from being too boomy or hissy/fizzy.


Edit: also good reference for transistor voltages
 
Back
Top