Flabulanche Questions

[HKimball]

So, I had some questions regarding the flabulanche and some modifications I wanted to make. I noticed that ROG made a "high gain fetzer valve" (the omega circuit) which I was interested in not so much for additional gain but rather for the variable input loading - additionally I wanted to move the gain control from between q1 and q2 to the source of q2 - the idea is to essentially convert the flabulanche to a cascaded boost festival.

I want to add input loading to the flabulanche to make the volume knob cleanup as righteous as possible. I know it's great already from the videos but I want to experiment because I'm weird like that. In no particular order:

1. If I use a mpf102 in q1 and feed the circuit 24v, can I substitute a fixed value resistor to ground for the gain pot, and add a potentiometer to q2 to adjust the gain via q2 bias (like the fatpants does)? Or would that be too much gain? I want light clipping at noon on the "gain" (slash fat) control.
2. If I add the omega input impedance control, will I need to change the rest of the cascaded fetzer part of the circuit to a higher resistance? (For example, using 50k trim pots on the drains). I ask because the omega article seems to imply that it's specifically sensitive to changes in input impedance - so I was curious is the high circuit impedance (pre-buffer) is necessary to get the most out of the "range" control.
3. if I wanted to implement a 'body' switch, should I just wire another cap in parallel with C2 (and a 10m resistor between the two) and use a spst switch? (again, like in the fatpants) Or should I add this to Q2?
4. If I use higher resistance drain trim pots, will I have to change the source resistors? Or will 470 work fine?

Thanks in advance for any help, I'm going to try and answer these questions myself through research in the meantime.

[midwayfair]

Hrm. There's a lot of things you've asked here that are looking to implement duplicate functions.

additionally I wanted to move the gain control from between q1 and q2 to the source of q2

The purpose of putting a gain control on the source is that it doesn't require components to decouple the gain control from the biased pins of the transistors, and because your gain control won't load the devices unduly. That's unnecessary in this design -- the gain control is already decoupled from Q1, and the loading is designed into the pedal. It also forms the gate biasing resistor of Q2. If you were to replace it, you'd have to add the biasing resistor, and then to make the source variable, you'd have to replace the source resistor with a pot of appropriate value, probably replace the trimpot with a larger value, and then add the 22uF onto the wiper. At that point, once you bias the FET, you'll have a variable gain that only attenuates about -6dB, because that's the gain boost you get when bypassing the source resistor on a FET that's been properly biased. The current gain control can go as low as -20dB. You'd be stuck on the right hand side of the dial, and I can tell you that I pretty much never set the gain above noon as it is.

I noticed that ROG made a "high gain fetzer valve" (the omega circuit) which I was interested in not so much for additional gain but rather for the variable input loading

Before I get to the input loading -- you can use the omega's large drain and source capacitors. The MOSFET in the omega is simply an output buffer with some specialized function (that isn't so different from having a FETzer valve biased a little high); the MOSFET doesn't increase the gain. The biasing affects how the FET clips; the 4u7 maxes out the gain.

the idea is to essentially convert the flabulanche to a cascaded boost festival.

That's what it is already! It's four stages that boost the signal by ~6dB each (the first is +12 for frequencies near 1KHz).

I want to add input loading to the flabulanche to make the volume knob cleanup as righteous as possible.

If you're looking for Fuzz Face type of clean up, you simply won't get that in this circuit. The Fuzz Face has a slew of things that make it behave like this, but the big ones are (a) negative feedback circuitry; (b) the use of a BJT, which can be biased all the way down to 0.5V, which means that it's far, far more sensitive to the input signal size. The Fuzz Face also treats the pickups like a current source instead of a voltage source, which is not something you can duplicate here.

2. If I add the omega input impedance control, will I need to change the rest of the cascaded fetzer part of the circuit to a higher resistance? (For example, using 50k trim pots on the drains). I ask because the omega article seems to imply that it's specifically sensitive to changes in input impedance - so I was curious is the high circuit impedance (pre-buffer) is necessary to get the most out of the "range" control.

The size of the drain and source resistors isn't the big reason the Omega boosts that much. In part it's that it's a J201 whose gain has been maxed out. A J201 is perfectly capable of hitting the power rails. Even without a source bypass capacitor, the J201 will boost by ~13dB. Add a source bypass cap at idea biasing and that's almost 20dB. Increase the size of the source resistor so that the cap has a bigger effect and you get the Omega's gain stage.

The Omega's input loading is there because it's how they chose to implement a Range control. It cuts bass as you turn it down because it forms a low-pass filter with the 3n3 cap. You could add all that stuff ahead of the 33K, but I'm not sure it's going to do much. In any case, if you choose to have the Omega's bass control, you probably want to make the source bypass cap on Q1 a larger value.

1. If I use a mpf102 in q1 and feed the circuit 24v, can I substitute a fixed value resistor to ground for the gain pot, and add a potentiometer to q2 to adjust the gain via q2 bias (like the fatpants does)? Or would that be too much gain? I want light clipping at noon on the "gain" (slash fat) control.

First you need to be aware that getting 24V means using a 12V supply. Don't try to give the LT1054 more than 15V. (You can give it 15V REGULATED.) Then go check the Fetzer valve calculator to see what sort of gain levels you're looking at with the 2N5457 on 18V, and compare that with the MPF102 on 24V. Then put in the values for the 2N5457, because you have two more of those in the circuit.

I think you asked before about the Fat control. The circuit must be designed around using that kind of gain control, because you need to be able to bias the FET so that the variable source resistance doesn't bias it into cutoff on the low end or oversaturation on the high end. I know you can get this range with a 2N5457, barely, and a J201. The Fatpants simply doesn't work with lower gain FETs, though.

Note that the Fat control from the Fatpants changes the biasing. The biasing is important in this circuit because it affects the way the signal is clipped. Altering one of the FETs bias as a gain control may produce unsatisfactory results. This isn't an issue in the Fatpants because the clipping there is incidental, not intentional.

3. if I wanted to implement a 'body' switch, should I just wire another cap in parallel with C2 (and a 10m resistor between the two) and use a spst switch? (again, like in the fatpants) Or should I add this to Q2?

If you want to audibly affect the bass, do it late in the circuit. The circuit is already full bass after Q1. Any cut to bass before the MOSFET stage will primarily affect distortion. The reason the bass cuts are there on and right after Q1 is because the distortion will sound very tubby without them. This is simply good gain staging. Also, your Range control is already going to be a bass control if you do it like what's in the Omega.

4. If I use higher resistance drain trim pots, will I have to change the source resistors? Or will 470 work fine?

For this question, I'm going to simply refer you to the FETzer valve article, which answers your question in depth.

[HKimball]

Hrm. There's a lot of things you've asked here that are looking to implement duplicate functions.

As I see now you're quite right... Unfortunately I get obsessed over OCD things that don't really have much of an effect on sound as it were.

The purpose of putting a gain control on the source is that it doesn't require components to decouple the gain control from the biased pins of the transistors, and because your gain control won't load the devices unduly. That's unnecessary in this design -- the gain control is already decoupled from Q1, and the loading is designed into the pedal. It also forms the gate biasing resistor of Q2. If you were to replace it, you'd have to add the biasing resistor, and then to make the source variable, you'd have to replace the source resistor with a pot of appropriate value, probably replace the trimpot with a larger value, and then add the 22uF onto the wiper.

I had planned on doing exactly this - I have attached a drawing because I feel like it would be easier to understand than my haphazard rants lol.

At that point, once you bias the FET, you'll have a variable gain that only attenuates about -6dB, because that's the gain boost you get when bypassing the source resistor on a FET that's been properly biased. The current gain control can go as low as -20dB. You'd be stuck on the right hand side of the dial, and I can tell you that I pretty much never set the gain above noon as it is.

I actually had planned on using it in conjunction with a few other modifications to reduce the gain - I would replace the gain control with a couple of resistors to approximate the gain control if it were set to noon (for example - I'll have to experiment with their values should I go this route) at the gate of q2.

Before I get to the input loading -- you can use the omega's large drain and source capacitors. The MOSFET in the omega is simply an output buffer with some specialized function (that isn't so different from having a FETzer valve biased a little high); the MOSFET doesn't increase the gain. The biasing affects how the FET clips; the 4u7 maxes out the gain.

After reading the section of the fetzer valve article about harmonic content (which is why I ordered a 340R resistor to use a mpf102 in q1) I'd prefer to use the lower-resistance values, but I wasn't clear as to whether the efficacy of the range control depended on the resistance of the circuit. My question was phrased poorly - what I meant was "do I need to use those values to get the most out of the range control?"

That's what it is already! It's four stages that boost the signal by ~6dB each (the first is +12 for frequencies near 1KHz).

This is true... I wanted something like a treble booster running into an echoplex. The treble booster to reduce bass, the echoplex to add texture. Essentially I wanted a pedal that acts like a sort of preamp rather than an OD pedal - cutting some bass, adding some texture to the mids, and pushing my tube amp.

If you're looking for Fuzz Face type of clean up, you simply won't get that in this circuit. The Fuzz Face has a slew of things that make it behave like this, but the big ones are (a) negative feedback circuitry; (b) the use of a BJT, which can be biased all the way down to 0.5V, which means that it's far, far more sensitive to the input signal size. The Fuzz Face also treats the pickups like a current source instead of a voltage source, which is not something you can duplicate here.

This brings up another question I had - re: the MOSFET section of the circuit - isn't there some negative feedback already in the design? I had seen on a fender amp schematic a presence control (which was a variable NFB loop with a capacitor to increase the amplified volume of treble frequencies relative to bass at the power amp stage) and was curious if I could do something similar in conjunction with the existing tone control. Essentially a dual ganged pot that rolled off highs before the power amp and increased the highs in the NFB loop at the same time. I know my language isn't precise here - my understanding is that NFB reduces overall distortion and makes the onset thereof more binary. And, it's been a while since I've seen the schematic, but IIRC the fender control in another part of the circuit would have rolled off bass rather than high frequencies? Something about the nature of NFB meant it had to work in reverse to get the desired effect. I know that's a lot of mumbo jumbo and I apologize for it.

The size of the drain and source resistors isn't the big reason the Omega boosts that much. In part it's that it's a J201 whose gain has been maxed out. A J201 is perfectly capable of hitting the power rails. Even without a source bypass capacitor, the J201 will boost by ~13dB. Add a source bypass cap at idea biasing and that's almost 20dB. Increase the size of the source resistor so that the cap has a bigger effect and you get the Omega's gain stage.

When reading the ROG article, I somehow got it into my head that the Omega used a 2n5457, and the claim on the site - 30-35db - made me curious as that was well in excess of any gain figures I've heard about the 2n5457 before (on 9v, I'm sure you could get there by upping the voltage.)

The Omega's input loading is there because it's how they chose to implement a Range control. It cuts bass as you turn it down because it forms a low-pass filter with the 3n3 cap. You could add all that stuff ahead of the 33K, but I'm not sure it's going to do much. In any case, if you choose to have the Omega's bass control, you probably want to make the source bypass cap on Q1 a larger value.

These are the droids I was looking for - I wanted to know if it would be a worthwhile endeavor to add the range control to affect the EQ of the pedal. If it's not going to do much then I figure I can just add the fat boostered/paul cochrane passive bass rolloff to an input buffer or something and get a similar effect.

First you need to be aware that getting 24V means using a 12V supply. Don't try to give the LT1054 more than 15V. (You can give it 15V REGULATED.) Then go check the Fetzer valve calculator to see what sort of gain levels you're looking at with the 2N5457 on 18V, and compare that with the MPF102 on 24V. Then put in the values for the 2N5457, because you have two more of those in the circuit.

This is what I do with my fatpants - I use an lt1054 with a 12v supply (and a 12.1v zener diode) because I like the sound more at 24v internally than at 18v internally. I've heard but cannot confirm that the mpf102 is near it's limit at 24v and maybe this makes it sound better to my ears? At any rate I know I want to do the same with this circuit - not because there's a valid scientific reason to do so but because it's something different and I'm weird like that.

I think you asked before about the Fat control. The circuit must be designed around using that kind of gain control, because you need to be able to bias the FET so that the variable source resistance doesn't bias it into cutoff on the low end or oversaturation on the high end. I know you can get this range with a 2N5457, barely, and a J201. The Fatpants simply doesn't work with lower gain FETs, though.

Note that the Fat control from the Fatpants changes the biasing. The biasing is important in this circuit because it affects the way the signal is clipped. Altering one of the FETs bias as a gain control may produce unsatisfactory results. This isn't an issue in the Fatpants because the clipping there is incidental, not intentional.

This is exactly what I wanted with a fat control in this application lol - having read aquatarius (I think that's the guy? He's the one you linked to in discussing the compressor circuit you use with the flabulanche) he mentions that he got FETs to clip more like a tube amp using lower source resistors than those proposed by ROG. However he doesn't provide any formulas for figuring out those values and I don't own an oscilloscope so I can't actually test the JFETs myself. The purpose of the "fat" control in my proposed circuit is not so much to change the gain level (I plan on using my guitar's volume knob for this) but to alter the timbre of the pedal and the clipped frequencies. Ideally I'd use it in conjunction with the range control (and the dual ganged treble/presence control I mentioned earlier) as a sort of tone control affecting the character of the pedal.

If you want to audibly affect the bass, do it late in the circuit. The circuit is already full bass after Q1. Any cut to bass before the MOSFET stage will primarily affect distortion. The reason the bass cuts are there on and right after Q1 is because the distortion will sound very tubby without them. This is simply good gain staging. Also, your Range control is already going to be a bass control if you do it like what's in the Omega.

Okay, so having different source bypass capacitors on q1 is unnecessary. That's excellent news. I'm not looking for a tubby distortion at all - quite the opposite - and not having to worry about a 'body' switch is great.

So, here's the circuit with my proposed modifications (excluding the charge pump):

edit: it's loading upside down, trying to fix now
double edit: will have to fix later today



If you can't read the notes I'd be more than glad to clarify anything and everything.

As always, thank you for your help - I can't tell you how much I appreciate it.

[TGP39]

Flipped....

[midwayfair]

I had planned on doing exactly this [...] I actually had planned on using it in conjunction with a few other modifications to reduce the gain - I would replace the gain control with a couple of resistors to approximate the gain control if it were set to noon

Yes, but WHY? Why would you replace the gain control with one that has only a tiny fraction of the same range? If you want to know how restrictive that will be on this design, grab another distortion pedal you've built and put the gain pot all the way up at 5:00. Now put the gain pot at 3:00. That's about how much range you'll have if you replace it with the source bias control and hardwire the gain pot spot to max like you've drawn. Is it JUST because you want to use the gain control you saw in the Fatpants? You can ADD that without removing the current (actually useful) gain control from the pedal. There's only three knobs on the stock design.

My question was phrased poorly - what I meant was "do I need to use those values to get the most out of the range control?"

You don't have to use those exact values but you do have to use their mathematical equivalents if you want the same range. You can calculate it here: http://sim.okawa-denshi.jp/en/CRlowkeisan.htm

This is true... I wanted something like a treble booster running into an echoplex. The treble booster to reduce bass, the echoplex to add texture. Essentially I wanted a pedal that acts like a sort of preamp rather than an OD pedal - cutting some bass, adding some texture to the mids, and pushing my tube amp.

I hope this doesn't sound rude, but I honestly have no idea why you'd want to use the topology of a pedal that's designed to create distortion if you don't want a distortion pedal. If this is what you wanted, you only need the first two stages. This is an awful lot of work when you only need 1/2 of the layout.

This brings up another question I had - re: the MOSFET section of the circuit - isn't there some negative feedback already in the design? I had seen on a fender amp schematic a presence control (which was a variable NFB loop with a capacitor to increase the amplified volume of treble frequencies relative to bass at the power amp stage) and was curious if I could do something similar in conjunction with the existing tone control. Essentially a dual ganged pot that rolled off highs before the power amp and increased the highs in the NFB loop at the same time. I know my language isn't precise here - my understanding is that NFB reduces overall distortion and makes the onset thereof more binary. And, it's been a while since I've seen the schematic, but IIRC the fender control in another part of the circuit would have rolled off bass rather than high frequencies? Something about the nature of NFB meant it had to work in reverse to get the desired effect. I know that's a lot of mumbo jumbo and I apologize for it.

The 1M bias resistor is negative feedback, but only incidentally. It's a biasing resistor.

In Fender amps, the presence control impedes a cap to ground hanging off the cathode of a tube that's also connected to the negative feedback. The presence pot forms the second leg of a voltage divider with the negative feedback resistor. When you lower the impedance to ground, the tube's gain for high frequencies increases and high end content is removed from the negative feedback. (When you remove frequencies from negative feedback, you get more of them in the final sound.)

There are lots of places with negative feedback in the circuit already. The source bypass cap on Q1 is a form of negative feedback. The compressor circuit is a form of negative feedback. The Q5/Q6/10uF/33K series is also negative feedback, and it's also the primary source of clipping in the pedal. The negative feedback in that arrangements ONLY appears when the signal is large enough to clip Q5 and Q6, which are diodes (the threshold is about 2-3V). Then it's limited by the 33K, which appears in parallel with the 1M for large signals.

But here again, I'm not sure what the point of adding a "presence" control like you describe would in a pedal that already has a treble control near its output. The treble control here is basically in the same place an amp would have its treble control.

[HKimball]

Yes, but WHY? Why would you replace the gain control with one that has only a tiny fraction of the same range? If you want to know how restrictive that will be on this design, grab another distortion pedal you've built and put the gain pot all the way up at 5:00. Now put the gain pot at 3:00. That's about how much range you'll have if you replace it with the source bias control and hardwire the gain pot spot to max like you've drawn. Is it JUST because you want to use the gain control you saw in the Fatpants? You can ADD that without removing the current (actually useful) gain control from the pedal. There's only three knobs on the stock design.

Yeah... I dunno. I think I'll follow that and add it without removing the gain control. I want a pedal that will distort, sure, but I want to be able to change the character of that distortion. If I'm not mistaken there are some settings that sound "raunchy" if you bias the transistor to a much lower voltage, like 8v instead of 11v.

Think of the "gain" control as a "character" or "response" control more than a gain control. My understanding is/was that the bias of the transistor affects not just the gain but the "deformation of the sinusoidal waveform upon excitation" (the Aquataur article about his umble modifications talk about how he does this).

I wanted to take out the existing gain control so I could limit the design to four knobs - "range," "timbre," "treble," and "volume." I almost never fiddle with the gain settings on my OD pedals once I have them set to where I like them - I much prefer to use my volume knob. I had planned to use the fixed resistors to set the maximum gain level to where I wanted it and then free up that real estate so I can have more knobs that affect the character and response of the pedal.

You don't have to use those exact values but you do have to use their mathematical equivalents if you want the same range. You can calculate it here: http://sim.okawa-denshi.jp/en/CRlowkeisan.htm

Thank you for the link - I'll be sure to plug in the requisite numbers and figure everything out.

I hope this doesn't sound rude, but I honestly have no idea why you'd want to use the topology of a pedal that's designed to create distortion if you don't want a distortion pedal. If this is what you wanted, you only need the first two stages. This is an awful lot of work when you only need 1/2 of the layout.

It's not rude at all - I am asking quite a broad array of questions concerning things I don't fully understand... Of course I'm going to sound idiotic. That's the price I pay for being too lazy to teach myself more of this stuff. On the contrary you've been very patient despite my persistent stubborn disposition lol.

The 1M bias resistor is negative feedback, but only incidentally. It's a biasing resistor.

In Fender amps, the presence control impedes a cap to ground hanging off the cathode of a tube that's also connected to the negative feedback. The presence pot forms the second leg of a voltage divider with the negative feedback resistor. When you lower the impedance to ground, the tube's gain for high frequencies increases and high end content is removed from the negative feedback. (When you remove frequencies from negative feedback, you get more of them in the final sound.)

There are lots of places with negative feedback in the circuit already. The source bypass cap on Q1 is a form of negative feedback. The compressor circuit is a form of negative feedback. The Q5/Q6/10uF/33K series is also negative feedback, and it's also the primary source of clipping in the pedal. The negative feedback in that arrangements ONLY appears when the signal is large enough to clip Q5 and Q6, which are diodes (the threshold is about 2-3V). Then it's limited by the 33K, which appears in parallel with the 1M for large signals.

But here again, I'm not sure what the point of adding a "presence" control like you describe would in a pedal that already has a treble control near its output. The treble control here is basically in the same place an amp would have its treble control.

The "presence" control would function differently in kind than the existing treble control - my understanding is that by introducing the frequency-specific negative feedback you change the rate and levels at which different frequencies clip. Because the audible effect of clipping occurs in the treble frequencies, I wanted the treble frequencies themselves to be less clipped - rather than simply rolled off late in the circuit. So yes, there would be more treble frequencies as a result, but the intended effect would be to decrease the actual clipping of those frequencies.

Additionally, in conjunction with the compression circuit, I wanted the effect to be clearer-but-still-compressed high frequencies and more distorted midrange/low frequencies.

I think what I'm going to do is this:

- Leave all the PCB mount controls untouched
- Change q1 to a MPF102
- Add those micro-shaft potentiometers for any additional controls should I decide to go that route
- Practice more instead of thinking of quirky ways to make circuits less useful lol