Red Llama slightly dulled top end - any remedies?

[alex_spaceman]

Put together a Red Llama clone and really love the tone and feel overall but it does seem to slightly dull out the very top end compared to the clean sound. I play a Jazzmaster into the TB channel of an AC30 so yeah, I like my tone bright.

What should I be looking at to get some of that top end sparkle back? The 51pf (47pf in my case because standard values) and 100pf caps seem like good candidates, but could changing the value of the volume pot have any influence on this as well? Schematic calls for 10k, which is pretty low and I went 20k in my build as that's what I had lying around.

Any input would be much appreciated! This pedal seems to play really nice with the AC30 and would really love to tweak it to my taste as I would no doubt end up getting a lot of use out of it!

[Mark Hammer]

Drop the feedback cap in the 2nd stage from 100pf down to something from 39-51pf.  It won't make a world of difference, but it should help.

[amptramp]

The combination of the 1 megohm feedback resistor and the 100 pF capacitor gives a rolloff of 1592 Hz.  This will make things seem dull.  Drop the capacitor down to 33 pF to move the rolloff up to 4825 Hz.  You really don't need anything above 5000 Hz out of a guitar, so if it works with this value without any tendency to oscillate, leave it there.  The first stage has a rolloff that varies with the gain setting and the use of a pot that has a high capacitance to ground means that the 51 pF should remain.  The feedback lead capacitor is intended to counteract the capacitance to ground at the inverting input and the pot has a high capacitance to ground.

All of this is based on the beavisaudio schematic.

[alex_spaceman]

Thank you both for your input and yes, apologies, thesis the schematic! http://beavisaudio.com/schematics/Way-Huge-Red-Llama-Schematic.htm

Will drop the 100pF and go from there. Out of interest, how would I go about calculating the rolloff frequency in this context? Not that I want to get mathematical about the value choice (tuning by ear between 47p and 33p should more than suffice to all intents and purposes) but would certainly help me understanding how this works and hopefully avoid having to ask such questions/be able to offer help in return in the future

[antonis]

Out of interest, how would I go about calculating the rolloff frequency in this context?

R5/C5 (as well as VR1/C3) form a High Pass filter which here behaves as Low Pass one due to its placement inside the negative feedback loop..

In case you've puzzled, read "Simplified Inverting Amplifier Filter Circuit" below:
https://www.electronics-tutorials.ws/filter/filter_5.html

[tzdvl]

The Red LLama circuit was based on the original Craig Anderton Tube Sound Fuzz:



The changes of the coupling caps reduce the low-frequency response in the Red Llama, but should not affect the treble response.
The increase in the feedback caps (C1 and C2 in the schematic) in the Llama WILL noticeably cut the highs.
As stated in the post above, the reduction in highs is dependent on the gain (treble reduced when gain increases).

I experimented with the TSF/RL circuit a couple of years ago.
I breadboarded the basic circuit, switched the components, and checked the changes in response on my scope, as well as the audible effects.
I recorded the effect of each cap value at gains of 1 and 10.

Basically, C2 had no noticeable effect on treble response (at any gain) until C2 was 150pf or more.
The effect of C2 = 150pf at gain = 10 was inaudible (-1dB @ 16KHz).
The effect of 330pf was BARELY audible at gain = 10.

Here are my measurements:



As you can see, C1 is responsible for any reduction in treble response.

So, if you want to maximize treble output, I would use the values recommended by Craig Anterton for the Tube Sound Fuzz circuit, 10pf for both C1 and C2.

[tzdvl]

The Tube Sound Fuzz circuit was published in Craig Anderton's book Electronic Projects For Musicians.

When I was building my Tube Sound Fuzz (and doing the measurements in my post above)
I was surprised to find that C1 had a large effect on treble response, and C2 had essentially no effect.

In the original TSF project, on page 171, under the "Modifications" section, Mr. Anderton writes:

[For] "De-emphasized high-frequency response: Increase the value of C2. A value of 50pF starts rolling off the high frequencies at approximately 4000Hz."

Based on my measurements, I believe this is a misprint?
The roll-off seems to be an effect of increasing C1, not C2.

[alex_spaceman]

The Red LLama circuit was based on the original Craig Anderton Tube Sound Fuzz:



The changes of the coupling caps reduce the low-frequency response in the Red Llama, but should not affect the treble response.
The increase in the feedback caps (C1 and C2 in the schematic) in the Llama WILL noticeably cut the highs.
As stated in the post above, the reduction in highs is dependent on the gain (treble reduced when gain increases).

I experimented with the TSF/RL circuit a couple of years ago.
I breadboarded the basic circuit, switched the components, and checked the changes in response on my scope, as well as the audible effects.
I recorded the effect of each cap value at gains of 1 and 10.

Basically, C2 had no noticeable effect on treble response (at any gain) until C2 was 150pf or more.
The effect of C2 = 150pf at gain = 10 was inaudible (-1dB @ 16KHz).
The effect of 330pf was BARELY audible at gain = 10.

Here are my measurements:



As you can see, C1 is responsible for any reduction in treble response.

So, if you want to maximize treble output, I would use the values recommended by Craig Anterton for the Tube Sound Fuzz circuit, 10pf for both C1 and C2.

Well, that was hugely helpful, thank you ever so much for chiming in (pun very much intended)! As I am currently short of low value ceramics I decided to simply swap the 47p and 10p and that's already done a great job at livening up the top end! It still seems to dull out a touch (that is, in the context of a very jangly rig) at max gain, but that probably wouldn't be my application for it anyway, so think it's now time to think about boxing this one up

[Mark Hammer]

The unit puts out a VERY big signal; enough that sticking some passive filtering between the second stage and volume pot won't result in enough of a drop in level that you'd treat it as a disadvantage.

Instead of monkeying around with the feedback caps, consider inserting a 2-pole lowpass filter between the 10uf cap and volume pot.  As well, consider using a DPDT-centre-off toggle to get three different preset rolloffs.  The advantage of a 2-pole LPF is that the rolloff is steeper.  Certainly not any sort of brick wall, but noticeable.

How to implement?

Put a pair of 1k resistors in series, just before the 10k volume pot.  After each one run a pair of caps (in series) to ground.  The common of each half of the toggle goes to the junction of each pair of caps.  One outside lug goes to the cap/resistor junction, and the other outside lug goes to ground.  When the switch is in the middle position, the two caps are in series, such that their combined capacitance is reduced.  The perk is that they always have a path to ground so there is no pop/click from using the toggle.

Given 1k series resistors, I'll suggest .15uf (150nf) and .056uf (56nf).  Their series capacitance works out to .041uf.  Using 1k, that works out to 2-pole rolloffs starting at around 1060hz, 2840hz, and 3900hz.  No frizz, but nice bite and midrange honk.  Naturally, you can use other resistor and capacitor values.  I'm just illustrating one possibility.  You'll still have gobs of volume, but three distinguishable voicings.

[antonis]

How to implement?

Hey, Mark..

Why not to implement unused inverters as unity-gain buffers for each RC active low pass filter..??
(two hex inverters strapped fior unity gain in series should work just fine..)

[Mark Hammer]

Sure.  I just figure that the fewer semiconductors in/along the way, the less noise.