A question about 2nd stage hi frequency roll-off in 4049 design

[AM]

Hi,
I'm working on a distortion circuit design and trying to adjust the first and second stage hi frequency roll offs. My design is based the 4049 chip and it takes various elements from Mark Hammer's 49er, Anderton's fuzz, Llama etc. I decided to keep the second stage the same as the Llama, put a buffer at the front and tweak the first stage. SWTC and some other things will follow the second stage.
In the Llama design he first stage hi frequency roll off is depending on the gain pot setting. The second stage is fixed at around 1.6 khz.
I am trying to figure out the ideal resistor-capacitor combination for the first stage now.
My question is:
If I set the gain at minimum I have a high frequency roll off at the first stage around  32khz. Since the second stage has a fixed roll off at 1.6 khz doesn't that mean that everything above that is getting cut anyway? In other words, no matter where I set the gain pot nothing above 1.6khz is gonna make it to the output, right? Am I missing something here?
Thanks
Schemo link:
http://gaussmarkov.net/layouts/redllama/redllama-schem.png

[tackleberry]

Its not cutting the treble just allowing some to not get amplified so much to tame some of the high freq clipping you can get. This is from ROG site and its pedal using the 4049.

The Double D is born! We were able to develop a two inverter channel, "Jiggle" with a lot of classic amp-like sounds at various settings of the gain knob. The other channel, "Bounce" uses three inverters to crank the gain up higher. Caps were used over the inverters to mellow the clipping out. We've implemented Mark Hammer's suggestion of rolling the treble back more in earlier clipping stages and less in subsequent ones. This helped bring out the amp-like sounds and removed the harsh, "fizzy" high end of the bare circuit. There is certainly no lack of highs at the output. The typical "piercing" highs of classic amps dimed are still present. Dynamic response is really good and grit can be controlled from the guitar volume knob.

[WGTP]

The roll off at 1.6MHZ is only a first order filter that is at -3 db at 1.6MHZ and continues to roll off at 6db/octave after that, so it isn't a very fast roll off.  When you have two 1.6MHZ filters cascaded, the -3 point is at 64% (1.024MHZ) and rolls off at 12db/octave.  You probably need to experiment with the 2 different roll off points.  Using sockets is very handy. 

The 2nd stage will be the only one effecting the low gain settings because 32MHZ is too high to have much effect.  At mid gain the 1st stage will start having some impact and at max gain, both will be in full effect.  The cmos distortions generate a lot of high frequencies at max distortion, so that combo may work out perfect as is. 

Since the cmos distortions have a reputation for getting farty, you may want to try socketing c1 as well to control the low frequencies.  Larger caps make it more of a fuzz and smaller ones more of a distortion IMHO.   

[AM]

The roll off at 1.6MHZ is only a first order filter that is at -3 db at 1.6MHZ.....
.............The 2nd stage will be the only one effecting the low gain settings because 32MHZ is too high to have much effect.

Hi, thanks for the answer. One correction in what you wrote. It's Khz not Mhz the measures I get. Do your observations still apply in this case?
Right now I have a 220pF at first stage, together with a 22k resistor and a 500k pot. At max gain I get 1386 hz there. Is that too low?

Since the cmos distortions have a reputation for getting farty, you may want to try socketing c1 as well to control the low frequencies.  Larger caps make it more of a fuzz and smaller ones more of a distortion IMHO.   

After building similar circuits a gazillion times (well... , not that many but 4-5 for sure) I have found the values that work the best there for the sound I want which is somewhere in between. I think of it as NOT an overwhelming fuzz. It's definitely still a bit too bottom heavy to be called a normal distortion.

[Mark Hammer]

I'll do the red face for my forum colleague who typed mhz when they obviously meant khz  .

The thing to keep in mind is that "distortion" means you are adding harmonic content, well in excess of what was there originally.  So, often what can seem like a very low lowpass corner frequency still lets through a large amount of treble content.  If there had been no treble boost whatsoever, you would be correct in assuming that a 1khz rolloff would yield fairly dull tone.  But in this instance, it is simply compensating and attempting to restore the tonal balance.

[AM]

I'll do the red face for my forum colleague who typed mhz when they obviously meant khz  .

Hey Mark, no problem at all. My intension wasn't to point out the mistake as something embarrassing at all. I was trying to figure out if the observations are still correct.

The thing to keep in mind is that "distortion" means you are adding harmonic content, well in excess of what was there originally.  So, often what can seem like a very low lowpass corner frequency still lets through a large amount of treble content....

Since you wrote, I'll take the opportunity to ask your opinion about this. We had an exchange of posts a long time ago when I asked a couple of questions regarding one of my first Llama builds. I remember you said then (among a lot of other helpful information)  that if I wanted to keep some bite in the circuit I shouldn't go above 100 pF in any of the two stages.
Well, with this latest build I went twice as much. As I said earlier, with my current values on this new circuit I get a roll off at 1386 Hz at full gain and 32.8 KHz with gain at zero.
I thought about putting a 100pF cap there. that would give me 3 Khz at maximum gain but then I would get 72.3 khz when gain is at minimum and i don't know if that's too much. Can that be an issue?
Bear in mind that I have replaced the 100k fixed resistor with a 22k one and the gain pot with a 500k one.
Second stage is left as it is in the original schematic. would swapping the 220pF cap for a 100pF one make a noticeable difference?

[Mark Hammer]

The potential to have awkward filter rolloffs is precisely why I personally prefer to smack the front end of an invertor with an op-amp gain stage, leaving the filtering and gain of the invertor stages fixed.  Far simpler to just stick a non-inverting op-amp with a gain of 20-50 ahead of the first invertor stage and plunk a level pot between them to adjust how hard the invertor is driven.  You will also note that the extra op-amp stage provides yet one more point of intervention for productive filtering to take place, at both the low and high end.

One of the perennial difficulties in any distortion/overdrive circuit is that wide-range gain/drive adjustments are often accompanied by inconvenient bandwidth changes.  In your case, achieving the "cleanest" sound introduces far too much treble and the potential for a hissy output.  The op-amp-plus-invertors arrangement, with an attenuator between them, allows for the bandwidth to remain fixed over a broad range of drive adjustments, something which I personally find more workable.  The 4049 excels at producing warm grind, anyway, so leaving the top-end rolloff fixed where you like it is a good idea.

[WGTP]

Well... I'll use the convenient excuse that I'm still recovering from my daughters wedding.  Yeah, that's it...  As always, Mark has an excellent point.  Some folks like the effect of cutting the treble as the gain increases, but it may be difficult to balance it where you have optimum treble at all gain settings.  That is why tone controls are nice.  Same with the bass. 

Not to be a show off, but here are a variety of inverter circuit snippets I assembled for you cmos pleasure. 

http://www.aronnelson.com/gallery/main.php/v/WGTP/Red+Rooster.GIF.html?g2_imageViewsIndex=1

[Mark Hammer]

Congrats on the marriage of your daughter.   Always a good excuse!

The problem with the gain adjusted by the invertor is that the range of resistance-change required to produce the desired overdrive tone, AND an option for clean, results in too much alteration of treble.  If you put the cap value higher, you lose bite at high gain, and if you put the cap value too low, you get hiss and radio reception when gain is too low.  The op-amp front end sidesteps all of that.

[AM]

Congratulations !!! And best wishes!

I see the point in having the op-amp front end. I have to admit, until now I thought taming high frequencies while boosting gain was the best way. Well, it turns out it's not...
Some observations in my results so far:
1. The full distortion of the circuit is really fat sounding. There is still enough hi frequency content but it doesn't have the trebly attitude a lot of people have criticized it for.
2. It retains the fuzzy flavor of the original circuit but it's somehow less loose and I think a bit smoother. I'm not saying that it turned into some polite sounding pedal. It's just a bit smoother.
3. I don't hear more hiss than my previous builds. I don't have the other pedals anymore. They all went to different people as presents but I remember their sonic attributes more or less. There is definitely some hiss but it's more or less the same amount with all previous pedals.
4. There is no major coloring in the main tone. I mean it's a fuzz circuit, there is of course some coloring but nothing huge. It just adds more harmonic content on top of the main tone.

Mark you are making me wanting to build something like you 49er now! The thing is I was hoping to keep it kind of simple. A bit like building an IC based fuzzface.  A few components and just two pairs of volume / gain knobs on a switch so I can stomp between two different gain settings while keeping the output level the same. A buffer at the front and maybe a tone control (although I don't think it needs it anymore). All housed in a single box with two footswitches.
Your idea is really interesting and I can clearly see its advantages in the way you describe it. Most of the distortion circuits have been designed following the "taming high frequencies while boosting gain" principle right? I mean tubescreamers, bluesbreakers, fuzzes, etc.  Am I right? I get my ideas by just looking at what is out there and tweaking. I'm not as knowledgeable as you so there are a lot of limitations which is one of the reasons i try to keep it simple.


I guess it would be useful to record a short clip playing a couple of chords and some single notes so we can judge the results so far.
I will need to find a place to upload them though.

[Ben N]

No DIYer can refuse any request on his daughter's wedding day.

[AM]

Ok, Latest update. I ended up replacing the 500k log gain pot with a 1Mohm linear one and played around with C1 and C4 a bit more. These changes made a difference in the way the gain increases and also, I let a bit more bass content hitting the first stage now but scooped some out right after it. With gain at zero it still acts as a clean boost but the full gain comes much sooner now. At about half way on the gain knob it's almost maxed out. The interesting thing now is after that it just boosts some highs and sustains a bit more. So although in theory the more I dial in the gain the more I also attenuate highs with my cap/res choise, in practice I boost highs while keeping gain, volume and noise levels practically the same. I guess there must be so much high frequency content at this point that even with a roll off point of 708 Hz at full tilt, all I do is just restoring it to a level that is not over the top.
I think I like the results. Out of all the circuits of that kind I built this is the most distortion-like while still keeping the fuzzy flavor I like.
I play with humbuckers so there was always some issues there before.
I can also just set the volume knob at the level I want and then just increase distortion, treble content and sustain using my Guitar's volume or the pedal's gain control without boosting volume too much. I really like this last one. It saves from needing two channels or two boxes for different gain levels.
I noticed that with my amps it sounds just right (and unrighteous  )  and maybe a bit on the bright side unless I roll my guitar's tone back. Straight to the mixing desk through a cab sim is a bit dull. Maybe the circuit would benefit from a switch to go between 47 pF and 220 pF at the first stage.
Thanks everyone for helping.