CMOS pedal question

[pokus]

Hey there,
I've got a short question. In cmos based overdrive pedals with chips like the 4049 or 4069, the unused inputs of the single amplifiers are tied to V+ through a resistor, like in the tube sound fuzz or red llama. Now I want to switch between using two and three of them. On pole is switching the input of the 3rd one from my signal path when used, to V+ resistor when unused. The output is switched between the output cap and no connection.
Now my question is:
I would like to put a constant feedback resistor (about 100k) and a small pf cap across the input and output of that single amplifier to control the gain of the stage. So are there any issues when the input is switched to V+ while the output is left open but is also connected to V+ and the input through that feedback resistor? Or will the IC work properly? I really just have small knowledge of what is going on in a cmos chip.
Here's the tube sound fuzz schematic so you will get an overall idea of what I'm talking about.





Thanks for help!

[anotherjim]

You can do that. Actually you can switch the input to + supply or 0v with or without a resistor. Make sure the + connection is not a higher voltage than the chip Vdd pin.
If you have a feedback resistor you have a high current consumption in that inverter, but if you connect the input to either supply rail, it will put that stage into its lowest current consumption which would be a good thing.

[pokus]

Great! Thanks for clarifying! So I can save a pole on the switch by making that feedback resistor constant.

[anotherjim]

Well, the inverter is just what it's name implies but it's a property of this type made with a pair of MOSFETs that when you add a feedback resistor it will bias itself to a mid supply voltage and act as an amplifier. Because both MOSFETs are conducting in this state it takes more supply current. However, it's still an inverter in that state so if you connect the input to either supply rail the output should swing to the opposite rail, then only one MOSFET is turned on and the supply current taken becomes negligable.
There's always the risk of getting switching pop/click noises from what you propose. See how it goes and if you have trouble, show us the circuit you have and we'll see what can be done about it.

[dschwartz]

What I'd do instead is using a spdt to switch the main output from the previous inverter output, or the next inverter output. No need to "clamp" the unused inverter.
Less chance for pops...

[pokus]

Don't know if I get it right, but I want to switch lug3 of the volume pot from 2nd inverter output cap to 3rd inverter output cap. While 3rd inverter input is switched from V+ (doing nothing there) to 2nd inverter output. Just wanted to know if the resistor I put between in and out of 3rd inverter is causing any trouble when the input is tied to V+ and the output is not connected to anything but the (100k)resistor.
Switching noise is never nice, but as it shouldn't be switched through a footswitch when playing, I can accept a little noise.


Another thing bothers me a bit. I really like the sound of the Red Llama, especially in higher gain regions. But in really low/minimum gain territory there's this unpleasant crackling sound when it just distorts a little. It's not completely "in your face" present, but it kind of bothers me and is not what I expected from an overdrive. As in all the videos of it, that I watched, I can't hear a single pedal having those issue. It's hard to describe, but to me what it adds sounds quite similar to a slightly wrong biased jfet. Stands out mostly when multiple strings are ringing out.

Just wondered if someone can confirm/deny this??

I'm talking about the original Red Llama circuit, that still hasn't anything to do with the mods I explained above. As it works and I checked it multiple times my build should be fine. I even changed a few cap and resistor values, but still sounds unpleasant in low gain mode. Could it be the chip? I used the 4049UBE one.

[ThermionicScott]

Another thing bothers me a bit. I really like the sound of the Red Llama, especially in higher gain regions. But in really low/minimum gain territory there's this unpleasant crackling sound when it just distorts a little. It's not completely "in your face" present, but it kind of bothers me and is not what I expected from an overdrive. As in all the videos of it, that I watched, I can't hear a single pedal having those issue. It's hard to describe, but to me what it adds sounds quite similar to a slightly wrong biased jfet. Stands out mostly when multiple strings are ringing out.

Just wondered if someone can confirm/deny this??

I appreciate you asking this, because I'm on a similar path and noticing the same thing at the clipping point.  A tiny little fizz/crackle riding on top of the note as it fades out and becomes clean again, right?  In fact, now that I'm listening for it, I can hear my Ibanez TS-5 Tubescreamer, Boss OD-3, and modded Boss MetalZone all do it to a slight degree as well.  I feel like I can hear a bit of it in this Blackstone Mosfet Overdrive demo, at the 1:00 mark:

I'm left to wonder if something about my guitar (Rickenbacker 330 with Hi-Gains) or amps (Fender BF-style preamps) is accentuating that effect, and I should just learn to accept/ignore it, or use different gear... or maybe I just spent too long playing clean/acoustic, and forgot the little things that happen with distortion, who knows? 

[pokus]

I appreciate you asking this, because I'm on a similar path and noticing the same thing at the clipping point.  A tiny little fizz/crackle riding on top of the note as it fades out and becomes clean again, right?  In fact, now that I'm listening for it, I can hear my Ibanez TS-5 Tubescreamer, Boss OD-3, and modded Boss MetalZone all do it to a slight degree as well.  I feel like I can hear a bit of it in this Blackstone Mosfet Overdrive demo, at the 1:00 mark

I'm left to wonder if something about my guitar (Rickenbacker 330 with Hi-Gains) or amps (Fender BF-style preamps) is accentuating that effect, and I should just learn to accept/ignore it, or use different gear... or maybe I just spent too long playing clean/acoustic, and forgot the little things that happen with distortion, who knows? 

Yeah riding on top of the note is a very good term for it. I'm playing a Jazzmaster with also quite high output PUs into a Orange Thunderverb, so I guess it's not your gear causing that effect. Although we are both using relatively high output pickups 

I know this noise from some fuzz effects in low gain mode, but those are fuzzes and not overdrives like the Red Llama. Of course it's based on the tube sound FUZZ, so it has a few fuzz-like sounds swinging with it, but I've never heard that crackle in any demos of the Red Llama. This is what makes me think there's something wrong with the chip or else.

When I'm driving the clean channel with it, it occurs more than when I'm using the dirty channel with just a little gain. But I guess there's just more coloring going on there overlaying that effect. Nevertheless it's still a bit hearable there.

[Mark Hammer]

This is part of why I prefer to use an op-amp stage ahead of the 4049, rather than rely on the inverter exclusively.  It provides one more stage to tone-shape what is hitting the inverters, particularly with respect to shaving off any "frizz-potential" from the input signal.  For that matter, if a person wanted a smoother sound, you could even pop some diodes in the feedback loop of the op-amp to compress the dynamics of what's hitting the inverters a little.

Finally, remember that these are inverter stages, so adding a 3rd to the existing 2 will flip the phase of the output.  In a great many contexts, that won't really matter.  But if you ever try to run some things in parallel, it can yield cancellations.  Of course, most pedals will attempt to preserve phase thru to output, but not every single one does.

[pokus]

This is part of why I prefer to use an op-amp stage ahead of the 4049, rather than rely on the inverter exclusively.  It provides one more stage to tone-shape what is hitting the inverters, particularly with respect to shaving off any "frizz-potential" from the input signal.  For that matter, if a person wanted a smoother sound, you could even pop some diodes in the feedback loop of the op-amp to compress the dynamics of what's hitting the inverters a little.

I've read some of your thoughts about that. Forgot to mention that I had already put a micro amp booster circuit in front of the Red Llama. To me it added some kind of compression, but can sound really nice driving the inverters.
But when setting the boost stage to about the gain you will get from the RL alone this crackle sound is still there (even with much lower feedback resistor for inverter2). Of course the micro amp is just a very clean boost, doing almost no tone shaping for what is hitting the inverters. So maybe I will try to cut some more highs in the opamp stage, but as I've already tried to do that with a cap after the 100k input resistor to ground I don't know if that would help. The feedback diodes could be a nice addition.
Thanks for mentioning all that.

Finally, remember that these are inverter stages, so adding a 3rd to the existing 2 will flip the phase of the output.  In a great many contexts, that won't really matter.  But if you ever try to run some things in parallel, it can yield cancellations.  Of course, most pedals will attempt to preserve phase thru to output, but not every single one does.

Yes I kept that in mind, although for my use it should be fine.

[anotherjim]

I'm not sure if the RL can do really clean. The second inverter has a lot of gain, very dependent on what the impedance of its input capacitor happens to be and that gain will be higher for higher frequencies. Haven't tried this, but it might be worth making the 2 inverter sections identical and have a dual gang 1M pot controlling both gains and the second amp also having an input resistor, maybe not 100k -  perhaps only 10k to 47k.

[pokus]

I'm not sure if the RL can do really clean. The second inverter has a lot of gain, very dependent on what the impedance of its input capacitor happens to be and that gain will be higher for higher frequencies. Haven't tried this, but it might be worth making the 2 inverter sections identical and have a dual gang 1M pot controlling both gains and the second amp also having an input resistor, maybe not 100k -  perhaps only 10k to 47k.

Didn't really thought about the capacitor impedance here. Is it possible that the impedance of a cap, which has the same value but has a different insulating material, is that different it can make an audible difference in regards to gain? So that the coupling cap between the stages is what keeps my build from sounding like the original ones? The dual gang pot for both stages is a good idea, I will try that later. Not sure if I done some experimenting with the input resistor yet, but that goes to my list, too.


I tried the feedback diodes, not at the op amp but at inverter2 and things changed completely!!!!
The crackle now is gone, at least the breakup sounds way more pleasant in low gain mode.
I put out the feedback cap here, as it's not really doing anything to the sound except when going higher than 200pF. And put in some germanium diodes back to back. It somehow cuts a lot of bass (why?) and the higher strings now sound amazingly present without any crackle (similar to when you roll your guitars volume back on a fuzz face). Unfortunately in high gain territory it sounds a lot different and not that good to me, even when bringing a lot of bass through larger input caps back in. But the crackle keeps gone.
So I'm just about to guess it's pick your poison. 

[Mark Hammer]

I've never been able to get a dual-inverter drive to do anything "clean", though admittedly I haven't tried extra-special hard.  The best it will do is mild coloration. 

An op-amp front-end can smash the first inverter stage pretty hard; as hard as you like.  If one does choose to use that, and wants to be able to adjust how hard, I recommend using variable resistance in the feedback loop, rather than the ground leg, as the Micro-Amp does.  You want additional gain to be accompanied by a reduction in treble content.  Adjusting gain via the ground leg resistance ends up reducing bass content.

So, for instance, using the basic Micro-amp circuit, replace the 500k+2k7 on the ground leg with 1k.  Instead of the 56k+50pf in the feedback loop, use 2200pf and a 10k-fixed + 50k pot.  That will provide for an 11x to 61x gain from the front end. At 11x gain, treble rolloff begins around 7.2khz.  At max gain, treble rolloff begins around 1.2khz.

The rationale is that the harder you push the inverters, the more harmonic content will be produced (that is, after all, what "distortion" is).  By reducing the treble content of the input signal, higher gain accentuates lower-order harmonics.  A 1.2khz rolloff sounds severe, but trust me, there will be no shortage of "bite".

[anotherjim]

You can temper the effect of feedback diodes by fitting a resistor in series with both of them (any end). Or you can raise the threshold at which they act by adding more diodes in series. It's a common trick if you want germanium character with higher forward voltage that the added dides can be ordinary silicon. [DrStrangelove] Ze Germanium character will dominate ja! [/DrStrangelove].

Probably the loss of treble when pushed is that a conducting diode is nearly a short circuit, the input starts to see the output as a ground termination and the input capacitor becomes a low-pass fillter. Resistance in series with the diodes will reduce that effect.
When feedback diodes are used on opamps, it's with a non-inverting amp (tube screamer) and there is always x1 clean signal passed through indepenant of the diode action and you get distorted signal clamped by the diode forward voltage mixed with the clean. It can't work like that with an inverter.

[ThermionicScott]

I'm not sure if the RL can do really clean. The second inverter has a lot of gain, very dependent on what the impedance of its input capacitor happens to be and that gain will be higher for higher frequencies.

I've never been able to get a dual-inverter drive to do anything "clean", though admittedly I haven't tried extra-special hard.  The best it will do is mild coloration.

My TSF/RL breadboard build reminds me of my 5F2A Princeton clone amp with the negative feedback disconnected.  Clean at the lowest volumes, but you can really hear all the low-order harmonics build up and make the sound "richer"/"hairier"/"less distinct" (whatever terms you like) as you approach the clipping point.

[BJF]

Hi There,



Some perhaps interesting points about using CD40XX inverters as linear amplifiers.

In 1976 when CD4000 family was new there was a Swedish amplifier called U66 built with a preamp entirely with inverters used as amplifiers. Something the designers at that point noted was that by making a feedbackloop from output to input resulted in both input and output resting at half B+ and with an input resistor
and inverter could then be used as linear amplifier. However gain expression for this amplifier got to be complex because open loop gain was partly low and dependent on B+ voltage so that at 5V open loop gain was the highest but falling rapidly over 1KHz ( due to phaseshift a closed lop amplifier would then oscillate, while at 12V's open loop gain was less than half but at a 3dB bandwidth in excess of 10KHz, at 15V's bandwidth was even higher but raw gain dropped considerably. So for linear amplification with reasonable bandwidth powersupply voltage must lie between 8-12V due to that GBP would greatly depend on powersupply voltage.
In it's crudest form an inverter is one P channel and one N channel connected as totempole

In 1976 a CD4049UBE or CD4069UBE could be used as amplifier with just two resistors and it would automatically bias at half B+.Also NAND gates like 4001/4011 UBE could be used as amplifiers by conneting the two inputs. It is worth noting that modern CD4049UBE and CD4069UBE have a more eleborate input protection system and for this reason they will not bias at half potential but a resistor from input to ground can be used to bias output at half potential-the value of this resistor depends on feedback network and roughly around 0,5M-2M
The U 66 amplifier used 12V's to make highest bandwidth at highest headroom
B+ to hex inverter can be controlled with a resistor connected B+ and VDD and then VDD needs to be decoupled and with a large electrolytic since PSRR is not so great.

So due to that GBP is dependent on the B+, B+ can be used to set the desirable bandwidth and also feedbackfactor at higher frequencies: for this reason B+ will set the overtones produced under overload conditions

The buffered version of these have extra parallelled totempoles to increase fanout and due to additional phase shifts oscillation will occour under feedback conditions.

To consider is that total headroom is just shy of supply voltage compared to input peak to peak and can therefore be worthwhile to limit input pk-pk or strong peaks will upset linear and unlinear operation. A way around this is to limit pk-pk by means of clamping diodes within feedback loop of preceeding stage
Practical limit is below 2V pk-pk and antiparallell switching diodes will do just fine

Notes on feedbacknetworks:
While it is perfectly fine to hook up a tone control network around an inverter and limit is the low raw gain while +-18dB can be done; it is not desirable to place clamping diodes within feedbacknetwork again due to the low raw gain:closed loop gain has a complex expression due to the low GBP.

INternal isolation bewteen the inverters is not that great so a limit of three adjacent stages used per chip is practical - all six inverters could be used as long as fan out is within limits and risk of over hearing is minimized and this will depend on the stiffness of the powersupply.

Fan out is the current that an output can sink or supply


At your service
BJ
BJF Electronics
Sweden