Do the NAND/NOR gate Boogey!

[edvard]

I was perusing some of my old Forest M. Mims notebooks and stumbled upon the fact that any NAND/NOR gate can be wired as an inverter.
If so, then any of the popular gate IC's (4001, 4011, etc.) can be used to make OD/Distortion pedals (a la 3-legged Dog) with up to 4 amp sections.
Vis:


Hooray, those 'useless' chips now have stompbox possibilities!

Then I started thinking...
Since the Dr. Boogey is essentially 4 gain stages with some tone shaping and gain controls mixed in, would it be possible to build one using a N* gate IC?

Gimme your thoughts while I start breadboarding...

[Brymus]

Yes,I have been toying with a CMOS distorter ala 3LD, HH,RedLama,Tube Sound Fuzz for quite awhile.
The ROG Double D is an excellent sounding pedal IMO,in the same high gain arena as the Dr Boogie.
Also IMO CMOS high gain pedals require far fewer parts to achieve the almost/same sound.
Just be sure to properly de-couple the PS to avoid oscillations/squeals.

PS another idea I had (which has already been done before of course) is to use the un-used sections of the IC for actively EQing the pedal.
there isnt a whole lot of info on this,but I am pretty sure it is doable.

[edvard]

What intrigues me about this idea is that with the 3-Legged Dog and other 4049-based circuits, you go throwing away at least half the chip because 2-3 stages usually gives you more than enough gain, and there's six of them on there.
Most Dual-Input N*-Gate chips would have only 4 stages to deal with and the pinouts are identical so you could compare different ones for sound/current draw/price, etc.
Also, I'm becoming not a big fan of fiddling with trim pots to get good sounds out of FET circuits...

Using the extra inverter stages of a 4049 circuit for active EQ sounds very interesting, but I haven't run across any circuits.
Do you know of any schematics I could reference?
I remember trying a similar idea once with an op-amp based circuit by simply running the normally grounded part of the tone stack to the inverting input instead.
It worked well, although I didn't experiment enough with it to see if it overcame the volume drop common with many tone control configurations.

Still breadboarding...

[edvard]

OK, I found your CMOSerizer circuit and the ROG Mr. EQ, so now I'm off the breadboard and into theory-land 

"... It's full of stars..."

[earthtonesaudio]

The unbuffered 4049 and 4069 have only one complementary pair per gate.  The NAND and NOR schematics I've seen generally have several FETs in series, which would mean more gain when wired as a linear inverting amplifier.  I would expect this would make them behave more like the buffered versions of the '49 and '69... which is to say more harsh and hard clipping.  (Although I kind of like that sound.)

[anchovie]

The unbuffered 4049 and 4069 have only one complementary pair per gate.  The NAND and NOR schematics I've seen generally have several FETs in series, which would mean more gain when wired as a linear inverting amplifier.  I would expect this would make them behave more like the buffered versions of the '49 and '69... which is to say more harsh and hard clipping.  (Although I kind of like that sound.)

I've interpreted it that the multiple FETs are in parallel, but total numbers of P and N are different. In an old forum post by RG that I don't seem to be able to find today, he said that this makes the amplification non-linear and therefore when overdriven the clipping is asymmetrical. If you have a look at the graphs on the datasheets for CD4001 and CD4011 you'll find one for the transfer function of the gate.

Just like with 4049 and 4069, it's important to use gate chips with a UB suffix. These aren't as common as buffered NAND and NOR gates.

[Brymus]

Sorry no advice,just wanted to clarify the CMOSerizer is WGTP's not mine 
But glad your having a go at this what you uncover should help me along as well.

[edvard]

Sorry no advice,just wanted to clarify the CMOSerizer is WGTP's not mine 
But glad your having a go at this what you uncover should help me along as well.

*facepalm*
Thanks for the clarification, and props to WGTP...
Don't know how much help I'll be, I have too many questions myself.

The unbuffered 4049 and 4069 have only one complementary pair per gate.  The NAND and NOR schematics I've seen generally have several FETs in series, which would mean more gain when wired as a linear inverting amplifier.  I would expect this would make them behave more like the buffered versions of the '49 and '69... which is to say more harsh and hard clipping.  (Although I kind of like that sound.)

I've interpreted it that the multiple FETs are in parallel, but total numbers of P and N are different. In an old forum post by RG that I don't seem to be able to find today, he said that this makes the amplification non-linear and therefore when overdriven the clipping is asymmetrical. If you have a look at the graphs on the datasheets for CD4001 and CD4011 you'll find one for the transfer function of the gate.

Just like with 4049 and 4069, it's important to use gate chips with a UB suffix. These aren't as common as buffered NAND and NOR gates.

Hmmm... According to this datasheet there's equal P's and N's going on, and only two of each at that, with none going in series.
This one, however shows 4 of each with what appears to be some in series (can't quite follow the signal path), is that what you're talking about?
Is that why the 'UB' may sound better then the 'BC'?
Since the pinouts appear to be the same for all Quad 2-Input N* Gate chips, this would be a great experimentation circuit; just pop in a different chip for different sounds (HOW different would be the ultimate question...).

Time to get breadboarding again... 

[earthtonesaudio]

Yep, the two datasheets you linked show the unbuffered (UB) and buffered (BC) respectively.  With the unbuffered version at least, tying one input high will essentially turn it into an inverter just like a 4049.  Tying the inputs together, however, will give you some sort of asymmetrical response, because the two P FETs are in parallel and the two N FETs are in series, so the output will source 4x the current that it can sink.  Could be interesting sounding for distortion.  Might make the bias point a bit weird.

Alternatively, you could exploit this weirdness and use one input plus feedback resistor in the "normal" way, but then add something to the other input, such as another feedback resistor or even an additional input source.  (Modulation input, anyone?)

[caspercody]

Not sure if this is what you may be looking for, but on the Emma Reezafratzitz (6) stages of a 4069. One stage looks to be used for a tone control.

[edvard]

Yep, the two datasheets you linked show the unbuffered (UB) and buffered (BC) respectively.  With the unbuffered version at least, tying one input high will essentially turn it into an inverter just like a 4049.  Tying the inputs together, however, will give you some sort of asymmetrical response, because the two P FETs are in parallel and the two N FETs are in series, so the output will source 4x the current that it can sink.  Could be interesting sounding for distortion.  Might make the bias point a bit weird.

Hmm... Very interesting

Alternatively, you could exploit this weirdness and use one input plus feedback resistor in the "normal" way, but then add something to the other input, such as another feedback resistor or even an additional input source.  (Modulation input, anyone?)

If it'll behave in an analog sort of manner, yeah that'd be cool.
You could also possibly do gating, compression, etc. but that's a little over my head ATM.

Not sure if this is what you may be looking for, but on the Emma Reezafratzitz (6) stages of a 4069. One stage looks to be used for a tone control.

Sounds interesting, I looked for a schematic but they are nowhere to be found, can you point me to one?

[caspercody]

It is on the other website (FSB). Or send me a PM with your email address, and I will send you the schematic.

[edvard]

Found it...
Rather like I expected, just a cap and pot over the feedback path of one of the stages.

Breadboard report:
I found that to get any sound out requires both gates to be tied together as input, as shown in the bottom circuit clip above.
Started out with 0.1 μf cap -> 100k resistor -> 1M resistor -> 0.22 μf cap.
Nice clean sounds, we're good so far.
Hooked up another stage with the same setup but used a 0.47 μf on the tail.
Nice and loud, a bit of breakup, piercing in the high notes.
Substitued 10M resistors in place of the 1M's just to pull some crunch out.
The resulting distortion was very hard-edged but not saturated and I immediately noticed a rise in the hiss level.
Ignored it for now but made a mental note to deal with it later.

So, I think I'll stick with the 100k/1M resistor pair for now and see how it fares after stringing it up with a Dr. Boogey topology.

[edvard]

Holy smokes!!
This thing draws ~18ma!  
Is that typical for CMOS designs?
I remember Craig Andertons Tube Sound Fuzz was (ideally) supposed to draw ~7ma, but I'm still a bit surprised.
Please tell me I'm reading my DMM wrong!

Never mind.
I had an output grounded.  
Now it reads ~8 ma.
Not the greatest, but better than 18, yeesh!

[edvard]

OK, I've got a little farther on this thing.
Earthtonesaudio, you were correct.
When using a 4001, the bias point is around 2 volts plus change, while a 4011 gets you somewhere around 6-7ish.
I had to use a big (1k) resistor and 470uf cap on the wall-power to get it to stop motorboating, so it's actually biasing down in the lower sixes.
Haven't tried to get tricky with using the other input, but I will as soon as I get a circuit finalized and posted.
Stay tuned...

P.S. So far, it sounds awesome...

[caspercody]

Do you have a schematic, or sound clips to see/hear?

Thanks
Rob

[edvard]

Currently working on both.
Hopefully in the next couple days, I've got a full work schedule

[WGTP]

I'm really happy with the cmos distortion.  Using the gyrators from Mr. EQ is working well, although some of the values have been changed in the first inverter for more gain.    Thru studying STM's early BMP simulation and other ROG CMOS circuits, I realized you can add a resistor/pot to ground at the cap that goes to ground and vary the level and Q (I think) of the boost/inverter. 

With the CMOS Obsidian, I did what your suggesting and used 3 inverters to replace the 3 discrete Mosfet stages.  Should work for other stuff, but the gain isn't as high, it still sounds good.  Using large resistors creates gain and noise as does multiple stages.  I suggest you keep the resistors between stages as small as possible if you want maximum gain, or don't us any.  Note the open loop of the second stage of the CMOSerizer.  I guess it is biased by the inverters on each side of it since there are no caps between stages (there is not a large resistor, does that reduce noise?).  Caps are not needed between each stage in some instances.  The voltage of the 4049 effects the "character" of the distortion and I usually run mine on low 9v batteries around 7-8v.  Cool stuff.   

[caspercody]

I have been curious to know how these cmos work. Is there a article/book to read about how to use cmos for distortion applications? Or do you just try different components?

[earthtonesaudio]

I think this forum is the best resource out there.  Just search for 4049 or 4069 based distortions.  Members WGTP, and STM come to my mind because they've done quite a bit with them.  Also there are some projects on runoffgroove.com worth looking into.

A CMOS inverter is a P and N MOSFET with drains and gates coupled, and sources going to positive and negative supply pins, respectively.  Negative feedback together with the symmetry of the two FETs causes the output to bias roughly near the midpoint of the supply, and from there you can basically use any circuit that would work with an inverting op-amp, with the exceptions that there is no non-inverting input and the gain is much lower than a real op-amp.  Alternatively, you can think of the CMOS inverter as a single transistor with no need of drain/collector or source/emitter resistors.  This might be a more accurate model because the input and output impedance characteristics are closer to a single transistor stage.