4016 switches as amplifiers report

[gez]

Out of interest I wired up a single stage of a 4016 switch as a linear amplifier today.  Lo and behold it worked!  

At low gain it sounded extremely transparent and very quiet.  Thinking it wouldn’t have that high a Transconductance, I decoupled the AC feedback and ran one stage flat out.  How wrong I was, this put me straight into the Metal zone!  The distortion isn’t as smooth as CMOS inverters, but it’s still better than diode clippers IMO.

I’m still toying with all this, but I just thought I’d pass this on as it’s a useful thing to know about - four high gain, decent sounding MOSFETS all in one chip (anyone getting any ideas yet?!).

[Tim Escobedo]

Wow. I'd heard of switches used as digital inverters, but I'm surprised they could be biased into a somewhat linear range. I thought they were designed specifically NOT to do that!

I'm gonna have to give this one a try.

[Peter Snowberg]

Very cool Gez!

I love misapplication of CMOS chips.

Take care,
-Peter

[gez]

I should have added some detail on how to do this (always something I forget!)  

I nominated the y connection as a ‘drain’, the x as a ‘source’ and connected it to ground, wired up a 10k drain resistor to V+ then used a feedback resistor from drain to gate (enable connection).  From a 9V supply the ‘drain’ biased at just under 4V.

This idea was an afterthought, I had the chip on the board and was just stripping it down when this occured to me - hence the simple bias, quick and easy to see if it worked.  I'm sure things could be improved upon.  

How you wire the amp up to set gain is entirely up to the individual.  I decoupled the AC feedback, but had to make the decoupling cap smaller than I normally would to avoid oscillation (pretty tetchy in this respect).

What amazed me was the sustain from just one stage, though it has to be run flat out to do this.

[gez]

PS.  I split the drain resistance up, along the lines of a fuzz face, when I ran the thing at full pelt.  I also stuck a pot at the output to control volume.

[David]

Come on, give it up!  Let's see the schematic!    

[bioroids]

Hi Gez

I was having some fun, using this 4016 as a heavy distortion, with the control pins as inputs and the switch pins as source & drain like you said.
I thought I had invented something and then bam! I see it on the forum. Well, the Nobel prize will have to wait ... :evil:

I should have posted it before to get some credit! but didnt have the time

Anyway, I had a hard time trying to get rid of the oscilations, so I was planing to build it properly (instead I have it on the breadboard). I'm gonna try your suggestions for it.

Luck

Miguel

[gez]

Come on, give it up!  Let's see the schematic!    

I have no problem with sharing a schematic, though I only spent 5mins on this circuit prior to packing up today so I’d like to spend a little more time on it before posting anything.

I’ve only just learnt how to upload stuff to the space provided by my server.  Do I have to do something for others to view my files?  If so what? (I’m with AOL).  I’m pretty clueless when it comes to computers…

I’m more interested in using this idea for things other than fuzz.  It will be interesting to hear how a twin-T sounds.

[gez]

Anyway, I had a hard time trying to get rid of the oscilations, so I was planing to build it properly (instead I have it on the breadboard). I'm gonna try your suggestions for it.

I was wondering if you’d tried this because you mentioned in one of your posts that there’s an area where channel resistance is variable.  That was one of the things that got my mind ticking.  Having glimpsed at the schematic I figured it could be done then just forgot all about it until today, when I was packing up.

At high gain oscillation is a problem.  I wired up two 1M resistors from ‘drain’ to gate and decoupled the junction with a 22n cap.  As things stand it oscillates with both pickups on but not when only one is selected.  Turning the volume pot of my guitar down a touch stops it.  Reducing the decoupling cap further would probably solve the problem outright (I’ve found this often works with inverters, though you loose a bit of low end).

PS.  We’ll claim the prize together!

[smoguzbenjamin]

There's always such a thing as a disclaimer "This circuit hasn't been tested and if it doesn't work, you've got work to do!"

You get the idea

[gez]

There's always such a thing as a disclaimer "This circuit hasn't been tested and if it doesn't work, you've got work to do!"

You get the idea

Fair enough, here you go!

http://hometown.aol.co.uk/Gezpaton/Switch+Fuzzer.gif

I have no idea if this can be viewed by others.  Perhaps someone can tell me what to do if you can't see it?!

I just used any old connections.  This circuit is just to give you the general idea.

Edit: I'm off to bed now, so if there are any problems viewing this...tough! :twisted:

[puretube]

the circuit is plain to see, at least from germany, here...
(I`m off to bed now, so didn`t breadboard it...)

[Marcos - Munky]

Cool, Gez. How many stages did you used in your built?

[gez]

Cool, Gez. How many stages did you used in your built?

Only the one so far, as I said in my previous posts this is just a sketch.  I'd be surprised if you didn't get oscillation with more than one stage, you'd have to reduce gain to do that.

The fuzz sound is a lot of fun, but I'm more interested in running these at lower gain (normal set-up of feedback resistor and input resistor).  The sound was very transparent when I did this so it might be nice for a twin-T wah...or two...or three...or four...

[gez]

PS.  If you get oscillation try reducing the 22n to 10n or lower (will kill the low end a bit though) OR stick a resistor in series with the input to form a divider with the 1M resistor seen at the input.  

Another approach would be to reduce the value of the other 1M resistor - the one which connects to the ‘drain’ - to deliberately load the output (though you'll probably have to increase the value of the decoupling cap too).  

Reducing the composite drain resistance would work too, but might affect bias.

[bioroids]

I see you used a different setup, maybe it works better. On my setup, I got rid of the oscillations by conecting a huge (1000uf) cap between +9v and gnd, though I still had a low passed noise left. Should breadboard this again to check.

Luck

Miguel

[smoguzbenjamin]

A 1 mF cap!!!!! :shock: That's an amazing amount of charge!!!

[Taylor]

I found this searching around for fun things to do with a 4016. I was really intrigued, but the link to the schematic was dead (5 years of link rot). I asked Gez and he sent me the file:



And here's a mirror at archive.org in case photobucket dies or the folders get restructured:

http://www.archive.org/details/ScuzzBox

A very interesting idea! I'm really intrigued by abusing CMOS stuff.

Edit: just breadboarded it, pretty nice. It's actually a fairly conventional sounding heavy gain distortion. Sounds somewhat like the Tube Sound Fuzz and its ripoffs. I'm surprised no boutique guys have come out with anything using this technique.

[StephenGiles]

It would be nice to force a 4016 gate or two into a VCF configuration - I wonder!

[Cliff Schecht]

At low gain it sounded extremely transparent and very quiet.  Thinking it wouldn’t have that high a Transconductance, I decoupled the AC feedback and ran one stage flat out.  How wrong I was, this put me straight into the Metal zone!  The distortion isn’t as smooth as CMOS inverters, but it’s still better than diode clippers IMO.

Think of this like an op amp vs. comparator. One can be used as the other but it isn't necessarily in your best interest to do so because each was designed with a specific purpose. A comparator is like a very high gain op amp which helps the device to switch faster. The same ideas are analogous to what you're doing here. This is why switching transistors are typically designed to have a very high transconductance, they will switch faster at the cost of lower linearity (not of interest in a switch) and higher power dissipation at elevated switching speeds. For your purposes, the high transconductance makes the switches saturate easily and is probably giving you a lot of third-order distortion considering the typical CMOS switch topology is a push-pull pair on the output.