Fun with CMOS - part 1 - The CMOS Raincoat

[Chuck D. Bones]

I've been messin' 'round with some low-noise CD4049s I picked up recently. First I tried the ROG 22/7. Not my jam, mostly because I don't dig the BMP tone. So what do you get when you take all of the diodes and some of the resistors out of a BMP? You get a Raincoat. Not nearly as thick and compressed as a BMP. The circuit below is similar to a Raincoat, but uses CMOS inverters instead of BJTs. I also changed the tone network, reduced the 1st stage gain a bit and added a FAT switch to cut the bass going into the distortion stages.

 

[Giorfida]

Very interesting! Kind of a take on the Red Llama/Anderton Tube Sound Fuzz/EH hot tubes… except with more cascading stages.

 

[Chuck D. Bones]

I'm pretty sure Anderton was the first to publish a pedal circuit using CMOS inverters. I had read about running CMOS in the linear mode in a trade journal back in late '70s and around the same time saw Anderton's Tube Sound Fuzz. The only tricky part is finding decent CMOS inverters. The stuff TI cranks out now is noisy. I have some NOS RCA and National Semiconductor parts that are dead quiet.

 

[cooder]

Colour me interested. So what's the spiel with noisy or not noisy CD4049s? Where's the good guys, where the shady ones to avoid?
I know, very technical way to ask in'it?

 

[Chuck D. Bones]

AFAIK, the only company producing new CD4000-series ICs is Texas Instruments. Their recipe is noisy when the CD4049 is operated in linear mode. If you can find old RCA or National Semi parts, they are quiet. Strangely enough, the CD4007s made by TI are quiet. That's why I prefer to design with those. The only advantage of the CD4049 (or CD4069) over the CD4007 is the CD4007 only contains three inverters. The 4049 & 4069 contain six. If and when we find a circuit we really like, the smart thing to do is design it with a CD4007. In the case of the circuit above, it would be easy enough to replace the first stage with a BJT or opamp. For the time being, this is just me playing around.

 

[JTEX]

Interesting. Back in the 90s I *almost* tried a distortion based around logic gates, but somehow managed to procrastinate for 30 years.

 

[BuddytheReow]

Great stuff here, Chuck. CMOS chips may be worth a Boneyard article to understand how the inverters work and why the topology is different than, say, an inverting opamp. I know I would be interested in reading about it.

 

[zgrav]

are you getting any dynamic range out of this configuration when you take out the diodes?

 

[Chuck D. Bones]

are you getting any dynamic range out of this configuration when you take out the diodes?

Yes, more dynamics than with the diodes. It still limits when driven hard, but that's to be expected.

 

[zgrav]

Yes, more dynamics than with the diodes. It still limits when driven hard, but that's to be expected.

seems like something fun to breadboard.

 

[Giorfida]

One thing I will say is that these ICs are VERY static sensitive. In fact a 4049 is the only IC in nearly 40 years of this that I’ve fried due to static discharge.

 

[Chuck D. Bones]

It's true. All MOS transistors and ICs are static sensitive. That's why we put 10V zeners in front of BS170's in pedal circuits. CD4000-series parts have some amount of built-in ESD protection, but it only goes so far. Store them in a conductive bag and don't handle them when the humidity is low.

 

[JTEX]

I used to have to wear antistatic straps or be fired. Both on wrists and shoes. Also, mandatory hand moisturizer. They'd also regularly wax the floor with slightly conductive wax. Static damage is real, and not always immediately noticeable.

 

[Robert]

I used to have to wear antistatic straps or be fired. Both on wrists and shoes. Also, mandatory hand moisturizer. They'd also regularly wax the floor with slightly conductive wax. Static damage is real, and not always immediately noticeable.

Oh yeah, been there.

The wrist straps and bench top ESD mat had alarms that would go off if either became disconnected or loose, and in extreme cases required a supervisor to turn off the alarm after inspecting.

We weren't allowed to have anything on the workbench except the piece of equipment we were working on... any paperwork or manuals needed to be on a separate bench.

The lab was also temperature and humidity controlled. Every once in a while the humidifiers would go offline and they'd come running in with mop buckets to drench the floor and bring the humidity back up.... and I tell ya, working on 480V 3-phase equipment while standing in a puddle made my nerves bad.

Anything that could have potentially been exposed to ESD was assumed to be defective.

The irony is that most of what we repaired / calibrated sat unused in a warehouse until the following year when it was lugged back in and calibrated all over again...

 

[JTEX]

The irony is that most of what we repaired / calibrated sat unused in a warehouse until the following year when it was lugged back in and calibrated all over again...

That's exactly the kind of soul-destroying stuff that makes one hate one's job. "Why bother? Rushing to finish this so it can go sit on a shelf and serve nobody."

 

[Chuck D. Bones]

Scanning Electron Microscope picture of ESD damage. All that splatter is from part of the metal layer that melted inside an IC from one tiny static spark.

 

[Robert]

I tested the ESD theory when I was a kid...

I'd walk around the house dragging my feet on the carpet to build up as much static as possible, then while holding an LED by one lead I'd touch the other to a doorknob. The LED would briefly blink, one last time...

 

[JTEX]

^ at least that poor LED died doing what it liked to do.

 

[airbud]

i finally did some intro research on these chips to fix my damn heterodyne receiver. great information, this is an entirely new frontier. thanks!

 

[cdwillis]

How do you calculate the gain in the stages that don't have an input resistor? That's something I never understood about the Red Llama with its similar setup. The UA1 should have 6.66~ gain (220k/33k=6.66~) I'm assuming, but what about UE1?