Anyone know of an introduction to using CMOS for audio?

[Taylor]

I've found plenty of info on the basics of Rs, Cs, Ls, and opamps, but haven't found much as an introduction to the whys and hows of using things like CMOS inverters as gain stages. I'm not looking to clone anything, I just want to understand the concepts. Perhaps I just don't know the right words for Google searching, or maybe there's not much out there because this is a "misuse" of technology, which EE types probably don't have a use for.

[Processaurus]

I imagine this forum might be one of the few places that would have that kind of information, as like you said, it is a mis-use of the technology to use unbuffered inverters as clipping amplifiers.  Also, as amplifiers, most of the audio world would avoid them, as they are noisy, and the interesting clipping characteristics wouldn't be that handy, say, for hi fi.  Maybe just search here for 4049UBE or just 4049 and see what comes up, also the data sheet should have a schematic of the arrangement of all of the transistors, etc. inside the chip.

[R.G.]

The first CMOS inverters were simply a complementary MOS devices, one N and one P, with their gates commoned for the input and their drains commoned for the output. The P source went to V+, the N source to V-. In this configuration, they worked well enough for slow speed logic.

It also made for a decent enough amplifier if you managed to hold the gates near the center of the power supply and put a signal onto it. The actual ideal bias voltage varied from chip to chip, but if you used the output voltage to feed the input in negative feedback, it worked well enough. You could also split the feedback resistor, put a capacitor to ground in the middle and get all the normal gain. The normal gain was about 24-32db, depending on manufacturer, process, and supply voltage.

Using gates other than inverters was chancy, because the arrangement of inputs to get AND, OR, NAND, NOR, XOR, etc. did not necessarily leave you with an input that was good for analog signal input. But inverters worked. Usually, as long as they were unbuffered.

Later CMOS logic was buffered. That is, at least two more inverter stages were added after the input to the gate. This forced the final output to saturate HARD either high or low, even with lots of loading; it switched faster too. Buffering ran the gain up, but also made it chancy to get it to act linear.

Unbuffered CMOS inverters were not very linear to start with. They nearly always had some soft distortion in linear use. This got harder the closer you pushed it to the power supply. In addition, tuning both N and P channel devices on at the same time conducted current directly from the V+ to the V-, so the devices could get hot at high supply voltages.

Read "CMOS Linear Applications", www.fairchildsemi.com/an/AN/AN-88.pdf
I saw this first as a National Semi app note, but Fairchild seems to be hosting it now.

Enough?

[maarten]

The Dutch Elekt*r ran several (rather short) articles about audio applications for CMOS back in the seventies. I saw some of them in Dutch, but probably they were translated and used in the other language editions as well. Maybe you still can find them (or other magazines like PE) in a library?
Maarten

[Taylor]

Enough?

Yes! Thank you. This will give my pea-brain plenty to chew on.

There's a cool box made by Dwarfcraft called "The Great Destroyer" using inverters to get some very unusual sounds. The guy got dragged though the mud at "the other place" for his build quality and the fact that the circuit is obviously "stumbled-upon" rather than designed. The interesting thing is he uses the buffered 4069 chips, which as RG mentions is supposed to be bad for audio apps. I assume this has something to do with the pretty strange sounds of the effect, as well as the relatively thrown-together design.

I'm not really too into straight distortion sounds, but it seems like perhaps there are some stranger ways to abuse these chips - and I like that you get 6 inverters per chip.

[earthtonesaudio]

Buffered inverters are fine to use, but different.

A sorta Big Muff-sounding distortion:
Schem: http://www.aronnelson.com/gallery/main.php?g2_view=core.DownloadItem&g2_itemId=22795&g2_serialNumber=1
Clip: http://www.aronnelson.com/gallery/main.php/v/Schematics-etc/earthtonesgallery/4049Buffy.mp3.html

[Joe]

As gain stages, they are very similar to inverting opamp stages. CMOS chips are always noisy, which is a drawback.

[doitle]

Buffered inverters are fine to use, but different.

A sorta Big Muff-sounding distortion:
Schem: http://www.aronnelson.com/gallery/main.php?g2_view=core.DownloadItem&g2_itemId=22795&g2_serialNumber=1
Clip: http://www.aronnelson.com/gallery/main.php/v/Schematics-etc/earthtonesgallery/4049Buffy.mp3.html

Wow is that over the top!

[dschwartz]

ironically, the most silent distortion pedal i ever made is the hot harmonics....

[Boogdish]

Thanks a lot for that link, RG. 

[cpm]

Read "CMOS Linear Applications", www.fairchildsemi.com/an/AN/AN-88.pdf
I saw this first as a National Semi app note, but Fairchild seems to be hosting it now.
Enough?

i am printing this for commuting readings 
http://www.fairchildsemi.com/an/AN/AN-9010.pdf