CMOS Mosfet Distortion Sim

Started by WGTP, March 10, 2010, 05:50:21 PM

Previous topic - Next topic

WGTP

OK, so I'm spanking Joe Davison's "Obsidian", 3 mosfet distortion with a BMP tonestack built on a small Radio Shack breadboard sitting on top of an Epiphone Valve Jr. head and sealed 12" truck woofer cabinet and thinking how great it sounds.  The down side is that there are parts everywhere on the breadboard.  3 stages of mosfet boosters consisting of 1 mosfet, 2 caps, 4 resistors, plus 2 resistors between stages, and then the BMP tonestack.  That's 23 parts not counting the tonestack.  The Obsidian is definitely worth the complexity, it's an earlier version of the BOR.

Then I looked at the breadboard that contains CMOS designs also sitting on the amp.  A 16 pin 4049 or 14 pin 74HCU04 (that is designed for audio and seems to have a crisper sound) that contain 6 stages of complementary mosfets is the heart of this breadboard. 
So I start thinking about simulating/emulating/approximating the Obsidian with 3 stages of the CMOS chip and 1/3 as many parts. 

First, the complementary mosfets stages in the CMOS chip are not the same as a single CS mosfet stage.  I don't really understand all the technical differences.  I think the CMOS stages have less input impedance, gain and bandwidth, so the design will need to address that as much as possible.  In both cases, noise can be a problem, so the resistors needed to be as small as possible while still keeping the gain up. 

3 identical stages consisting of a single 470k resistor in the feedback loop were used.  No caps between stages are needed.  The input cap controls the bass continent.  A 100k resistor was added at the input to raise the impedance (OPTIONAL).  10k resistors between stages are used as in the original, but may be omitted for additional gain.  The tonestack was modded slightly to bring out more treble content. 

47 to 100pf caps can be used in parallel with the 470k resistors for treble/edge reduction if needed.  LED's can also be added in the FBL for a different sound.
The first 470k resistor can be replaced with a 1meg pot for a gain control, or a pot to ground in the circuit path can be used as in the original, but this adds more parts.

The voltage greatly affects the "character" of the unit.  The 470 ohm resister could be a 2K trim pot or external control to use with the Gain and Output controls.

Then I see the Blackstone Appliance Overdrive/Distortion and BOR schematic on another forum and that gets me thinking about a variation with 4 stages... and additional tone control circuitry...hummm  :icon_cool:

http://www.aronnelson.com/gallery/main.php/v/WGTP/CMOS+Obsidian.jpg.html?g2_imageViewsIndex=1
Stomping Out Sparks & Flames

ACS

Have you had a look at the Double D and the 22/7 over at runoffgroove.com?  Both of these are based off the 4049 and produce some brilliant sounds - the 22/7 in particular is a great representation of a BMP...  There's some good technical discussion in there that may lend some ideas...

caspercody

Or check out the Hot Harmonics. Looks like the Double D, and Hot Harmonics use a Jfet stage into (3) CMOS stages. Both get a high gain distortion. The others I found (Hot Tubes, Fallout, Blackstone, and 49er) use a IC stage to drive two to three CMOS stages. But do not get as high gain as the first two mentioned.

brett

Hi
Quotethe 22/7 in particular is a great representation of a BMP
The BMP is basically two inverting gain stages and a simple tone control.  There are big caps in the feedback paths that cut heaps of highs and making it creamy.

From memory, the total gain is a couple of thousand, so try two inverting stages of (22k input resistor and a 100k feeback resistor in parallel with a 470pF cap, which gives  fc=3.4 kHz and gain of about 2 x 40 = 1600 below the fc).

NB You can also use zero input and feedback impedance in a spare inverter to produce nice compression (non-inverted).  Strange, but true.

cheers
Brett Robinson
Let a hundred flowers bloom, let a hundred schools of thought contend. (Mao Zedong)

dschwartz

hey wgtp, i just checked your gallery....impressive work!!
#once is not enough" looks very interesting, like an OD chop suey..looks like you enjoy breadboarding a lot..have you made actual pedals from those?

about the cmos ckt, looks very simple and effective, well that´s what we look after with inverters..
can you describe how it sounds? brown? fuzzy? spongy?
----------------------------------------------------------
Tubes are overrated!!

http://www.simplifieramp.com

WGTP

#5
I made the Tube Sounding Fuzz when the book first came out, many years ago.  In the last several years, I have tried most of the CMOS stuff mentioned (great stuff) and am happy with what I can get out of one now.  Yes, I like to breadboard and I haven't built a pedal since I first got one.  It's cool to see something posted here and then slap it on the breadboard for a trial.  I have 3 small RS models, 1 has CMOS, 1 op amp and 1 discrete stuff.  For some reason mixing different distortion stages together has alwasy appealed to me.  I think that is more like the reality of having a 100w tube amp cranked.   :icon_eek:

The 22/7th is a jewel as are the others.  Just remove the diodes and change the 33k resistors 10k or 15k if I recall.  Sebastian got me thinking of the inverter stages as more like conventional op amps and off it went.

I mostly posted this as sort of a stretch of the Sim/em concept.  Why not Sim/em SHO or Obsidian stages.  It's mosfets even, just complementary ones.

Also Puretube and Fez building a discrete inverter stage was interesting.  Brett has posted some cool stuff too.

Each stage has a gain of 47=470k/10k theoretically for a total of.  This seems to be plenty.  Using a smaller input resistance boosts the whole deal and tweak the input cap.

It sounds like a mosfet overdrive/distori ton.  Zvex used the word "blizzardy".  I prefer mosfets to jfets for the glassy, edgey, air ripping Raging Marshal Stack sound.   :icon_cool:

I have built a 4 stage unit that I like with each gain stage at 10x. :icon_cool:
Stomping Out Sparks & Flames

Joe

I think its a great idea, and definitely a lot simpler. I became discouraged with CMOS because of the noise but maybe metal-films would help?


WGTP

I have tried to keep the resistor values low, using 10k & 100k rather than 100k and 1M, but haven't tried it playing really loud.  I try not to do really loud anymore due to age and hearing lose (but it still feels good).   It my be hissing like crazy, but I can't hear it.   :icon_wink:
Stomping Out Sparks & Flames