Sallen Key issues, and also hello I'm new!

[cherler]

Hey er-body! I'm working on a super wooly bass fuzz kinda circuit for for a project and I wanted to throw an active low pass filter on it. I chose a Sallen Key and the filter and I'm getting some weird issues with my Q control. Here's my circuit:

http://i.imgur.com/iRO7L0V.png

When I increase the resistance in the negative feedback loop, I'm not getting what I expect. I loose output and I get these weird squelchy bursts of treble if I use a heavy attack on the guitar. Any thoughts? I guess I could have this set up right and the knee for the filter just happened to be at higher frequencies and when enough content around that range comes through the filter oscillates a little?

Also I'm new here! I like weird noises, fuzz, beer and over simplifying things!

[PRR]

Welcome to the forum!!

I have never seen an image 15,000 pixels wide and I'm not seeing this one well enough to comment on your apparent question.

The FETs probably want gate resistors. The 4.5V voltage divider will need a bypass cap or you will have sneak-through which could be your filter problem.

Part-numbers would make discussion easier (we can't point-at your paper or chalkboard).

[Transmogrifox]

Yes I had to download and resize this image.  +1 one on part designators (numbering)

There is one blaringly obvious problem, and that is DC bias:

Look at the 100k resistor to ground from the inverting terminal of IC1A (Pin 2).  That should not go to DC ground.  2 ways to deal with it:
1)  Connect 1uF capacitor in series to ground for DC blocking
2)  Make a voltage divider with 2 200k resistors, one going to +9V and the other to ground.

As you increase the pot resistance, the DC bias at the opamp output (pin 1) goes to the rail and you increasingly get really ugly nonlinear/noisy behavior.

Easy to see with a DMM.  Connect a DMM to Pin 1 of IC1A and you will see the DC output is not 4.5V as you expect.  I'm surprised you get any sound at all even with the pot at 0, if that 100k series resistor is actually there in your circuit.

You really should have a limiting resistor in series with each one of these 100k pots in the frequency setting location, something the op amp can drive (>1k).  Also the 4u7 becomes interesting in terms of feedback loop stability when those pots get down below 1k.

Nasty things will happen when you turn those pots full clockwise without series resistances. 

[cherler]

Thanks so much guys! Next time I'll actually go into Eagle and just increase the font size instead, oops.

The DC bias for the inverting input makes a lot of sense actually. All my electronics knowledge comes from robotics, microcontrollers and sensors where you see the negative feedback loop I made for a lot of essentially DC signals. The capacitors makes a lot of sense.

I've been planning on adding some series resistance with the frequency pots, I've just been keeping the knob away from all the way down cause I'm super lazy. The same goes for the gate resistors, I'm gonna add a gain control soon that'll put in some gate resistance.

I never actually knew the capacitor on voltage dividers was for AC bypass. I thought it was similar to a decoupling cap used to smooth out any ripple, I'm glad I got to learn something today!

Thanks again guys! I'd say this was a very productive first post, for me at least.

[Transmogrifox]

I never actually knew the capacitor on voltage dividers was for AC bypass. I thought it was similar to a decoupling cap used to smooth out any ripple, I'm glad I got to learn something today!

The answer is it is both.  "A decoupling cap to smooth out any ripple" is by definition an AC bypass....but technicality aside I know what you mean.

You were probably thinking it's there for noise filtering only -- to prevent power supply noise from getting into the audio path.  It serves the job as a power supply noise filter AND it rejects AC signals from different parts of the circuit. For example, when the voltage divider is used as a DC reference, the bypass cap prevents AC signals from coupling from one amplifier to another and injecting AC signals where you don't intend them to be injected.

In the case we're considering, using a capacitor in the grounded leg of the opamp feedback loop, its purpose is a DC blocking capacitor.  That way your op amp looks like a unity gain follower for DC, but it's an amplifier with gain for anything that wiggles.

[anotherjim]

Schematic shows the input cap reverse. If you've fitted it that way you will likely get odd behaviour when you use the guitar volume control. That's because a reverse electrolytic cap leaks current and the volume control will be varying that leakage current as you turn it. In fact, if your gate bias resistor for the first FET is the usual 1M to ground, the input cap can be a much smaller non-polarized type (Film cap),say 10nF. Probably all your interstage coupling caps can be smaller value film types.