Post e-books, knowledge for *fixing* the squilling noise in high gain pedals!

Started by Angelo777, May 02, 2010, 12:50:15 AM

Previous topic - Next topic

Angelo777

Hello everyone!

Its been a while since my last posts .... being busy with Phil reverse engineering the iSP Decimator and building my extreme distortion pedal  :icon_biggrin:

I'm really into extreme high gain pedals...

Please help me and the others members in this forum by posting any knowledge you may have about the noise and the squilling noise that we get in extreme high gain pedals when we palm-mute the strings of the guitar!!!!

Any books that you know that contain knowledge about it.... any ideas... PLEASE ENLIGHTEN US ALL  :)

Thank you all in advance!!!!! I'm pretty sure that a lot of members are totally interested in learning all these.... plus me :P

Best Regards,
Angelo777

philbinator1

"Hows are we's?  We's in the f*cking middle of a dinners meal!  Dats hows we am!" - Skwisgaar Skwigelf

R.G.

The noise you get is because because you're getting unwanted feedback.

There are three requirements for sustained oscillation. These are
1. A feedback path from output to input. There will always be some losses in this path, so the "gain" from output back to input is always less than 1.
2. Enough forward gain in the circuit to make up for the losses in the feedback path so the total gain from the circuit input to the output and back through the feedback path is greater than 1. This lets the circuit provide its own input.
3. Signal phase change from circuit input, through the circuit, through the feedback path back to the input so the phase of the fed-back signal reinforces the input, instead of opposing it.

These three are pretty universal. All oscillators work this way. It's just the details of the three factors that vary.

Notice that item 1 is always present. In fact, there are many feedback paths from output to input all the time. These include but are not limited to
- capacitance between input and output wiring or circuit traces; there are always many, many of these
- resistance in the "ground" wires
- resistance/impedance in the power supply wires
- intentional or unintentional circuit components that leak output back to the input

If you keep turning up the gain (Item 2) then anything will oscillate at some point. Full stop. This is what you're doing. There is always a gain beyond which it will oscillate.

Item 3 is also always present. The feedback paths are a mixture of resistive, capacitive, and inductive side effects of there being wires and circuit boards, and even if you have used an inverting circuit, there is some frequency where the phase is changed to positive feedback by the reactive nature of the parasitic capacitances and inductances.

In theory, stopping oscillation is easy. All you have to do is either (1) turn down the gain or (2) make the losses in the feedback path be high enough so that the loop gain from input through the circuit and back through the feedback path is less than one.

In practice, turning up the gain is what you want to do, so you can't do that. You're stuck with 2.

You need to find the feedback path(s) that are causing the squealing and introduce signal losses that keep the overall loop gain under 1. If you've read anything here, you've noticed that I'm always harping on the need for short wiring paths, clean and decoupled power supplies, and proper grounding. That's because those are what make your circuits oscillate if you don't do them.

Shielding is what you do when you don't know what else to do or when you have exhausted every other possible remedy. Shielding works by introducing more attenuation into the feedback path by shunting some of the feedback signal to ground. Shielding works, but it is a very expensive remedy in many cases. In most cases good layout and wiring can eliminate shielding.
R.G.

In response to the questions in the forum - PCB Layout for Musical Effects is available from The Book Patch. Search "PCB Layout" and it ought to appear.

Gus

There is also signal coupling via the power supply.
EDIT I mean this in addition to R.G.s"resistance/impedance in the power supply wires"  The PS battery etc is also a a part of the circuit.

GibsonGM

  • SUPPORTER
MXR Dist +, TS9/808, Easyvibe, Big Muff Pi, Blues Breaker, Guv'nor.  MOSFace, MOS Boost,  BJT boosts - LPB-2, buffers, Phuncgnosis, FF, Orange Sunshine & others, Bazz Fuss, Tonemender, Little Gem, Orange Squeezer, Ruby Tuby, filters, octaves, trems...

Angelo777

Hi everyone :D

QuoteSignal phase change from circuit input, through the circuit, through the feedback path back to the input so the phase of the fed-back signal reinforces the input, instead of opposing it.

Ok... so first of all i use double shielded wires everywhere. I was thinking of putting a cab simulator at the end of the circuit to cut major db from some frequencies that oscillate like crazy...

Though R.G could you please give me some more info about the thing you said about signal phase change ?

Thank you in advance!!!

Angelo

Angelo777

By the way.....

I just noticed that my pedal has realistically unlimited sustain ... is this has to do to the overloaded feedback which 1) sustains infinetly the notes and 2 ) gives that horrible noice ???

Thanks!
Angelo

R.G.

Quote from: Angelo777 on May 02, 2010, 02:57:01 PM
Though R.G could you please give me some more info about the thing you said about signal phase change ?

If the output of an amplifier goes up when its input goes up, it is a NON-inverting amplifier. If the output of an amplifier goes down when its input goes up, it is inverting.

A non-inverting amplifier has an output that is a replica of the input, but bigger. An inverting amplifier has an output which is a bigger replica of the input but the mirror image - that is, when the input signal goes up, the output goes down, and vice versa.

But in-phase (like from a non-inverting output) and out-of-phase (like from an inverting output) are not the only two phases possible. Inductors and capacitors can cause phase shifts of between zero degrees (in phase) and 90 degrees (halfway between in phase and out of phase) for each cap or inductor. Exactly what phase is caused by an inductor or capacitor depends on the signal frequency and the rest of the circuit feeding the cap or inductor. So the phase of a signal fed through an inductor or capacitor varies with changing frequency.

A non-inverting amplifier's output is already the correct phase to drive its input into oscillation. All you have to do is let enough of it get back to the input, and it will provide its own input, and oscillate.

An inverting amplifier's output is in the direction to reduce its input, so conceptually at least, you can feed 100% of the output of an inverting amplifier to its input, and not have it oscillate. This is exactly what happens with emitter followers. The output is in effect fed back to the input.

But no amplifier has an output phase which is constant with all frequencies. Especially at high frequencies, the parasitic capacitors leak a little signal back to the input. This is by definition going to be between 0 degrees and 90 degrees of phase shift depending on the frequency. However, inside the amplifier there are also parasitic capacitors. These internal caps may have caused a phase change with increasing frequency too, but inside where you can't get at it. All amplifiers do this. It's unavoidable. So an external capacitor may add just enough signal leakage and phase shift to make the thing oscillate. It will oscillate only at the specific frequency where the total amount of phase shift through the amplifier, and back through the feedback path is exactly correct to make the fed-back signal in exact 0 degrees/360 degrees with the input.

The way amplifiers are stabilized is to note that to oscillate, they have to have both a forward gain from input to output of greater than unity, and they have to have a phase shift of less than 180 degrees (noninverting amps already have 360 degrees to start with!). If either of these fails, it will not oscillate.

But you can't control the amount of phase shift. It is what it is, generally from the internal capacitances that you can't do anything about. What you can do is to lower the gain below unity before the frequency where phase shift is 180 degrees. Or, as I stated before, if you have enough gain, anything will oscillate at some frequency. You can not simply increase gain without limit and not build an oscillator. The gain of every oscillator is infinite - it has an output for no input.

From your replies I think you're headed for some frustration. You cannot simply double shield everything and then keep increasing gain. At some point it will oscillate again. And you can't simply notch out the offending frequencies. Filters to generate the notches also add phase shift, an this will only push the oscillation to some different frequency.

Google "feedback amplifier design".
R.G.

In response to the questions in the forum - PCB Layout for Musical Effects is available from The Book Patch. Search "PCB Layout" and it ought to appear.