Anderton Comparator Fuzz

Started by soulsonic, April 27, 2008, 03:54:44 AM

Previous topic - Next topic

Aph

Okay, it looks like the resistors work as a ground reference or something.
With reference to ground, they keep pin 4 at -4.5v and pin 7 at +4.5v. Without them, the pins are lopsided.
I haven't gotten the scope out, so I don't know if this affects the output symmetry.

caspercody

I hooked up the ground on the Fuzz circuit to the ground from the power, and nothing.

I then cut the ground wire I just installed, and it works, but only powered by itself.

Here is only difference from your posted schematic:

I forgot to put in the 1K resistor on the LT1054

I used a 3pdt switch, and added a led and 1k resistor. The ground from the power goes to the ground connection on the LT1054 board, and to the middle lug on the 3PDT switch, and that is it.

All ground points shown on Fuzz schematic are connected together, but not to power ground.

Thanks
Rob


Aph

#22
Do you have pictures?
You're not building the LT1054 circuit outside of the box and trying to plug it in through the power jack, are you?
I've confirmed this circuit.

amz-fx

This is an odd design, that was meant to be powered with a battery.

The ground of this circuit is biased to 4.5v and this is connected to the in/out jacks, which might be causing some of the problems. What is ground to the circuit is only internal to it, and outside connections have +4.5v on them.

I have redrawn the circuit so that this can be seen better:



I had planned on fixing this last night but something came up. I'll try to get it breadboarded tonight to verify.

regards, Jack

diffeq

Quote from: amz-fx on February 07, 2018, 08:09:53 AM
This is an odd design, that was meant to be powered with a battery.

The ground of this circuit is biased to 4.5v and this is connected to the in/out jacks, which might be causing some of the problems. What is ground to the circuit is only internal to it, and outside connections have +4.5v on them.

I have redrawn the circuit so that this can be seen better:



I had planned on fixing this last night but something came up. I'll try to get it breadboarded tonight to verify.

regards, Jack

Wouldn't it make more sense to bias the input at +4.5 in front of a buffer, then run it into comparator?

rankot

#25
Or use DC voltage invertor to provide -9V, so opamp can be run at ±9V, while ground remains common, as already proposed?
  • SUPPORTER
60 pedals and counting!

Aph

Ground is ground. The circuit I posted Works!
I have it breadboarded right now and it is daisy chained to several other pedals.

amz-fx

Quote from: Aph on February 07, 2018, 09:30:12 AM
Ground is ground. The circuit I posted Works!
I have it breadboarded right now and it is daisy chained to several other pedals.

Your LT1054 circuit will work fine, just as you mentioned, since it is putting the Vbias at essentially 0v.

The original circuit wants to (weakly) pull the ground up to some arbitrary voltage and this can cause problems on certain power supplies, as some have already noted. I'm not sure why, but I have a theory or two  :icon_biggrin:

Best regards, Jack




Aph

Thanks for the clarification, Jack.
Just as a note... The 1K resistor in the LT1054 circuit lowers the voltage to get about +4.5v and -4.5v (referenced to ground), otherwise the voltages would be +9 and -9, which works... But it seems to sound better with the lower voltage.

rankot

Why not like this:



Single IC (TL072 or similar dual op amp), input buffer biased to vref and that's it.
  • SUPPORTER
60 pedals and counting!

caspercody

I can get a picture on Friday night

The LT1054 is inside the box and only commutes to this pedal.

I am using a One Spot power supply

amptramp

Quote from: R.G. on April 27, 2008, 10:26:22 AM
I always liked that one too. The 741 might not be the best opamp to do that with these days, as there are quieter ones, and also opamps that have a cleaner entry/exit from overload. The opamp is always in overload, either banged against the top power supply or the bottom.

The only thing I didn't like about it is that there's a DC transient at the start of every note. The resting state of the op amp between notes is always either full-positive or full-negative, so the output cap is charged to the full power supply one way or the other. When the note comes along, the DC level on the output changes to half the DC supply, and so the output after the output cap has a transient to be passed along.

I think if I were working with that today, I'd do a side chain which provided a "note present" gate signal that starts the note off a bit softly. That would open the note with a low gain to the DC transient and soften the "snap" a bit.

The other thing that would be interesting is to use a real comparator like the LM339 or one of its eight-pin variants as the comparator. This might produce lower noise.

There is the possibility of using a window aperture comparator which has a low, middle and high output level so it starts at the zero (or middle level) voltage and has three possible outputs.  This avoids the rail-to-rail snap between states because everything starts at the mid level.

Aph

Quote from: rankot on February 07, 2018, 03:46:06 PM
Why not like this:

Single IC (TL072 or similar dual op amp), input buffer biased to vref and that's it.
Breadboarded.
Yes, it does work, but it's not as "responsive" and has even less sustain than Anderton's.
Make the first stage a gain stage?

amz-fx

Here is what I came up with:



Breadboarded and verified to turn a small sine wave into a square wave.  :icon_mrgreen:  Will work with a daisy chain power supply. Use a TL071 for best performance.

regards, Jack

Aph

Quote from: amz-fx on February 07, 2018, 10:55:09 PM
Here is what I came up with:

Breadboarded and verified to turn a small sine wave into a square wave.  :icon_mrgreen:  Will work with a daisy chain power supply. Use a TL071 for best performance.

regards, Jack

Thanks, Jack. Interesting design.
But... (here it comes), it's still too comparator-ish. The sustain is pretty good. It's a little brighter than Anderton's (but I think I prefer his: slightly darker). Also, the original circuit has quite a bit of sustain and doesn't "splatter out" as much when it finally does die. A lot of comparator circuits can sound too digital. Anderton's circuit sounds more "analog". This all may be due to the difference in IC's... I don't know. Maybe I'll try your version with a 741.

amz-fx

Thanks for the report Jody.

It is related to the IC chip since I could see a definite difference in the output at low levels on the o-scope when I tried it with a 741. The lower output impedance of my version probably contributes to the bright sound too.

I listed the TL071 because it was more synth-like and comparator-ish  than the 741 since I thought that was the object (make it buzzy) :icon_mrgreen:

Not only did I fix the ground issue and lower the output Z but also made it non-inverting.

Thanks again for testing it!

Best regards, Jack

Aph

Yes, it's a useful sound. Putting a compressor in front of it helps the sustain and adding a chorus and wah at the end makes it even "synthy-er".

So, I tried a 741 in several combinations of your circuit and Ranko's circuit to try and get the "Anderton sound" while keeping it daisy chain compatible.
No go. They all worked to some extent, but I think the way he applied power has a lot to do with the sound and sustain!

amz-fx

Thanks for the report. Very interesting.

Does your solution with the +/- 4.5v supply using the charge pump sound more like the original battery version?

regards, Jack

Aph

Quote from: amz-fx on February 08, 2018, 02:32:49 PM
Thanks for the report. Very interesting.

Does your solution with the +/- 4.5v supply using the charge pump sound more like the original battery version?

regards, Jack

As far as I can tell, they sound identical... with this observation: The original power adapter I used gave me around +/- 4.9v with the 1K resistor in the charge pump circuit. I then switched to an adapter that was giving about +/- 4.2v. But this was a few weeks later, so I thought "still sounds the same". Just now I tried a fairly new battery that was supplying about +/- 4.7v and I COULD tell a very slight difference (battery sounded just a touch more "bitey"  :icon_rolleyes:). So, I lowered the charge pump voltage drop resistor to 820 ohms which gave close to +/- 4.9v. Then there was no difference that I could hear.
Sounds like voodoo, I know  :o

amz-fx

#39
Rather than depend on the voltage drop across the resistor, here is an idea for a specific voltage drop by using a zener diode. The 4.3v drop across the diode lowers the V+ from 9.0v to 4.7v. Any noise contributed by the zener should be filtered by the 10uF on pin 8. Of course, the zener type could be changed to give a different voltage. The 1N748 is 3.9v for example.



regards, Jack