Do I need any more components at this output stage?

[Balthazar]

Hi

I'm trying to make this guitar effect box that uses a state variable filter, which just means 3 op-amps that together form a low pass, high pass and band pass filter. The schematic for such a filter, with a UAF42 IC is here:



From the little knowledge I have of electronics, I'd think that the signal coming out of this chip, from the HP, BP and LP terminals, would be ready-to-go, so to speak. That I wouldn't need any more components after it before I send it to the tip of the output jack. It's coming from an op-amp, and is supposed to have a low output impedance, which is supposed to be great from what I know before (non stomp-box related). Since I will filter out any DC bias at the input, the output is not supposed to have any DC bias to be filtered out, and LTSpice simulation shows absolutely no bias at the output.

From other threads I see that someone recommend some components to reduce pops from bypass switches and the like, so I thought I'd mention that I'm planning to add one selector switch to select the type of filter, and one bypass switch, like this:



Yet, I have searched through a lot of schematics of different stomp boxes (via this site), and can't find one single schematic where they send the output of an op-amp right to the output jack. They always have something more, usually involving a capacitor in series with the op-amp output and at least one resistor to ground, like this:

http://www.montagar.com/~patj/harmswtn.gif

Often they also have a resistor in series between the op-amp output and the capacitor. Or after the capacitor. But perhaps all these circuits do have a DC bias to filter out? Or does this almost compulsory output capacitor do something more too?

So, do you recommend any more components in the circuit after the filter chip?

[Balthazar]

PS: The UAF42 schematic above is from the datasheet for the chip, here: http://www.ti.com/lit/ds/symlink/uaf42.pdf

I've redrawn it in a graphics program because everything is a copyright issue these days.....

[Keppy]

The schematic assumes a bipolar supply. If you're using 9v single sided supply, you'll have to create a 4.5v supply and use that in place of ground on the schematic. You'll also need blocking caps in & out.

[Keppy]

Bear in mind as well that the input impedance of this circuit is equal to Rg, so if that's a low value you'll want a buffer in front as well. Ideal input impedance for guitar gear is usually about 1M.

[Keppy]

From other threads I see that someone recommend some components to reduce pops from bypass switches and the like

These "pulldown" resistors go to ground before the input cap and after the output cap. Often a volume pot acts as the output pulldown resistor.

Often they also have a resistor in series between the op-amp output and the capacitor.

This is to protect the output of the opamp by restricting the current on the output during power up/down. Usually not needed, but a good failsafe for those unusual times. Plenty of designs use this, but plenty of others leave it out.

[Balthazar]

Thanks for your help. I'm sorry, I should have provided more context. I have posted about this project earlier on the forum, and decided to use a charge pump to provide a bipolar +/- 9v supply. From other members suggestions, I have chosen to use the auxiliary op-amp that comes with the UAF42 chip to buffer the input. Rg will be a 3.3k resistor in series with a 50k pot, so it will be too low for anput impedance without the buffering. Earlier posts about the project are here:

http://www.diystompboxes.com/smfforum/index.php?topic=100720.0

And here:

http://www.diystompboxes.com/smfforum/index.php?topic=100679.0

The hand drawn schematic in the first link contains an error in the input stage. I have changed it to one 2M resistor to ground, followed by one capcitor in series with the signal, and one 1M resistor to ground after it. It will be about 667k input impedance I think. Does that sound OK?

[PRR]

If the outputs stay in the box, and DC has been considered, just do it.

You *may* find that the "zero bias" outputs really sit at +0.01V or -0.007V, stray "small" DC error. And when you switch from one to another you get 0.017V jumps. Relative to 0.02V-0.2V signals, this is loud POPs, annoying. Even in bipolar work it is not uncommon to cap-couple often.

If they go OUT of the box, it can be wise to throw in a few hundred ohms of series resistance. That helps protect against shorts (shorts happen, specially when you walk on cables), although the chip may already be hard to kill with a short. More urgently, cable capacitance against the amplifier supersonic roll-off can often cause supersonic oscillations with 100p to 1,000p of cable, which is 3 to 30 feet, which is just what you usually have. A few hundred ohms between chip and jack lets the amp do its thing without dragging the whole cable.

[Balthazar]

Thanks, that was very valuable input. The output is going out of the box. I've simulated it again in LTSpice with no signal, and it shows a DC offset of about 10µV between the channels. But these models do not take into account all the imperfections of the components, wire, etc. Therefore I will take into account DC offset between the channels. I suppose that means a seperate output capacitor for each of the channels? You are absolutely right, all the schematics I've found so far with bipolar power supply still includes an output coupling capacitor.

Also, the UAF42 chip is spec'd to handle a short circuit at the output, but I don't want any oscillations. I'll add a resistor too.....