Weird Buffer Behavior

[karis12]

I recently built myself a simple, single op-amp buffer using Jack Orman's (amz-fx) Basic Buffers article on his website, specifically the voltage-divider biasing type. I used a 220nf input film cap, a pair of 2.2Meg resistors, and a 10uf tantalum output cap. This is meant to be an always-on device (no switch), which only turns off when the input jack is unplugged.

Upon first trying this with a TL061 I noticed that when plugging something into the output jack first before the input, there is a shocking amount of noise (sometimes white noise, and sometimes a slow sine wave) that is present with the signal.

Doing the opposite (input jack first before output), or hitting the foot switch on the next pedal in line gets rid of it instantly. Moreover, replacing the TL061 with a NE5534 took care of the problem entirely - now, whether I plug into the output or input jack first, no noise is present.

Doing some further investigation I found out that grounding the output jack when the noise becomes present gets rid of the noise, but adding a 1Meg pulldown resistor (not present in the original schematic by Mr. Orman) does nothing at all.

Any experience with this, or similar issues? I'd really like to use op amps with lower power consumption, especially since I don't really hear an appreciable difference in noise performance between the TL061 and the NE5534, when the former is functioning correctly.

[garcho]

specifically the voltage-divider biasing type.

Basically, all single-supply op amps use a voltage divider to provide a voltage "bias" for audio signals. There is a ton of information on this forum and elsewhere online regarding what this is and why we use it. Any schematic you see that has points marked "Vref" or "Vb" is referring to this voltage bias. Ask if you can't find any or don't understand.

which only turns off when the input jack is unplugged

That makes it a good candidate for a DC adapter power supply instead of a battery.

Moreover, replacing the TL061 with a NE5534 took care of the problem entirely

The '61 can be noisy but not really like how you're describing. It's a JFET, the 5534 is a BJT, both should work. There might have been another issue not related to op amp choice. Do you have any other op amps to try? 741, tl071, etc.?

Doing some further investigation I found out that grounding the output jack when the noise becomes present gets rid of the noise, but adding a 1Meg pulldown resistor (not present in the original schematic by Mr. Orman) does nothing at all.

Grounding the output jack would definitely get rid of the noise, as it should get rid of any and all signal. Way different than adding a pull-down resistor, which is there to prevent switch "popping". Can you be more specific about what you did?

I'd really like to use op amps with lower power consumption

Why is that a priority for you?

[amptramp]

The TL061 is optimized for low current consumption at the expense of noise performance.  Just to compare it to the TL071 which is the most common op amp used in DIY stompboxes:

TL061

Input voltage noise: 42 NV/sqrt Hz
Supply current: 0.02 - 0.025 mA/amplifier
Slew rate @ unity gain: 1.5 - 3.5 V/µsec

TL071

Input noise voltage: 18 NV/sqrt Hz
Supply current: 1.4 - 2.5 mA/amplifier
Slew rate @ unity gain: 8 - 13 V/µsec

The TL061 is often used for LFO's because the current is low enough to avoid a "tick" from current spikes but is too noisy for signal amplification.

[karis12]

I followed this schematic, save for the input cap being a 220nf, and the resistors 2.2Meg.

The reason why I'm prioritising low power consumption is because I intend to run it primarily via 9V battery. I've actually finished building it, and everything fits snugly.

Here's what the thing looks like inside the 1590a enclosure:


I forgot to mention that I tried using a TL071 as well, and with identical results (weird noises when plugging jacks a certain order), so I'm thinking something about the schematic, or perhaps one of the parts I selected, doesn't play well with the TL series op amps.

[antonis]

garcho told you to ground Output when pedal is OFF...

You can easily obtain that via a "switch type" output jack (if you don't use a 3PDT Switch..)

[karis12]

garcho told you to ground Output when pedal is OFF...

You can easily obtain that via a "switch type" output jack (if you don't use a 3PDT Switch..)

I happen to have such a jack in my parts box, but it's not particularly robust. I'll look for a better one on my next parts shopping day.

[slacker]

It might be oscillating, try putting a 100 Ohm resistor in series with the output, between "out" on the schematic and the tip of the output jack.

[induction]

garcho told you to ground Output when pedal is OFF...

A bit of confusion here. He didn't tell you to ground the output. He pointed out that grounding the output will kill all signal, not just the noise. It's not clear that a different kind of output jack is going to help at all here.

Note that there is no power filtering on that schematic. Even with a battery, a 100u filter cap between the V++ rail and ground can help prevent oscillation. A low value resistor (e.g. 100R) between the V++ source and the positive leg of the filter cap wouldn't hurt either.

[antonis]

garcho told you to ground Output when pedal is OFF...

A bit of confusion here. He didn't tell you to ground the output. He pointed out that grounding the output will kill all signal, not just the noise. It's not clear that a different kind of output jack is going to help at all here.

Quite right..!! 

So, it's me that tell him to ground the output..
( a "floating" 10μF Cap is an appreciable noise maker..)

[karis12]

UPDATE: I solved the problem! Turns out the double-sided foam tape I used to secure the board to the enclosure failed to insulate the underside, causing a short. I just put on some clear tape on the side of the box, and redid the foam tape, and the problem was solved. Now I can use a TL061 without the strange noises, and it doesn't matter whether I plug into the input or output first. Thanks for all the help!

[PRR]

> put on some clear tape

Tape won't last. In a recent discussion, milk-jug plastic was suggested as fairly tough, long-lasting, flat (unlike detergent bottles), and readily available.