# Op-amp boost pedal tonal noise

I am working on a boost pedal, but I can't seem to find the cause of a humming noise in the circuit.

I tried bigger caps for C5 and C6, but it didn't solve the issue. For power I use an 18V model train power supply (it's the only supply I have now.) I divide it for the TL072. The 6V is there for bias.

Spectrum of the hum:

• Perhaps the problem can't be seen from the schematics. Also the voltages in the schematic are not correct. 18V VCC divided by 3 is 6V, not 4.5V. If the op-amp positive is 9V, where does it come from, and what is the negativr supply for it? Why there are no bypass capacitors for op-amp supply? Is any part of the circuit floating? Capacitive coupling via power supply? Jun 16 at 10:06
• Is it 50/60 Hz humming? Jun 16 at 10:30
• I think it's not 50/60 Hz it's higher, but i can't measure it. Jun 16 at 10:39
• Spectrum looks like mostly 100Hz and 200Hz, so that would be 2nd and 4th harmonics of your 50Hz mains. A train power supply is probably going to have no filtering or regulation at all. You need to either get a better supply or add filtering and possibly regulation. Will this eventually run on batteries? Jun 16 at 14:38

You use model train power supply. Those supplies often contain no capacitors to filter the rectifier output. I assume that the OpAmp is fed with 100 Hz pulsing DC, a proper voltage regulator is required.

To start, I would change C5 and C6 - a lot. As shown, the filter at C6 has a lowpass corner frequency of 22 kHz -ish. This will do nothing to reduce hum on the DC power source. As it is, incoming power supply hum is injected directly into the opamp input through R2.

You don't say how much you increased them, but I suggest a 100 uF electrolytic and 0.1 uF ceramic in parallel for each one. This lowers the corner C6 frequency to around 0.5 Hz. 60 Hz is about 7 octaves higher, so the hum attenuation would be around 40 dB, a 100x improvement. This is a rough guess number. Because there are two R-C filters in series, the actual attenuation will be better.

Also, as above, decrease R7 to 1 K or less. This changes the C6 corner frequency calculation, but things still will be way better than before.

You can try lowering the values of the voltage divider resistors, or better yet use a regulator.

As it is you are dropping 6 V across 4700 $$\\Omega\ \$$ which give you

$$\frac{6~V}{4700~ \Omega} = 0.00127~A$$

That's about what the quiescent current of the opamp is going to be, so what's going to happen is that your '12 V' is going to be more like 9 V or less, and with signal the opamp will draw more current so that voltage will be even less and fluctuating. This might be why you originally had labeled it as 4.5 V for the bias, with a 1.2 mA load from the opamp you'd get around 4.5 V across the bottom resistor of the divider. Did you base the original value on a measurement or did you just calculate it incorrectly?

This may be the cause of the noise you are getting or not, either way it's the first problem you'll need to address.

You'll also need to filter the power better, 1nF capacitors would more be used for RF bypassing than ripple filtering. C5 should probably be in the hundreds of microfarads range, you can get away with lower values if your 18 V supply already has good filtering, but even then you want the capacitors to have low impedance at audio frequencies.

As @Jens suggests "greatly" ... A good power is needed.

Here is a picture for "explaining" somewhat the "spectrum" of the output.
NB: it is only for one configuration of "variables".
Some "hum" (pseudo triangle) on supply (100 mVpp vs 18V), signal (1 Vpeak).
Probably "saturation" on output.

Here is without saturation (signal = 0.1 Vpp, but with the same "hum").

With a "clear" supply ... no saturated opamp output ...