Signal to Noise Ratio for audio circuit

Thread Starter

KrisUK

Joined Feb 27, 2009
9
I have 2 circuits of my own design I need to find the signal to noise ratio for but I'm not quite sure how to go about it.

Both circuits are guitar effects. One circuit is a cascaded 2 op amp circuit with an RC filter between op amps and 2 diodes as a voltage swing limiter to ground to clip the inputted signal. The other circuit is a negative feedback arrangement using 2 op amps with a RL and RC filter.

I have the noise data for the op amps but I don't know how to take into consideration the filters, diodes and negative feedback.

I thought I could try measuring the output signal with no inputted signal. There was a noise signal there but the oscilloscope was reading a voltage of 220mV and an incoherent signal. The 220mV also changed as I changed the V/div on the scope from 75mV to about 300mV. For a circuit that only has input signal levels of a few hundred milivolts, something like 220mV seems awfully high?

What's the correct way of working out the SNR?
 

Audioguru

Joined Dec 20, 2007
11,248
Sorry.
You saved your schematic as an uncropped huge JPG file type instead of as a very clear GIF or PNG file type.
The parts are too far apart for me to try to read the tiny writing.
The background is dark grey instead of white and is covered with dots.
Can you see it well?
 

Thread Starter

KrisUK

Joined Feb 27, 2009
9
Sorry.
You saved your schematic as an uncropped huge JPG file type instead of as a very clear GIF or PNG file type.
The parts are too far apart for me to try to read the tiny writing.
The background is dark grey instead of white and is covered with dots.
Can you see it well?
I'm afraid that's the background in Proteus :(


How about these images?

The layouts are the same but there's a couple of differences in resistor values.
 

Attachments

Audioguru

Joined Dec 20, 2007
11,248
Thank you for making a normal-looking schematic that can be seen. Proteus is sick.
I cropped the first one so it is not the size on my neighbourhood.
The opamps have no part numbers and no power supply.

The first opamp looks like a clipper (fuzz-maker).
In between the opamps is some kind of tone control.
 

Attachments

Thread Starter

KrisUK

Joined Feb 27, 2009
9
Sorry I forgot those schematics omitted the power supply and part numbers :(

Op amps are TL072 all round.

Power supply is ±9V.

And correct ;) Clipper with a tone control.
 

Audioguru

Joined Dec 20, 2007
11,248
I think the gain is way too high for a signal from a guitar's pickup. Then the opamps amplify their own noise and the clipping diodes reduce the signal level.
 

Thread Starter

KrisUK

Joined Feb 27, 2009
9
The gain is fine. I have both of these circuits built and fully operational with no volume mismatches.

A gain of about 100 is a typical value for clipping effects. A guitar output signal can be easily as low as 100mV. With a reduced gain, there isn't going to be much clipping happening.

I just need to know what I need to take into consideration to calculate the noise. I know the op amps are 5nV/root Hz but I don't know how to take into consideration the passive components.

Are you saying the op-amps are the only significant source of noise and the passive components can be ignored?
 

thatoneguy

Joined Feb 19, 2009
6,359
Physical circuit board layout and power supply usually contribute a good share of noise into an otherwise correctly designed circuit. Keep good power and ground planes, leave as much copper on board as possible when etching, but not to the point all signal paths are vias or jumpers.

With one of the circuits, the clipper, It would be difficult to measure as a signal/noise ratio with a standard noise analyzer, as injecting a huge amount of noise is the purpose of the circuit. "How it sounds" is an objective argument that reaches back to the first amplified guitars. Measuring matched Output of circuit with an impedance matched "Zero" input should show very little noise, only what is generated by the circuit design and assembly itself.

The measured noise on the scope at quiescent, what frequency bands was it in? 60Hz, audible range, RF?

Eliminating noise is a sometimes tedious procedure, and can come from passive components in certain circumstances/designs.

I'm still working on my infinite power wideband amplifier with DC to light bandwidth, 1ppm THD+noise, and stable down to 0.1ohm but I've gotten sidetracked..... :p
 
Last edited:

Audioguru

Joined Dec 20, 2007
11,248
A TL072 has a noise rating of 18nV/root Hz, not 4.5.
An OPA2134 has a noise rating of 8nV/root Hz and will sound quieter.

The high impedance of the pickup coil creates noise and the very high resistance values in your circuits create noise.

Metal film resistors are quieter than carbon film.

I am talking about hiss noise, not hum pickup.
 
Top