Guitar - Photodiode Amplifier, Signal-to-Noise ratio

[Audioguru]

The OPA380 has a max allowed supply voltage of only 7V. Your new schematic has NO supply voltage.
Your new recording has less hiss than the first circuit.

Your extremely low value input resistors are shorting and reducing the photodiode signal and should be removed.

 

[Yarrrrr]

The OPA380 has a max allowed supply voltage of only 7V. Your new schematic has NO supply voltage.
Your new recording has less hiss than the first circuit.

Your extremely low value input resistors are shorting and reducing the photodiode signal and should be removed.

I'm sorry I'll try to remember having the relevant voltage values visible in future schematics, I've never had to share these with anyone before.
There is a 5 volt linear regulator for the OPA380 circuit.

I will try your suggestions on my new circuit and get back to you with the results.

 

[betwixt]

Summarizing:
1. Use the photo-diodes as voltage sources, they need a DC current path to work properly, either formed by the feedback path in a transimpedance amplifier or by using a high value load resistor. Do not drive them into a capacitor because regardless of what filtering you want, it will not allow DC current to flow. Their output is tiny, driving their output into a low resistance will 'short out' the signal.
2. Control the LED current so the photo-diodes are working somewhere close to giving mid-range output current. You can monitor the light level by looking at a filtered output of the transimpedance amplifier and from there control the LED current.
3. Change the output stage so both XLR signals are identically driven but with one inverted relative to the other.

It might be worth researching optical pick-ups for vinyl records as the technology is very similar.

Brian.

 

[Yarrrrr]

Thanks to everyone who kindly gave me some tips yesterday I am making progress.

I've removed all input 100 ohm resistors, and the filter capacitor and resistor.
Will experiment with the LED current now

https://soundcloud.com/don-stroganotti%2Fpcb-test-2%2Fs-3RUePyGrQ4S

--- Updated Feb 4, 2021 ---

Update:

Halving the resistor for the LEDs gave me about 4.5 dB improvement in signal to noise.

https://soundcloud.com/don-stroganotti%2Fpcb-test-3-100-ohm-led%2Fs-1tN7sRwvppS

 

[Audioguru]

R1 is not needed and is shorting the photodiode output and increasing opamp noise. The - input of an inverting opamp is automatically biased at the same voltage as the + input by the negative feedback resistor.

Have you tried it with no parallel opamps? Then you will have 3 series opamps, 2 with low gain and one with no gain. Only a single opamp can probably do it.

Hey, my little Gateway ported 3" pc speakers with large magnets play your music like a real acoustic guitar!

Of course the signal to noise ratio is increased when the IR IEDs are brighter. Interference from AC light bulbs will be reduced.

 

[Yarrrrr]

Ah I forgot to remove R1 from the schematic after experimenting, it is not in the real circuit.

The last example on page 1 in the thread is a single op amp, so it will most likely work perfectly fine with a single one.

I have previously seen measurable difference in the noise floor when using an oscilloscope, comparing parallel with single.
But If I recall correctly that was with the 100 ohm input resistors in the circuit, so it might not be applicable any more.

I will most likely repurpose the extra op amps into an active high pass filter and use the two last ones for the balanced XLR stage. Currently the circuit is very susceptible to hand movement above the strings that quickly lead to clipping the signal so I need some sort of filter for low frequencies.

What more can I do to improve the signal to noise? I'm guessing I will hit a limit very soon with how bright I can make my LEDs without frying them.
Maybe there are more sensitive photodiodes.
I've seen there are two legged phototransistors, is it possible to try them in the circuit as is or do I need to do some modifications?

 

[c_mitra]

Instead of using paralleled inverting opamps with extremely low input resistor values, try using a single non-inverting opamp with a very high input resistance so that the tiny voltages from the photodiodes are not loaded down so much. Then reduce the gain.

Or use a reverse bias on the photo-diodes? Then reduce the gain.

I suggest both.

Use a single low noise op-amp. Multiple op-amps do not serve any useful purpose in mu opinion.

Use a single reverse bias photo diode as detector.

What you are going to do with the ambient signal that appears as noise?

Use a modulated (RF) IR source in the excitation. Use AC coupled amplifier and tuned amplifier and then demodulate.

You can use the diodes used in common remote controls.

 

[Yarrrrr]

I suggest both.

Use a single low noise op-amp. Multiple op-amps do not serve any useful purpose in mu opinion.

Use a single reverse bias photo diode as detector.

What you are going to do with the ambient signal that appears as noise?

Use a modulated (RF) IR source in the excitation. Use AC coupled amplifier and tuned amplifier and then demodulate.

You can use the diodes used in common remote controls.

If you check two posts above I have now tried without the input resistors which resulted in a significant improvement.

I'm not sure what you mean about the ambient signal.

And the part about modulated signals is beyond my understanding, will I need to incorporate a microcontroller?

 

[Audioguru]

Light bulbs are modulated by the 50Hz or 60Hz of electricity and compact fluorescent light bulbs operate at around 38kHz which might cause interference with your IR reception. A spotlight might saturate your photodiode and mute the audio.
They will be in your ambient.

 

[Yarrrrr]

Ah, thanks for the clarification.

The matched LED and photodiode at ~950nm seem effective enough for my lighting conditions. And covering the sensors completely in a fairly bright room does not change the noise floor as far as I can tell measuring with an oscilloscope.

 

[c_mitra]

I'm not sure what you mean about the ambient signal.

Just consider your radio/TV remote control. Light from an IR LED coming from a distance of 5-10ft (2-4m), reaches a detector.

The detector is covered with a black (IR transparent but opaque to visible light) film to reduce the effect of surrounding lights.

But this is not enough; ordinary light also have considerable amount of IR and they will also produce a signal at the photodiode.

In fact the amount of IR from the remote that falls on the photodiode is much smaller than the IR present from all other sources.

So what is the solution? Modulation, like one in a radio (or TV or your mobile phone) can help. But how?

Using any form of modulation helps to shift your desired signal to a higher frequency where you may have much less noise.

All IR remotes use modulation to avoid interference from ambient light (also from other remotes). The basic principles are very simple.

But if you want six IR LED source for six strings, you must have six independent amplifiers. You cannot combine them into one.

 

[betwixt]

Another factor that might cause a problem for you: In my country there are two languages English and Cymraeg, most public venues have translation equipment available. It usually works by having someone manually translate what they hear in one language, speak it into a microphone that modulates IR LEDs around the room. People wanting the translation wear headphones with IR receivers built in to them so they can hear the translation. The equipment is frequently left switched on when not in use and it floods the room with IR light. If you are in a similar situation and play guitar where one of these systems is installed it will almost certainly interfere with the sensors.

Brian.

 

[Yarrrrr]

All IR remotes use modulation to avoid interference from ambient light (also from other remotes). The basic principles are very simple.

But if you want six IR LED source for six strings, you must have six independent amplifiers. You cannot combine them into one.

I doubt ambient light is an issue for my use case, it seems to currently be below the noise floor of the circuit if it is.

Could you explain why six amplifiers would be better? Wouldn't the signals and noise get summed all the same?

 

[Audioguru]

Each of the six amplifiers could have its own bandpass filter for the single note played by each guitar string. Then the interference noise and all audio high frequencies will make the guitar play the 6 notes without noise and without a guitar's sound.
But a guitar player uses the frets to play many notes on each string then bandpass filers for each string will not work.

 

[betwixt]

The most obvious use of individual amplifiers is to allow feedback to the LED brightness so each pick-up is optimized for SNR and linearity.

I may be overlooking something here but where exactly are the LEDs and Sensors positioned? The optimum place for sensing movement of the string will be half way between the bridge and the fret being used so obviously some compromise would be needed but I would think near the base of the neck would be best. Then there is the problem that the string position may not relate to the light reflected from it, in fact lookup directly up at the string would show equal change whether it was displaced one way or the other so there would be a tendency to second harmonic distortion.

Brian.

 

[Yarrrrr]

I may be overlooking something here but where exactly are the LEDs and Sensors positioned?

The pickup is in the neck position to capture as much of the string movement as possible. The LEDs are aimed at a 60 degree angle towards the string above it.

 

[Audioguru]

Does a guitar string vibrate only horizontally or does it vibrate in a circle making vertical, horizontal and all in between directions as it vibrated? That would cause level changes in your circuit that assumes only horizontal vibrating.

 

[Yarrrrr]

I think it's an ellipse based on the direction you pick the string.

 

[Audioguru]

I agree that I have seen a guitar string swinging around in an ellipse. Then I think it will modulate the levels from your photodiodes. I do not know why it does not sound modulated since swinging vertically will vibrate the guitar bridge and body a lot more than if it is swinging horizontally.

 

[barry]

I agree that I have seen a guitar string swinging around in an ellipse. Then I think it will modulate the levels from your photodiodes. I do not know why it does not sound modulated since swinging vertically will vibrate the guitar bridge and body a lot more than if it is swinging horizontally.

It DOES sound modulated. That’s why different guitars and different players have different tone. It’s all about the harmonics, baby.