Diy optical pickup for guitar -- is it possible?

[sorveltaja]

Not really an optical pickup related, but more about how the external enclosure has turned out. Bit too dark and fuzzy photo, and drawn one:

 

I used chipboard to add some extra weight for it(otherwise I hate to work with that stuff). Kitchen aluminum foil was used for shielding to see, if it has any benefit for that noisy amp board.
The cover also has foil glued in it.

Cables from the guitar(pickups and ir-leds) will go through printed strain relief clamps. Later they will be replaced with two D9-, or a single D15 connector.

Using cd-rom ide-cable for outputs, as mentioned before, didn't work that well.
As a male part, I used pins soldered to perfboard, but when pressing/connecting the female part in, some pins popped out on the other side, with the copper spots they were soldered to.

Simple pin headers were used to replace them(as on the above pics).

That box has a 7809-regulator in it(as the testings were done with 9 volts from the variable power supply), and will be powered with 24V/1500mA power supply(from an old scanner).

Combined current drawn by the amp board, schmitt triggers, and ir-leds is only under 100mA, though.

By the way, when looking videos on youtube about hex fuzz, besides the commercial ones, there was one guy, that had done a diy-version(or so I assume) of it.
Of course I forgot to bookmark it, and now can't find it anymore.

He demonstrated it on several videos on his channel. A big let down, as on numerous other youtube videos also, is to show only the results, and to refuse/to not bother to reveal any of the details, of how it's done.

I mean, what's the point making those kinds of videos anyways.

[sorveltaja]

More of the interfacing hassle. Yes, it is (very)boring, but has to be done. Current version of the stuff:

 

Printed 'boxy' part on the guitar consist two sandwiched perfboards with pin headers, to plug in the sender- and receiver wires.
Another one is a strain relief clamp. They make it a lot easier to move the assembly around, as I have to do it often, to free some table space.

That wiring/connecting/interfacing is very finicky to work with. Apparently some of it needs to be changed at some point.

But for now, I'm going to test it as it is. I'll be surprised, if it works, as expected. If not, it's again a good chance to learn new project-concerning things.

[sorveltaja]

I managed to assemble the circuits inside the enclosure, and after stomping several bugs (which were actually stupid mistakes) out, they work, as expected.

Settings for the amp board/schmitt triggers/ir-leds are still same, that were used, when the circuits were on the guitar. As a sort of a reference.

The overall noise level, although not seemingly that high, is enough to reach the schmitt triggers. The receiver/phototransistor cables are shielded, except inside the small box on the guitar.

When covering that box with a hand, it decreased the noise just a bit, so that's where the 'leak' is. As a last attempt, I'm going to try to shield it with aluminum foil.

Whether it works or not, I kind of saw that coming(shielding problems/noises), as most, if not all, optical pickup manufacturers place the amplifying circuitry as close as possible to the source.

If only I had an ability to work with smd(surface mounted device) components, all the involved circuitry would be already 'stealth' enough to fit nicely inside/outside of the guitar.

After all, if the shielding of the 'box on the guitar' fails, it's probably time to redraw the involved circuits from the very beginning, as the current ones are very clumsy, and were just thrown together for testing purposes. Maybe even fusing them together to form a single circuit board, without unnecessary pin header -connections.

But time will tell, as Frankenstein would have said it.

[sorveltaja]

Shielding that small box in the guitar doesn't seem to have noticeable effect, to lessen the overall receiver/phototransistor noise.

Although the wiring for them is made of a shielded, individual microphone cables, there are differences between the pickups, in a way their noise outputs and -levels(with zero input signals) look.

Perhaps a shielded, single 6-core cable could be a better choice for that. But who knows, before testing. I'll probably add some of that to the next order.

After all, I really would like to keep the current setup, instead of redrawing/remaking the whole circuitry to fit in the guitar body.

So the next challenge is obvious; to lessen/attenuate the overall output noise levels with external circuitry, without knowing, if it's possible at all.

Schematics for simple noise gates are available on the net, and so far, I have tested one, that is based on the optoisolator(Led/Ldr pair). There might well be something in it, but it needs to be modified to suit for this project.

Probably not worth the trouble, but no matter what, I'll have to find out, how deep that rabbit hole is far I can get, before hitting the wall.

[sorveltaja]

While tinkering with the optoisolator based noise reduction for the amp board, I accidentally found the major source of the noises, that appeared, after the circuits were put to an external enclosure:
 

Those pots, that were added afterwards, seem to act like radio towers. Just by touching the metal wiper part, the noise immediately decreased a lot.
Although the noise source is located, I'm not quite sure yet, how to shield those potentiometers causing it.

But after all, I'm rather relieved, that the noise source wasn't that much about the cables, or other wirings on the guitar side.

The residual noise can be suppressed even more, by using something like the noise gate, that I've been testing.
At the moment, it's a bit quirky one, until I find a way to smoothen out the input signal driven led's response.

The noise gate circuit itself is quite simple, consisting one npn-transistor, that is fed by the signal from the amp board, to drive the led of the diy-optoisolator.

There are few other components involved also, but I'll have to do more testings to see, how they work.

[WeldingRod]

Capacitor to ground on the pot mounting nut...

Sent from my SM-G892A using Tapatalk

[sorveltaja]

WeldingRod, that's a good idea, but the pots are trimmers, and don't have mounting nuts in them. But on the other hand, it could be worth testing, if at first replacing one of them with actual pot, that has a metal cover and mounting nut, affects the noise level... hmm, we'll see.

The noise gate, that I've been tinkering with, is based on the circuit on the left(source: https://sound-au.com/project145.htm):

 

Although it is for switching, it could be abused used as a sort of 'automated volume pot', by substituting the Vctl input with an audio signal.
To drive the led, I've been using breadboaded circuit like this:

 

Component values on the above pic aren't necessarily same anymore, as I change them during testing, to see the differences.

So far, after plenty of adjustments, it seems to be almost usable. Drawback is, that it decreases the overall signal levels from the strings/pickups also.
To simplify its construction, instead of using each string to drive its own led, could be to drive all the leds with a single input from the traditional electromagnetic pickup.

Might seem odd, but actually there is no need create individual volume envelopes for each string/pickup, as the whole point is to use the device only for polyphonic(chords) purpose.

Monophonic stuff(like solos) could then be done by using plain old electromagnetic pickup, as needed.

To get back to the noise problem, quickest and dirtiest workaround is possibly to use a foot pedal/switch to control/kill the output, once the chords fade out.

I'm not sure, if I should even mention one of my mad mood ideas. But here it goes, anyways. The thing would be to use 4046-based phase locked loop(pll) -circuits for each string, to track the frequencies.

When the frequency is stable enough(like after plucking the string), the pll can lock to it, providing 'yes' -signal to a certain pin.

After that plucked note fades out, noise takes over. If the noise is random enough, the pll just wanders, and can't lock to it, providing 'no' -signal to that very same pin.

But in the end, I haven't yet tested, of how the externally installed circuitry sounds, with all the noises. Time to find that out.

[sorveltaja]

Time for an audio testing, and the result, although rough and noisy, wasn't as hopeless as I expected. Again, raw output from the schmitt triggers recorded to laptop:
 

testi_2.mp3.zip

All the six strings were plucked to form the chords/sounds, and to mask the noises. If partial chords were plucked, noises from 'idling' strings/pickups would mask them.

It seems, that every string/pickup needs their own, individual noise gates. That previous 'touching the amp boards trim pots with finger' -trick doesn't work anymore, as all the pickups are connected.
Infact, it makes the noise problem even worse.

As mentioned before, I'm willing to keep the current setup, as it's so much easier to move around, and to work with, as all the circuitry is in an external enclosure, and the cables between it and the guitar are secured with strain relief clamps.

So, on to the noise problem. I'm going to test an optoisolator-, and 4046(pll) -based ideas.

[sorveltaja]

After trying out some stuff about reducing the noise levels, I finally got this frustrated feeling in my stubborn head, that I'm beating a dead(or actually very noisy and farty) horse.

It's time to ditch the 386-based amp board, no matter what. When shortly testing two of the pickups with only the phototransistor amp-thingy(that are also in front of the 386-ic's), noise levels were already many times lower:

 

The above circuit will likely still be a part of the forthcoming circuitry. It seems to (probably) have enough output on its own, to feed the schmitt triggers. More testing with thinner strings is needed to confirm that, though.

On the other hand, the mechanical setup appears to work, as expected, apart from two of the thickest strings(e-a).
They are thick enough to almost cover the entire ir-beams from sender to receiver(which have 1,5mm diameter 'eyes'), providing erratic, octave-doubling effects.

Their optical components might need to be replaced with 3mm ones, but, as always, one thing at a time.

[sorveltaja]

When looking an alternative for the 386-based amps, I ended up testing a 4049(Hex Inverting Buffer and Converter)-based one (source http://electro-music.com/forum/topic-51275.html).
Although the schematic on that page uses two stages, I cropped the second one out, to see if the thing works with fewer parts:
 

After breadboarding, and some usual tests with thickest, and thinnest strings outputs, and with schmitt triggers, yes, it seems to be working. Not only that, but it's also a lot more responsive to gain adjustment, than the 386-amps were.

Nice thing about that 4049-chip is, that it has all the needed six 'amps' in a single 16-pin chip.

There is one serious precaution, though. I almost fried the chip(it got very hot), when forgetting to connect unused inputs to the ground(or to +, as the datasheet says).

As before, the noise wasn't much of a problem, when testing one string/pickup at a time... I'm not expecting wonders, as there is a big chance, that combined pickup outputs combine the noises also.

One possible way to minimise that, would be to mount the circuitry again, in the guitar. If that single-4049-chip-amp stuff works, the whole circuitry could then take a lot less room, than the previous 386-amp one did.

[sorveltaja]

The phototransistor- and 4049 amp board is ready for drilling:

 

Hopefully there isn't any errors, as I checked the layout multiple times. The size of the board is 72 x 68 mm. I could have made it smaller, but then I would have to (again)cram the parts in.

Besides that, I probably should have made it even bigger, as it's all analog circuitry. "The bigger and uglier it is, the better it sounds"

But seriously, as there are six 4049-based amplifiers so close together, their inputs and outputs should have some distance between them.

When doing the previous breadboard testing, they don't seem to do the hard clipping, as the 386-based amps did.

[sorveltaja]

New amp board:

 

I tested all the six sections, and they seem to work, as expected. As usual, square wave was fed to the ir-led of the pickup(from the earlier mechanical setup), and the phototransistor was connected to the circuits input:
 

 

There is one thing, that bothers me about that 4049-chip, though. While testing the sections one at a time(while others were not connected), it did hog constantly about 50mA(at 9V), and got warm.

Bit like on the breadboard, where I forgot to connect unused inputs to the ground, but not that bad.

From that came this shady logic conclusion, that the inputs might act like weathercocks, that aren't tied to anything during the storm. So they rotate at forever increasing speed, until they get glowing red, and therefore get very hot.

But enough of that. I took a look of other 4049-based circuits on the net(there are lots of them, as that chip was/is used in some commercial, and in many diy guitar fuzz pedals), and at least some of them have 1M resistor between inputs(that are in use) and ground.

Not a big deal, so I drilled holes for the 1M resistors, and soldered them to the circuit. But they didn't help at all. Am I missing something important here? 

I have tinkered with other cmos-ic's(mostly for guitar effect purpose, some counters, and pll) before, but never had problems with excess power consumption/heat.

Let it be said, that I read somewhere, that this chip(4049), when used as an amplifier, has a notorious reputation of having wildly different current consumption, even between individual chips from the same manufacturer.

I have two spare chips, and will probably test, how they compare.

But in the end, if the circuit works, regardless of current consumption and warming, it's okay by me, as I never intended to rely on batteries to use it. I'm more worried of how the chip lasts, being warm all the time it is used.

Also, one thing to test is to connect all the pickups to the circuit(to'occupy' all the inputs), to see if it makes difference.

 

[sorveltaja]

Audio test using the new amp board and schmitt triggers, again without any effects:
 

testi_3.mp3.zip

It seems, that changing the amp board was a step to the right direction. There is very little noise, when I tested it with the cables, and practically none of that noise is strong enough to get through the schmitt triggers.

Also less distortion(although some of it could add nice texture to the sound), as the 4049 doesn't clip the signal as hard as 386 amps did.

There is one downside, though. When plucking a single string, or playing solo, the outputs are very weak/lacking(or outright unpleasant). Don't know why that is. But when several strings are plucked(as chords), somehow the outputs 'morph' together, having a lot more powerful, and usable output.

What comes to the power consumption of the 4049 chip, I tested another one, and it acts just the same. I guess that's just the way it works in the current circuitry.
As the power supply has a lot more beef(24V/1500mA), than needed, I can live with it.

So next thing to do, is to fit the new circuitry to that external enclosure, that I made earlier, and some wiring.

[sorveltaja]

I've been eagerly waiting to get to the point, where the circuitry is worth putting inside the external box, and here we are:
 

No smoke or sparks, when I shortly tested all the six sections, they seem to be working just fine. Also no extra noises on the outputs.

That certainly doesn't mean, that the project is finished. It just means, that all the(well most of it) wire hassle is now out of the way.

Lots of aspects to test, including 'stealth' covers for the pickups.

Also to see, if I can make sense about that 'soloing'(playing single string) -problem mentioned earlier. There might be some strange phase cancelling stuff going on, that makes plucked single strings to sound so weak. When looking at individual string outputs on the scope, they are rectangular, and should sound, like fuzz does.

There is now less crosstalk between the strings/pickups/sections, when compared to 386-amps, but it still exists, and that alone could cause phase anomalies between the strings.
Almost like the next strings(that are muted) pickup(s) 'loot' most of the plucked single strings signal somehow.

Once I get into it, I'll make a diagram, which hopefully makes babbling above less fuzzier(pun intended).

After all, now it's good time to focus on testing/expanding/exploring the systems usability further.

[sorveltaja]

It seems, that the 'weak single string' -problem has something to do with the way I used to mix them together. So far, 10k resistor network like this, is what I've used:
 

I have a sneaky suspicion, that mixing sources, which have exact same output levels(like schmitt triggers in use) is the cause for that 'cancelling' -anomaly.
More proper audio mixing circuit is on my(ever growing) to-do list.

When the outputs are listened one by one(not mixed together), they sound a lot more solid, and the way they should.
So I guess it could be possible to use each output for individual purposes.

One example, of what I mean by that, is, that one(string's) output could be panned to the left of the stereo field, while other is panned to the right. 
Another example is to drive one output to echo effect, while other one goes to some kind of cmos-based device(like up- or down counter). 

Schmitt trigger circuit, that I use, has probably too low output levels to drive cmos-level devices, as it uses ~1,8V supply to lower the input thresholds. I haven't tested it yet, but I think that very mild amplification of the outputs should be enough to bring them back to cmos-level, if needed.

After all, there are six outputs to boot for what ever effect combinations. I could say, that it's like some form of an analog synthesiser, but naah, I'm not going to say it.


One thing, that might increase the sustain of the thinnest strings a bit further, would be to use Darlington phototransistors. But as they should be "side looker" -ones, there are only few types of those available.

Some of them are thicker(2,5mm), which makes them quite difficult to fit in, or require whole new mechanical setup, when compared to slimmer 1,5mm ones.

Anyways, it would be interesting to test them(1,5mm Darlington ones) with current electrical setup. Sourcing them seems to be bit difficult, though. But again, we'll see.
They aren't that necessary, but could add some 'kick' to the project.

[sorveltaja]

When testing breadboarded, simple single transistor amplifier to see, if the schmitt trigger levels could be increased that way to the cmos-level, there was suddenly this repeating noise pattern, where noise was on for 1 second, and 2 seconds off.

After a while, I was ready to open the external box, where the circuitry is, and to dig it all out. But no, luckily it wasn't necessary. When I removed the lid of the box, shortly after that, the noise disappeared.

After some head scratching, I noticed, that the 7809-regulator chip was rather hot. I checked its datasheet, and there it was: 'Thermal overload protection' caused repeating noises.
The whole circuitry draws 100mA, or less, so it was kind of surprising(at least for me), how hot it gets.

Now the regulator board hangs outside of the box, and all works just fine.

What comes to the transistor amp circuit, after some tinkering, it seems to provide enough gain/amplification. Form of the output isn't perfect rectangle/square wave, but as long as it 'triggers' external cmos-devices, it should be good to go. Next thing is to test that in practice.

Hopefully it's the last amplifier circuit needed, as there are few of them already in the main circuitry.



   

 

[BillTodd]

The whole circuitry draws 100mA, or less, so it was kind of surprising(at least for me), how hot it gets.

The dissipation is the product of  the the current (0.1A) and the voltage drop across the chip (~2V minimum for a 5v device so 7V input), so your device is trying to loose at least 0.2W of heat a TO-220 package will have a typical thermal resistance of 70'C/W  so expect at least +14'C over ambient without a heat-sink .

If your driving the regulator with more volts 9-12v expect  a lot more heat :-)

[sorveltaja]

Bill, no wonder then, that it gets that hot, as I'm driving it with 24V/1500mA supply. As the regulator chip is 9V 7809, 12 volt supply should be enough.

In fact, I used variable psu on earlier testings, and it was feeding the circuitry at 10 or 12 volts, and the reg. chip stayed a lot cooler.

Thanks for pointing that out .

[sorveltaja]

Now, that the transistor amp circuit is built, the output levels should be good to go, to feed other cmos-based devices, or effects.

It brings the schmitt trigger outputs(about 2-3 V p-p(peak-to-peak)) nicely to ~9 volt p-p output at 9 volt supply. Very quick and dirty, and the resistor values were chosen simply by tinkering first with potentiometers, to find suitable output. Capacitor values were chosen by looking on the scope(and what I had on my shelf), which made the output to remain closest to square/rectangular wave. Pretty scientific approach, eh?

But after all, the frequency range of the strings/outputs is below 1Khz(unless one insist on shredding on the thinnest string on the frets above 12th one).

Only one section is shown, but as the circuits before, it also has six of them:

 

 

One, or few other outputs could then be clipped down with something like 5V zener diodes, if needed. Don't know about that yet, as I'm going to take a bit of a break from the guitar/pickups -involved circuitry, and take a look, what kind of cmos-devices I have, that could then be driven with those pickups/outputs.

So far, I have fiddled with 4046 pll(phase locked loop) -ic, just by feeding it with function generator, and listening to the output.

I used to tinker with that chip years ago, so it isn't entirely new thing for me.

There are so many different component combinations/options to alter, of how it works/sounds. This is only one variation of them, that tracks the frequency of the signal fed to it:

 

That chip has a vco(voltage controlled oscillator) in it, that runs at tracked input frequency, which could be interrupted, making some kind of "sample and hold" -situation, where the played note could be freezed/sustained to play longer, than the guitar string vibrates. Almost like on keyboard, that plays the note as long as the key is pressed.

The idea for that came from: https://hackaday.com/2015/08/07/logic-noise-4046-voltage-controlled-oscillator-part-one/

There is a lot of theory based stuff, of what that chip does, how it does it, and so on, for those interested, but I just don't understand it.

But that chip(basic version CD4046, not sure about the exotic, high frequency variations) is a very good one for testing on breadboard, as it is practically impossible to fry it out, unless intended.

[sorveltaja]

It seems like I'm getting more and more drawn into testing, how different cmos-based contraptions sound. So far using breadboard, and function generator, and a small amp for listening.
I feel tempted to test with guitar/pickups, but the concepts/results are somehow easier to follow with simpler test setup.

This is going to be rather long off-topic babbling, and I should perhaps start another thread, considerig the 'after-effects'.

I was looking info about sawtooth generators on the net, as I've read, that they have a lot more harmonic content, than plain square waves do. But seemingly, making decent sawtooth generator isn't necessarily that simple.

I have, as long as I can remember, liked Hammond organ sound, and Moog(analog synth) also. I'm not actually looking for replicating those sounds, but to see, what could be achieved using simple circuits.

What comes to that sawtooth generator, I thought that 'does it really have to be in traditional sawtooth form, that was used on so many analog keyboards?'.

Then came the idea of some kind of primitive "sampling" technique. After more searching, I found a name for it - staircase generator.
There are schematics, of how to make a proper one, but again, I started looking for other(simpler) ways to test the idea.

Back to the shelf, to check, what chips, or other stuff I have at the moment. There was 4040(12-stage binary ripple counter), which was then test subject.
I cheated a bit, and did some simulations, to see, if it's worth breadbording.

Circuit itself:
 

The output waveform is not, what I expected, but nevertheless, it sounds like a cheapo analog synth. Good enough, so I breadboaded it, and yes, it sounds very much the same, as in the simulation.

One big drawback is, that the frequency of the input signal is divided down a lot. To get the output frequency to match the input one, I did some testings with 4046(phase locked loop), that could be used to multiply the frequency by the factor, that it's divided by. 

I'm not sure yet, if it works in practice. More testings ahead.

[Yarrrrr]

I stumbled upon this thread as I have been researching optical pickups myself and came to the same conclusion that there is barely any information about it. But what I was looking for was more of a drop in replacement for a normal humbucker pickup which led me to the oPik design.

I've got a little test setup for the thickest string: 

Oscilloscope output: 

Audio output: https://www.dropbox.com/s/dug6chet2pw6jv7/test%20%233.mp3?dl=0

I am currently working on building a shielded enclosure to improve the signal to noise ratio, and trying to make this all fit in the size and shape of a normal pickup.

Something like this: 

[sorveltaja]

Yarrrrr, welcome to the forum. It would be interesting to see your take on an optical pickup. If memory serves, there was some pictures/details in the oPik's patent, like an approximate angle between sender and receiver.

I remember using that patent's pictures as a basis, when I tested, how the infrared light reflects from the string(s), and what kind of output signal it produces. Obviously, when moving to thinnest strings, the output level decreases also. So yes, as good as possible shielding is needed.   

[Boop]

Hi,
Joined the group for this post!

Hopefully this thread isn’t dead, maybe this can lead to its resurrection.

Played with this a little bit some years ago.  Didn’t go much further than hooking up a photo diode and an IR led running into a guitar amp directly, or maybe it was a mixer?  Got output, was very excited, then life “got in the way”. 
But I thought I’d put these here from old Make magazine articles and associated links:
https://makezine.com/projects/infrared-string-bass/
https://www.slideshare.net/engrmali3/infrared-string-bass
There are more links that I’ll have to find, but this is what I have time to give right now.

 

[Yarrrrr]

Hi,
Joined the group for this post!

Hopefully this thread isn’t dead, maybe this can lead to its resurrection.

I got to the point where I have a working prototype that I am dissatisfied with. But with my lack of experience in electronics and the amount of effort I put into trying to improve it which amounted to nothing has me quite discouraged.

More details with pictures and audio examples over here: https://www.edaboard.com/threads/guitar-photodiode-amplifier-signal-to-noise-ratio.396736/post-1705241

[sorveltaja]

Hi,
Joined the group for this post!

Hopefully this thread isn’t dead, maybe this can lead to its resurrection.

Played with this a little bit some years ago.  Didn’t go much further than hooking up a photo diode and an IR led running into a guitar amp directly, or maybe it was a mixer?  Got output, was very excited, then life “got in the way”. 
But I thought I’d put these here from old Make magazine articles and associated links:
https://makezine.com/projects/infrared-string-bass/
https://www.slideshare.net/engrmali3/infrared-string-bass
There are more links that I’ll have to find, but this is what I have time to give right now.

 

Welcome to the forum, Boop. Yeah, it's rather good start to test a setup for an audible output. That's how I did. I don't remember how the progress went, but it's documented in this thread, stage by stage, of how I got the strong hexaphonic fuzz sound that I was after for a long time. Of course that isn't everyone's cup of tea, but hopefully in this thread are some things to consider, when one wishes to experiment with optical pickups. Remember that this project is based on a lot of trial and error testing. When something worked, I used it, and if not, ditched it.