Diy optical pickup for guitar -- is it possible?

[sorveltaja]

Ade and Pekka, thanks for the replies. Yes, when using the laser printer for the transfers, there seems to be almost always some fine pitting/holes on the traces. But as they are quite randomly distributed across the traces, they should in most cases still be usable.

To avoid, or prevent the pitting/holes on the traces, there might be one (although unsolved) resolution for that:
After the transfer is transferred to the shiny copper surface, the traces with tiny holes and pittings should be "impregnated" with some (so far)unknown substance.

What I mean by that, is that the transferred traces with tiny holes/pittings are like an exaggerated sponges.
To fill all the holes and pittings, likely some stuff in liquid form, that is able to be sucked in to all those tiny cavities, while not sticking to the smooth copper surface, could be used, and let cure, before etching.

What that stuff could be, so far I have no idea. Not sure, if all that makes sense at all, maybe it was just a brain hickup.

But in the mean time, I tested one more thing with the current mechanical setup. This time the output is from the schmitt triggers. Some equalisation to brighten the sound, and a hint of reverb was added:
http://www.elisanet.fi/kz1706/optical_pup/schmitt_trigger_test_.mp3

Those 4093-based schmitt triggers seem to work really well, cutting all the noises out. Major problem is the mechanical setup(as it relies on the reflection from the strings), which requires the ir-leds to be driven hard enough to make the trim pots quite warm.

For now, I'll put that mech setup aside, and get back to the earlier one, to see how it compares.

[sorveltaja]

When that previous mech setup was in the guitar, at the time I didn't think of the thing, that could extend its usability. But now it seems obvious: simply add reflective cover above the strings, that could be easily removed at will, to allow 'normal' access to the strings like this:
  [ You are not allowed to view attachments ]  [ You are not allowed to view attachments ]  

It should eliminate the effect of the waving hand above the pickups. Also, while testing that setup, there was momentary, audible peaks on the output, when the plucking hand was resting over the pickups.

Again another factor, which I didn't consider, while concentrating on other things with that setup. But when thinking of it, if there was a smooth, reflective surface above the string, it could add more reflections from that very string, and somehow improve the overall infrared 'flux' from the sender to the receiver, in form of increased output with lesser amperage used by ir-leds.

All that is just an assumption so far, but it seems that I have to revisit that setup to find out.

That setup uses 120 degrees between the sender/receiver, making it bit too wide/large. The narrower 90-degree setup could have similar output. But as always, plenty of testings ahead.

[philf]

......... But when thinking of it, if there was a smooth, reflective surface above the string, it could add more reflections from that very string, and somehow improve the overall infrared 'flux' from the sender to the receiver, in form of increased output with lesser amperage used by ir-leds. ..........

sorveltaja,

If you're going to try this wouldn't it be better to include a dividing wall between each string to prevent (or at least reduce any crosstalk)?

Phil

[sorveltaja]

Phil, that's a very good point. After a bit of sketching, the 90 degree setup, and then same with a cover:

  [ You are not allowed to view attachments ]  [ You are not allowed to view attachments ]  

I added 'legs' to one side, which rest on the base, so that the cover doesn't flex when the plucking hand rests on it. I might add legs to the other side as well.

Although on the picture, the surfaces, that face the strings, are a little bit curved, they probably should be flat, to get straighter reflections.
Do the flat, reflective surfaces still need some kind of fine adjustment option for alignment, can't tell yet.

What comes to the reflective surfaces, I guess they should ideally be mirrors. But that is really difficult to achieve, unless one has a very good skill in glass cutting/forming.
Instead of that, well polished, flat metal surfaces could do the trick. Maybe even shiny enough paint, or aluminum foil.

On the other hand, infrared light could have other reflecting properties, that may not be so obvious, when compared to the visible light. Haven't go into that too much yet, but the output differences should be measurable, when testing materials for reflection with ir-led/phototransistor(or photodiode) -pair. I'm looking forward to test it.

But before that, I have the older kind of mech setup in the guitar, so I'll test, how it works. Today I got the ordered missing ic's for the amp circuit, and if all goes well, it should be possible to test, how all the six strings sound together.

For 'mixing' the string outputs together, I simply used 'passive' one with 10k resistors for each output, when recording the previous audio clip. Schmitt triggers were used, and as their output levels are equally 'loud', not really need for adjustment between the outputs. Basically It's like three(later six) independent square -or rectangle wave oscillators.. well, oscillating together.

[sorveltaja]

After thinking of testing the rest of the older setup, mentioned in previous post, which was similar to this:

  [ You are not allowed to view attachments ]  

-- I already know, that it should work to some extent so far, and is something to get back to, if other mech/optical configurations fail.

I got curious, of how the reflections could possibly improve the infrared 'flux' of the 90 degrees -kind of setups, which otherwise require the senders(ir-leds) to be driven at higher amperage, to get usable output, if relying only for the reflections from the string.

So I sketched a simple setup to test, how different materials reflect that ir-light. It has a 5 x 10 mm 'window', where to put different materials. That setup is on the table, instead of guitar:
  [ You are not allowed to view attachments ]    [ You are not allowed to view attachments ]  

As a reference, I use a special, 8mm wide mirror from the old scanner. What makes it 'special', is that on the other side it is like normal one, that has a glass between the reflecting surface, and looker.
On the other side, no glass, only dead flat mirror surface. I tried to take pictures of it, but the camera sees it as a black/dark object.

But, after all, that kind of mirror isn't required/needed for the testing. Plain shiny aluminum foil gave about the same results, when the sender was fed with 400Hz square wave.

That was after rather quick testing, but the infrared 'flux' could be improved that way.

How all that stands, when that setup is in the guitar, with the actual vibrating string added to the ir-beam, I can't tell until more testing.   

[sorveltaja]

The new 90 degree mech/optical setup is ready for testing, for the thinnest string(0,25mm). For the reflective material, I glued aluminum foil to thin cardboard, so that it's easier to handle, and to cut to suitable size.
 
Then it was tested with the previously mentioned test-setup. Again, the input ---> output -results were close to mirror-like surface. But when looking that aluminum foil, it doesn't seem that 'mirrorish', though.

It could be carefully polished, but I'm thinking of replacing it with thicker brass, or aluminum, which should be a lot easier to polish, although they should be sawed and filed to the size.
Bit of work, but not a big deal.
  [ You are not allowed to view attachments ]  

The cover with alu foil glued in. There were thinner legs on the front side, but I managed to snap couple of them out, while sanding the upper surface(it was printed upside down).
But the overall construction feels rigid enough without them, so I snipped rest of them away also:
  [ You are not allowed to view attachments ]  

I certainly hope, that this setup, even with cover in place, works, as it feels somehow very 'natural', without the feeling, that something is protruding between the strings:

  [ You are not allowed to view attachments ]  


After some testing with the thinnest string(without schmitt trigger), there was output, but it wasn't easy to locate the factors involved(as I have a nasty habit of starting at the hardest point available ).

But after all, it looks promising enough for this setup; so the next step will be the thickest string, that has a lot more output to boot, to hopefully get a bigger picture, of how that setup works. 





 

[sorveltaja]

The cover is now ready for testing all the strings:

  [ You are not allowed to view attachments ]  

Aluminum foil was replaced with 0,25mm brass. The brass sheet was first treated with dremel and polishing compound, and after that, the pieces were sawed with a fret saw to size.
Bit of filing to remove any burrs, and that's it.

There is one thing though, when handling polished brass like that. To avoid any fingerprints on that shiny surface, straight after polishing, I covered it with a tape.
After sawing and filing, it was peeled off, before gluing the pieces in place.

Although the brass surfaces on the above picture might look like mirrors, they are actually quite 'shady'. But, after all, if aluminum foil performs well enough, polished, more solid metal surface should do even better. So fortunately, no need to go crazy with the polishing.

With this mech/optical setup, there are a lot of options to test. One being like this, where the distance, and angle between sender/receiver could be altered, while still using that very same 'base platform':

  [ You are not allowed to view attachments ]  

At this point, when the amplifier circuit is completed(although not perfect, but good enough for now), as well as the schmitt trigger one, a very basic kind of 'foundation' finally exists.

With that in mind, it allows one to dive more into other factors of testing, like producing a stable signal from the string(s), that could be a lot more easier to measure, than simply plucking the strings.

One of the reader mentioned earlier something, that reminded me of a 'sustainer', that is used to input the string(s) signals, and fed back to the speaker-like 'exciter'(under the strings), to make them 'sustain', or ring constantly.

I'm considering of making that kind of sustainer for a single string use, that could be fed with open string's fundamental frequency with a function generator.

One heck of a side project, but the concept isn't actually new for me.

[sorveltaja]

It seems, that the 'resonator' to make the string(s) constantly ringing, while adjusting/measuring output of the pickups, doesn't need to be that complicated, as I expected.

Using small 8ohm speaker and function generator, I tested, what could be the position, to get the strings to respond to speaker signal.

So far, at end of the neck, where the guitar's headstock usually is, turned out to be the best location. My guitar is a decapitated version, so it doesn't have a headstock, but I managed to tape the speaker so, that it doesn't touch the strings:
  [ You are not allowed to view attachments ]  

That way the speaker makes the neck and body resonating, and therefore, also the strings. I guess, that it would help to have a headstock, with a mass in it to make the neck resonate even more.
Even that small speaker works, although a bit bigger one could help with higher frequencies, that the thinnest strings use.

I have another small 8 ohm speaker, and maybe I'll add it in series or parallel with the one already in use, to hopefully get some extra resonating power.

As one could expect, thicker strings resonate/vibrate a lot easier, than the thinner ones, which require the speaker(s) to be driven harder.

I'm not sure, if driving the speaker(s) straight from the function generator is a wise thing to do, for a long time period. To play it safe, I'm going to breadboard one of the 386- based amps to drive them:

  [ You are not allowed to view attachments ]  

Some things, of how to get the strings to respond to the fed signal: the signal itself should be in sine wave -form, as the wooden parts of the guitar do not seem to respond to (somewhat unnatural) square/rectangular waves that well.

Also, the signal feeding the speaker(s) should be 'tuned' for each string's fundamental frequency, to get them to resonate. That is, when testing one string at a time.

Although the setup is fairly simple, a sine wave generator, that has an option for fine tuning for the frequency, is highly recommended.



 


 

[sorveltaja]

I tested the x200 gain configuration of the 386-amp, to drive the speaker, and it clips and distorts the fed sine wave, even at very low levels, making it useless for the purpose.
So next is going to be the mildest, x20 version.

I have also looked a way to improve the way, of how the vibration is transferred to the guitar body. There is a thing called 'vibrating speaker', which could be made using a sacrificial speaker.
As I have one, that's what I did:

  [ You are not allowed to view attachments ]  

I did cut the cone on the edges, and hot-glued random, printed left-over t-piece to the center of it. As the piece, that holds the strings, is made of aluminum, I used superglue to attach the assembly to it. Removal is easy, just knock it out.

It works bit better now. I have only three thickest string's pickups in the guitar so far(need to prepare rest of them also), but for them, the signal is more than enough. One of those strings, when tuned in, actually starts to buzz like sitar string, and the signal, that is fed, must be turned down, to get cleaner output. All that by abusing using only the function generator to feed the speaker.

As always, plenty of testings ahead, before usable end-results.
 

[sorveltaja]

Some observations, while testing. First the overall base setup for all the strings:
  [ You are not allowed to view attachments ]  

To avoid confusion about the guitar strings(with a standard tuning), they go from thickest to thinnest; e-a-d-g-b-e. Basically just a bunch of different gauge steel/nylon wires, in certain order.

But so far, some results of the three tested setups; first one seems to work on four of the thickest strings(e-a-d-g). That setup doesn't give strong enough output for the second thinnest string(b), though.

Second one is a bit of a mystery, as instead of increasing the output, when cover(with reflective surfaces facing the strings) is on, the output disappears(b-string).

Third one, again tested on a b-string, seems to be a step to the right direction, as it provides more output, when using the cover.
  [ You are not allowed to view attachments ]  

Distances/angles between the optical components on those setups are only arbitrary, but I think still a very good way to get a hunch, of how they behave, and affect the overall output.

[sorveltaja]

Addition to the previous testing theme; but this time it was done only with the hardest one - thinnest string(another e).
Angle between the optical components was changed, while the distance changes only a bit.

First one is a no-go, but gives good a tip, of what (or what not)to test next.

With the second one, I was aiming for an approximate angle, that would allow the ir-beam to actually reflect straight from the cover's brass 'mirror'. Afterwards it seems obvious, that the results were a lot better:

  [ You are not allowed to view attachments ]  

Needless to say, but although the four thickest strings(e-a-d-g) already give usable output with one of the earlier setup, that 70-degree setup will probably replace them, to squeeze even more output from them.

[sorveltaja]

The two thinnest strings(b-e) with the 70-degree setups were tested. There might be just enough output to keep their notes 'sustaining', to make the schmitt trigger's outputs usable.
That could be judged better just by plucking the strings, and listening the overall output.

As they are so thin, adjusting them for the sweet spots is rather nitpicking thing to do, though. If one doesn't mind the string material, nylon strings, that are generally a lot thicker, could be used instead. Personally I don't like the 'rubbery' feeling of the nylon strings, but who knows, I may well end up adding a set of them, when ordering the strings next time.
 
One thing was again confirmed with this kind of setup: the signal disappears immediately, when the string is bent. There could be a way to fix that, perhaps by using something like two pairs of senders/receivers per string, where the pairs are wired in series or parallel to form again one pair per string.

When reading back the thread, JohnHaine suggested diffuser. I had to look out for the definition of the word, and as I see it now, the light is 'scattered', making it to glow, instead of straight lighting, when it goes through the (shady?) surface.

I think I know at least one way(found accidentally at the beginning of the project) of making the ir-led's plastic surface to look not-so shiny(more like shady); simply expose it to superglue fumes, and that's it(maybe even dipping the led to acetone or xylene instead could work). 

Although I think it would greatly decrease the transfer of the ir-beam, it should be easy enough to test, how it works.

But in general, the project(while still having plenty of room for improvements) is slowly getting to the point, where it's time to make a simple circuit(6-to-1 -mixer), for the numerous 'audio-only' testings.

If that goes well, then there is yet another challenge: to get all the circuitry out of the guitar's body, to form a separate control unit.

[sorveltaja]

The whole setup, with almost all wires connected:  [ You are not allowed to view attachments ]  

As for the 6-to1 -mixer for the output of schmitt triggers, I built a stamp-size one, with discrete 10k resistors. But, memory served, and I found a bag of 10 x 10k resistor networks from the shelf.

It was quite easy to make a nice and tiny 'plug-in' unit out of that. It's barely visible in the picture, but it's the orange one, on the upper boards right side, that has schmitt trigger -ic's in it.

Makes me a bit nervous to start testing out a mess like that. I don't expect instant success, though, as there could be hiding some 'cumulative' bugs.

On the other hand, with (almost) all the wires running wild connected, it looks more complicated, than it actually is.

If(when) there is any kind of success, only then it's worth diving into the 'sub-project', to tidy up(or remake) the wirings for something like.. more permanent setup.
 


 

[sorveltaja]

After two days of testing, the output of the thinnest strings still lack usable output. Also there appeared a strange 'octave skipping' -effects on thicker strings, which should have enough output.

For that I suspect the not-so-optimal amplifier board. There is quite a lot of crosstalk between the strings, not only from the next string, but from all other strings. Meaning, that when observing only the thinnest string's output, others bleeded through as well.

The schmitt trigger -board was removed shortly after starting the testings, as it didn't make sense to use it with the amp circuit, that already produced not so good outputs.

So, back to drawing board. For the optical setup, I'm going to re-visit the one, where the sender/receiver 'stare at each other's eyes'. On the left is the older one, that should still work, and on the right is a possible new alternative.

I have some optoswitches, that have those kind of flat components inside. Should be an easy task to extract them from the optoswitch body, and test, if they are usable. When on the level with the string, they protrude only about 1,2mm above that string:

  [ You are not allowed to view attachments ]  

The housing will be sort of 'semi-open', where the component pairs are mechanically connected 'under-the string', to hopefully make them easier to install/remove. They need to be connected together, or the height adjustment becomes (not) very interesting. To prevent ambient, or plucking hand -effects for the ir-beams, again, a slim cover over the strings could be used.

One might wonder, why not to use the optoswitches, as they are? For a bass(at least for 4-string), that has more room between the strings, that could probably work. Otherwise they are too bulky, and the ones I have, the optical components have some play inside their housing, so not very good for precise height adjustment(as they should be secured with glue to stay in place).

As an opposite, one supplier on my list has smd(surface mount device) version of those available, but they are far too tiny for my paws/eyes.

I'm not sure yet, of what the amp board is going to be, though. 







 

[sorveltaja]

Sketch for a new setup for the optoswitch components:
  [ You are not allowed to view attachments ]  

Model of the optoswitch is H21A2 by Isocom. Components came out easily, by pulling the 'lid' open with a sharp chisel, without even breaking their housing:

  [ You are not allowed to view attachments ]  

Sender and receiver were push-fitted in place(perhaps some glue also needs to be used). One with yellow dot is the ir-led, while the one with red dot is a phototransistor. Printed part on the left is a testpiece of the to-be-printed base, to test the tolerances and height adjustment. So far it has gone well, and (hopefully)shouldn't take too long to get back to 'test-bench'.

It may need some modifications to fit the wirings, though. The concept will probably be same, as with previous setup, where the sender's wires come out from one side, and receiver's wires from the other side.

The components work, as expected. The physical specs of those flat components aren't usually mentioned in the optoswitches datasheets.
Although they can be simply measured, in some cases it could be better to see the actual manufacturers spefifications, and tolerances.

Sparkfun has datasheets of Liteon's LTR-301 phototransistor, and LTE-302 photodiode available on their site, as well as downloadable 3d-models, if memory serves. But generally all of those seem to be the same size, after all.







 

[sorveltaja]

New setup almost ready for testing:
  [ You are not allowed to view attachments ]   [ You are not allowed to view attachments ]  

Once the short-ish legs of the components are bent, they fit in just fine. That way, the wires that are to be soldered, stay outside of the base.

By using the smallest size heat shrinking tube, that I have, the legs shouldn't short between each other. But we'll see.

While sketching and printing the parts, I remembered seeing a picture of those flat components used in optical pickup somewhere. So that's where I got the idea.

It was Ron Hoag's invention. I have no idea, of how old that picture is, but if judging it by the thickness of the optical components, they look a lot thicker, so probably not very recent:

  [ You are not allowed to view attachments ]  

Again, making an optical pickup to look that clean, without wires running around, there really isn't too much options, other than to drill holes for the wires to go to the back side of the guitar.

After all, there is very little information available of that pickup/setup. I have found only few sources, one which is on the same site, as the above picture:
https://news.softpedia.com/news/Light-Music-Technology-On-Sale-The-Optical-Pickup-Patent-52851.shtml

If only he had 3d-printer (or affordable cnc) technology available back in the days, whole electric(and why not bass and acoustic also) guitar scene could be a whole lot different, than what it is today.

Generally, as mentioned before, this project doesn't follow the same path, that commercially available products, which concentrate on removing magnetic pull, created by traditional electromagnetic pickups, to get the cleanest possible output/sustain from the strings.

Bit of (ranty, sorry about that)background; for a long time, I've been a fan of the analog syntesiser sounds. But, not a surprise, that to get those, one needs to be able to play keyboards(which could make whole thing much more simpler).

Although I have had some keyboards in the past, they ended up gathering dust and sold, as they don't feel as natural to produce sounds(pluck), as guitar does(personal preference).

Other options available today, in form of commercial products:

- Get the costly 'authentic' old analog guitar synth with a hexaphonic pickup, that had already tracking problems, when it left the factory.
- Get the digital guitar synth with a hexaphonic pickup, that uses the newest technology, to emulate analog synth(could do the trick, if you got the money).

It seems, that working, true analog guitar synths haven't developed at all in the past decades, to make them easier to use, as there hasn't been enough demand for the big manufacturers, that have plenty of engineers, and other resources, to develop them further.

I'm well aware, that those kind of devices require special playing techniques to suit the technology. 

But no matter how good/bad it is, that analog guitar synth -stuff seems to be very marginal these days. Maybe it was always that way.

End of the rant, and back to the project.

 

[sorveltaja]

New setup was tested with three of the thinnest strings (g-b-e):

  [ You are not allowed to view attachments ]  

Outputs from the strings are now bit closer to be usable, than with previous setups. When they are mixed together after Schmitt triggers, chord harmonies are more audible, than on the lower frequencies (thicker strings e-a-d).

There is still that ugly sputtering, octave-skipping effect. One reason for that could be 'double clipping', where the amp board clips the signals first, and after that, the schmitt triggers clip it again.

On the other hand, actual analog cmos-based 'octaver' -device(divides the input frequency by 2, like Roctave divider)for the guitar, has a very similar behaviour, but it appears only when the signal fades out, or the input signal is too weak.

To get around that, perhaps an amplifier stage, that doesn't clip so easily, while producing huge amount of output, should be used. Have to find one first, that suits for the purpose, though.

There is also one stage, that is a part of the currently used amp board, which is placed before the 386-amp. It boosts the phototransistor output. Not sure, if it's an amplifier, or something else.
Generally phototransistors seem to have max. collector-emitter voltage, that is at least 30 volts. So far, I haven't explored it that much, and have used it only with 9 volts(like the rest of the amp board):

  [ You are not allowed to view attachments ]  

At this point, one might think, 'where's the beef? let's hear the results!'.

I was merrily recording an audio clip, using trusty Windows 7- based laptop, that has an ancient version of Goldwave installed. I know, but it just works for simple tasks(or so I thought).
As usual, I listened the output with headphones, while recording. All seemed/sounded fine.

But when listening the resulting, saved audio file, it had a strong, phase shifter-like effect going through the zero-point, that 'cancelled' most of the audio, making it useless.
Previous audio clip was recorded using exact same setup, with no problem.

I have also a desktop pc, and another laptop to test, so maybe i'll manage to record a new audio clip, that is worth posting.

   
   

[sorveltaja]

In the past few days I've looked a way to replace the 386-amp board with an op-amp based one. Test circuit(TL072):

  [ You are not allowed to view attachments ]  

It was just thrown together for breadboard, and probably lacks all of the some important components. But as it is, it gives enough output for testing.
Seems to work better by using a split power supply.

But in general, the lack of delivering enough output to keep the schmitt triggers 'squaring' long enough, to make the output to feel more like guitar-ish, still exists.

Then I had to rethink about the whole thing: I have already an overly powerful(and noisy) 386-based amplifier board, which should be more than enough for the purpose.
But in practice, is that much grunt for amplifying the phototransistor's output really even needed, as the end result is still far from usable?

So it was time to try a bit different approach to the problem. I remembered seeing something like positive- and negative input threshold voltages on the 4093's(quad schmitt trigger ic's, that are used so far in this project ) datasheet.

This picture is actually from 40106's(cmos hex schmitt triggers) datasheet, as it's much more clearer, than the old 4093's one:
  [ You are not allowed to view attachments ]  

One thing, that came to my mind, was to bring the input signal between the Vp and Vn, by using an offset, and then adjusting it, to see how low it would go to 'tickle'(trigger) enough both limits.

There might be something in it, but I'll have to admit, that after trying that with a function generator and a scope, there is much more in it.
No matter, which way I tried to test that out, I just couldn't get my head around, of how it could actually be done.

Threshold voltages, although only approximate ones, are on the datasheet, and as the supply voltage is lowered, they lower also.

So the next thing to test was obvious; decrease the schmitt trigger's supply voltage. At 1,8 volts, the input threshold was under 0,1 volts, when testing with a square wave.
Needless to say perhaps, but at that level, the quality of the output never meets any of the standards.

Besides of that, I kept on going, and tested the bugger with a g-string. Yes, the output was a lot lower, but for some reason, 'squaring' was there, and it just kept on doing it longer, than ever before.
And that was by using the above 072-op-amp based circuit.

So far, cons of using that low supply voltage:

- output waveform has some mutations(although not necessarily lethal), when compared to clean square wave
- needs to be amplified/gated afterwards, to get back to the cmos-level

Then the pros:

- overall bandwidth of the schmitt trigger chip decreases drastically(down to few kilohertz), which should/could be good enough to reject radio-like interferences
- less amplification is needed

There is a drawback, when using optical components, that are extracted from optoswitch(or at least the ones I have), though. The phototransistor doesn't have an ir-filtering in it, so it is easily affected by the visible lighting.




 

[sorveltaja]

I decided to take a final test for that noisy 386-based amp board, to see if it still has any use, with 1,8V supply fed schmitt triggers.
It appears to have, once its output levels are turned down enough.

Testing was done with three thinnest strings(g-b-e). There was a strong (4-5V) ~25kHz noise signal, that made the schmitt triggers go nuts.
In previous post I claimed, that the bandwidth of the schmitt triggers had decreased with lower supply voltage. Maybe it has, but not down to a few kHz.

Still I wanted to see, if that noise level could be lowered. By using an online low-pass filter calculator http://www.learningaboutelectronics.com/Articles/Low-pass-filter-calculator.php#answer1, I ended up making a quick test-plugin, that has 10k resistor and 1,5nF capacitor in it:

  [ You are not allowed to view attachments ]  

Its cut-off frequency is at ~10kHz, and seemed to do the trick. After that filter, the noise level went well under 1 volt.

After some tinkering, the strings' outputs had quite decent rectangle shapes. Although the scope showed also more acceptable sustain of the plucked strings, it's only visual result.

I hate to go ahead of things, but if the results are still there, when listening to the outputs, the project has taken a tiny step to the intended direction.

But after all, if that goes well, the current optical/mechanical setup still needs a removable (as slim as possible)cover above the strings to be drawn and printed.
And the schmitt trigger board needs its own dedicated regulator, likely LM317-based, variable one. I have those already at hand.

What comes to audio recording setup, especially, after failing with the previous audio clip, I feel like 'meh, what else is not working, as it should'. Eventually that needs to be sorted out, though. 

[sorveltaja]

Latest setup ready for height adjustments of the pickups:

  [ You are not allowed to view attachments ]  

To get reasonably stable outputs from the strings, for adjusting the height of the pickups for max. outputs, perhaps most time consuming part (at least on my guitar) is to track the peak frequencies, where the individual strings resonate, when using the small speaker as a resonator(mentioned earlier).

Using low enough outputs, that don't clip at the amp stage, makes it easier to find sweet spots.

I used to write those frequencies down, but they change from day to day. Not so much between the strings, but all of them have gone in frequencies tiny bit lower.
I guess, that's the way the wooden construction(mahogany body, maple neck) reacts to changes in air humidity.

But enough of that. Only after the height adjustments, it's worth to move on to the electrical side.

One thing, that popped to my mind, when testing, if it's possible to 'palm-mute' the strings, while the cover for the optical pickups is in place. 

It is almost possible, but as the printed cover is made of plastic, it has to have some extra thickness in it. If that cover was made of, say, out of 0,5 or 1mm brass sheet, it should allow a lot lower, 'slimmer' cover to be used.

What I mean by that, is that if one uses a traditional electromagnetic pickup, its usability could be expanded, when using it besides the optical pickup.

Making that kind of brass/other metal cover would require some serious skills, that I don't have, though.

On the other hand, one way could be making a mold, and casting some tin in it. Could be too smoky thing to do in my apartment. But, after all, who knows

[sorveltaja]

This time, the audio recording of the 6-to-1 mixed schmitt triggers was done straight to the laptops mic input. So no echo, equalisation or any other effects were used. Plain 'dry' output:
(zipped mp3 attached)

The schmitt triggers' output was bit too hot for the laptops mic input, so I did simply put a potentiometer between them, to the breadboard to turn down the excess distortion. That breadboard was on my lap, while plucking the strings, and kept on falling all the time..

But after all, I was just too impatient to test the audio output, to solder some wires for more usable/permanent level adjusting unit. 

When doing the previous, failed recording, when there was that weirdo cancelling effect, I used Zoom G3(guitar effects device) between the schmitt triggers and laptop.
I'll have to admit, that for a long time, I have been interested in exploring odd and strange audio effects. But that one, I can't even imagine, how to track it, or make sense, what it actually is.

Besides all that, came another idea of using individual covers for each string/pickup, like this:

  [ You are not allowed to view attachments ]  ]

I haven't tested that yet, but if it works, the parts could be printed, while being a lot more 'stealth', than single, bulky cover.

[sorveltaja]

At this point, it's (finally) time to move on to making the electronic circuitry(amp, schmitt trigger, and ir-led adjustment -boards ) in more easily form to operate. That means basically removing them from the guitar body.

As the guitar is to be used only at the living room, for starters, one meter long cables should do. The plan is to use 7 for the ir-leds(could possibly be made out of an ethernet cable), and 6 for the senders/receivers(using shielded microphone cable).

I already tested a single pickup with a single, one meter mic cable, and so far, it seems to work. The end result, with all the cables/pickups connected, could be lacking, what comes to noise levels.
But as always, there is only one way to find it out.

If (when)all goes well with the wiring, then comes time to test, of how the schmitt trigger outputs could be refined further, and to see, do they(running at 1,8V) need to be amplified/gated to feed external cmos-based devices.

 

[PekkaNF]

This time, the audio recording of the 6-to-1 mixed schmitt triggers was done straight to the laptops mic input. So no echo, equalisation or any other effects were used. Plain 'dry' output:
http://www.elisanet.fi/kz1706/optical_pup/testing_new_setup.mp3

......

Wow. I am surpriced that the output is that rich and even some warmth in it - like second harmonics or something "musical" distortion. Must be hard to find a sweet spot location for opto componenets.

[sorveltaja]

Wow. I am surpriced that the output is that rich and even some warmth in it - like second harmonics or something "musical" distortion. Must be hard to find a sweet spot location for opto componenets.

Pekka, yes, for some reason, that seems to be the nature of the true analog 'synth' -sound. Produced harmonics are certainly very complex, providing that 'something', which makes it sound the way it does.

So, that's how the analog hex fuzz sounds. There are some videos on Youtube about commercially available ones, but generally they have a lot more mellow outputs(probably to suit for larger user base), than mine has.

What comes to adjusting the height of the pickups for sweet spots, I'm not sure, but it could be done so-and-so just by listening the outputs, judging the result by how long the string rings, and therefore cause the schmitt trigger to 'squarify'.

To lessen the 'guesstimation' -factors, using a function generator and scope help a lot.

The process isn't that difficult, but can be tedious. Fortunately, it needs to be done only once(assuming the string heights remain the same).

[sorveltaja]

After the wiring for the new enclosure is done, I'm going to test the idea of the previously mentioned individual pickup covers.

At first, for the enclosure, I thought about ordering a Hammond-like aluminum one right away, but decided to hack it together from pieces of wood I have lying around.

As the project is in testing phase, I didn't want to bother with possible need of the isolated/insulated connectors, that metal enclosure might well require.

Could take some time, as I'm not yet sure, what kind of connector to use for the enclosure outputs.

One likely candidate is a cd-rom ide-cable(which I already have at hand), that has more than enough pins(39) for possible 'extended' future options(like modulating each ir-leds individually, using external signals, if needed).

Also each schmitt trigger would have individual outputs for an external mixer, and/or other processing purposes(to convert square wave to a triangle/sine one, or to feed some envelope followers/filters/oscillators etc.).

Lots of babbling, but in short: the enclosure mentioned will include only the circuitry used so far. Rest is about making easily accessible connections for further testing, mainly for breadboard.