Extending the soundscape of a hexaphonic guitar pickup

[sorveltaja]

Today, after finding this video, I got stuck in the idea for printing the reduction gears for the pot, after all:

As can be seen, fewer and bigger gear teeth are used. There is a link to Thingiverse for the 3d model in the video description.

There seems to be quite a lot of slack between them, which could probably be adjusted in the slicer settings. Haven't got much into that, so I rather redraw/modify the gears using 3d software.

When there are sun- planet-, and ring gears involved, trying to figure out the gear ratios makes me feel dizzy, so it's better to just print out some samples, and see, how it works.
Nothing critical about the gear ratios, so any result, that 'feels' good enough for the purpose, is worth testing with actual potentiometer.

At first, I'll be using simplified tooth shapes for sketching:

 

This will be a short sideproject, to see, if the concept is relatively easily implemented, as I don't feel like delving into it too deeply at this point.




 

[sorveltaja]

The "potentiometer gear reducer" -side project goes on: I noticed, that gears with straight teeth don't work so well in this case, so helical ones would be better.

Test setup for gears with 45 degree angle:

 

Idea came from this:

 

Printed parts:
 

So far, it seems to work as expected. For now, tolerances are bit tight, but it's just a matter of playing with the offsets of the gear profiles.

Also, as the gears are helical, they tend to 'climb' up or down, while turning back and forth. Some sort of support structure needs to be added to keep them in place.

Currently the footprint, or outer diameter of whole thing is 30mm.   

[Sea.dog]

I'd suggest printing them as Herringbone gears which have no side thrust. The downside is that the outer would need to be split so that it could be assembled from each side and then joined at the separation line. Or maybe the gears could be printed as two halves that come together with pegs to engage them. They'll still require some means of gluing or a snap-together fitting so that they don't part company in one direction.

[sorveltaja]

Seadog, thanks for the idea! I tested it and it works very well to keep the gears in place.

Assembling the thing, yes, it's a bit of dowside, but somehow I managed to cram the gears together.

With the gear sizes, that I've tested do far, the reduction ratio is close to 4(by eyeballing, while rotating). Sun gear has 5, planet gears 8, and ring gear has 22 teeth.

To simplify the thing, I guess only one planet gear(blue one) might be enough, as there aren't much of forces involved. With small, printed gears, there seems to be always hideous tight spots, so reducing the gear count is an attempt to minimise that: 

 

If more gear reduction is needed, there would be room for that also:
 

 

To get less wobbly action, the sun gear should probably use a metal rod(2mm in this case), attached to the green part.

As an alternative for using gears, plain friction based one, as on a picture on the previous post, should make the desingn a lot more easier to build. Using the steel balls would require very tight tolerances, and is therefore beyond the scope.

Again, one possible option could be to use something like rubber as a material for the planet gear, which otherwise causes the tight spots.

[Joules]

I think you are getting confused with the ball bearing pot.   They are no more precise to make than a 3 legged stool, the use of 3 bearings (the items of precision) allow the shaft to self centre, a bearing pot can slip.  It relies on two conical surfaces to work, the shaft is tensioned to maintain friction.  As the shaft and cone wear, the ratio of turns also changes, slightly.  Do you want repeatable location (geared) or high resolution of motion, the two are not the same.

The tight spots in prints will always be present in a cartesian printed part, they can’t produce accurate round components.  The best you can hope to achieve in extruded filament is using a delta printer as they have no XY backlash, but plenty of other setting issues.   Resin printers seem to produce some of the most accurate gears I have seen as they have such high resolution and consistent blob size.

Reducing your nozzle diameter, will also allow more accurate parts to be printed, but a higher level of printer tuning is also required.

[sorveltaja]

Joules, that's interesting. I thought that the ball mating outer metal surface had similar form, like on ball bearing. Thanks for clarifying.

What my aim to achieve with this, is simply to make it easier to find sweet spots, which seems to occur at rather narrow ranges of a 1M pot.
It's a case of listening for the audio output, so the actual resistant value isn't important.

[Joules]

For the benefit of anyone struggling with how this works, the parts contacting the balls need to be metal, but the rest could be 3D printed, such as the carrier that holds the balls and transfers their motion onward.   I know this form of reduction drive from telescope fine focus controls.  The drive can be configured push or pull by the direction of the cones, this is in fact the opposite to the potentiometer in question, only low load applications need apply   8-)

[sorveltaja]

Joules, thanks again. I'm thinking of making one(or more) at some point. Currently I have only 4mm bearing balls on my shelf, which would make the thing too small for my paws to build.

At least 8mm ones would be easier to handle, so I'll order a bunch of them, and other sizes too. But yeah, it'll be fascinating project to get into. Although the working principle is quite simple, it would require its own, separate building thread. 

In the meantime, some temporary (and bulkier in size) two-gear based solution will be used. Once I get to the point of building the enclosure for the vocoder, smaller is better, as I don't know yet, how many more pots/knobs/plugs it will have in its front panel.

[sorveltaja]

To get back to the original subject, it's time for another audio sample. First is the original, that has tango on the left channel, and steam train on the right:

 

Tango_and_steamtrain.mp3.zip

The very same audio file was fed through the vocoder:

 

Tango_train_vocoder.mp3.zip

To get the actual effect, both inputs (mic and instrument) should have signals, that have common frequencies(like above). I've tested both the tango and train tracks separately with animal sounds, like birds, horse, chicken and crow. Not much common frequencies between them.

One trick could be to slow down, or to speed up either of the channel's signal, to get them to the same frequency range. But that's just a tip of an iceberg of possibilities.

Current to-do list: build the fuzz, as it's working well with the vocoder.
Build the 'Super tone control', that has separate high- band- and lowpass controls in it:

 

It should add more flexibility for tone shaping of the input signals. If it does well, I'll build another one for the output also.

In the end, the enclosure, being an empty cassette deck, would be a lot better to work with, than the current cardboard mock-up. Also it would have its own +15/-15V power supply, instead of using the bench psu with wires and clips. In other words: a developing stand-alone unit.


   

[sorveltaja]

For testing the vocoder effects between different audio samples, there is an easy way, using Audacity's(free audio editing software) vocoder, which I discovered just recently.

I tried it with the same tango-steamtrain audio sample, that was previously posted. With default settings, similar effect is there, but has somewhat different 'pronouncing', when compared to the analog device.
------------------------------------------------

Today the ordered bearing balls arrived(for the potentiometer 'gear' reducer). There doesn't seem to be 3d-models of them available on the net, so why not make one.
So far, I have fiddled with different ball sizes to see, what the approximate ratios could be. Printing the ball mating surfaces/objects might work, or then not.

Preferably the ball mating surfaces should be machined out of metal, but I'd like to find out, if there are simpler ways to achieve acceptable results, without using the lathe. 

[Joules]

I will refrain from commenting further on the bearing surfaces, at the beginning of this project I didn’t belive the optical pick ups would work.

 

Well done on that.

[sorveltaja]

I didn't believe either, when I started that pickup project. There I was wondering: "nobody have done this, as far as I know, so it shouldn't work". Then I needed a proof, of why it shouldn't work.

That's how the project started, by just doing against one's better judgement. At some point, the 'shouldn't' changed to 'could', and later on to 'would'.

Don't know, if that makes any sense, but it seems to be the pattern, that I've noticed, when looking back to the recent, experimental builds.

[Joules]

Makes lots of sense, if you ever saw my 3D printed ball nut for the insane.  It started out as theoretical, proved impossible, til I got a 0.3mm nozzle.   Hence my suggestion for your small gear, good to see people push the boundaries and change our thinking and understanding on the way.

I know I said I wouldn't, but....   You might consider 3D printed press tooling.  Cut up some steel beer cans to make discs with holes in.    They could be pressed to form conical surfaces for your prints, allowing a metal surface for bearing contact, without machining.  Press tooling makes good sense as you potentially have multiple parts to make, assuming you have an Arbor press or the likes of.

[sorveltaja]

Measuring, how to fit the vocoder into the enclosure. Bit tight, so the slide pots between band pass filters and optoisolators can go, for now. If needed, they could be added afterwards:

 

I've had a far-fetched idea of adding some kind of an audio analyser to the setup, having an led bar display for each channel, to show and hold the peaks of average values of say, 5 sec. audio samples, to get the 'profiles' of certain sample's frequencies.

In practice, I guess it would get too complicated, if only analog components were used. An alternative could be to use one of those cheapo pocket size oscilloscopes instead.

--------------------------------------------------------
Getting/testing smaller nozzle has been on my to-do list for a while, so we'll see, once the current projects give room for it.

What comes to conical metal(or other material) surfaces, there is a slight chance, that they might not be necessary, as I noticed today, when fiddling with different configurations. Also the tension spring might not be needed. It seems that only the centerpiece, or 'sun gear', that rotates the bearing balls(planets), needs to be made of metal, having smooth surface.

Printed outer ring turned out to be adequate for testing(although smoother/machined surface would be better). Once there is sufficient amount of tension between the parts, they rotate just like planetary gears do.
Almost too simple, so there must be some serious obstacles ahead. To find out, what they are, I'll start a separate thread about the subject soon.

[WeldingRod]

Loving the cardboard aided design ;-)
That's how I did the guitar amp box with my son!

Sent from my SM-G892A using Tapatalk

[BillTodd]

I've had a far-fetched idea of adding some kind of an audio analyser to the setup, having an led bar display for each channel, to show and hold the peaks of average values of say, 5 sec. audio samples, to get the 'profiles' of certain sample's frequencies.

Can you program a PIC micro? (specifically 16F877)

I have a few 101 segment bar graph VFD devices and drivers (I think I'm out of PCBs but the gerbers are in the zip file) . Happy to donate one to the cause if you can use it :-)


https://drive.google.com/file/d/0BwXDcXKx2KzldldqSlZTS0hvT3M/view?usp=sharing

[sorveltaja]

WeldingRod, yes, delightful and universal prototyping material.

Bill, I appreciate the generous offer. Only problem is, that I don't know how to program in assembly langue. Back in the eighties I had the Sinclair Spectrum, which had an option to use assembler(was that same as assembly language?). Although I was fascinated by its sheer speed(sample programs from books and magazines), when compared to ones written in basic, I never understood its construction.

Later on, when internet started to grow, there were more info and samples of it available, but still I ended up scratching my head. So far, basic is the only language, that I can comprehend, and write simple math programs. I remember trying out some basic to assembler(or to machine code) translators at one point, but back then the results weren't too spectacular.

Since then, I haven't looked much into it, so in these days, more robust ways to do the translating could be available. As far as I know, there are also Basic Stamp -systems available, but as they don't seem to be as popular, as PIC-based ones, the prices are prohibitive, especially, if one needs several of them for separate projects.

Apart from the "language barrier", an interesting subject.
--------------------------------------

Today I managed to almost finish the fuzz and super tone control for the vocoder:

 

Again, the fuzz pcb layout was mirrored, which I noticed(again), while soldering the corner legs of the ic socket.

Anyways, It was a breeze to make a corrected one, as the board is so small.

The way, that the fuzz works with the vocoder is already tested, when it was on the breadboard. The super tone control will have to wait, until the vocoder itself is fitted into the enclosure.

The plan is to use as few switches as possible, to avoid unnecessary wiring hassle, when compared to the original, Paia version.

 

[RussellT]

You could try Mikrobasic for PIC.

I've used it a bit - seems to work well.

Russell

[sorveltaja]

Russell, thanks for the tip. I downloaded the demo, but haven't tried it yet.

Now the vocoder has an enclosure, although carboard is still in use for front panel:

 

There will be at least 7 pots more and one switch, once the fuzz and super tone control are installed. Then plug for the headphones, so the output doesn't need to be connected to the pc's soundcard, while testing.

[RussellT]

Russell, thanks for the tip. I downloaded the demo, but haven't tried it yet.

The free version is only limited by size of program - and the limit still gives you a lot of scope.

Russell

[sorveltaja]

Today more cardboarding:

 

Before putting the tone control and fuzz into the vocoder enclosure, it's easier to first test different combinations, to see, how they work. Knobs need to be printed, and power supply wirings added.

[sorveltaja]

Tone control and fuzz are now tested and working. Haven't used them with the vocoder yet, as I got bitten by the "guitar repair/mod" -bug.
__

The new guitar, that I have, needs some minor fret leveling to get the strings lower, without too much string buzz. It should be done right at the first shot, so I'm going to practice the techniques, using the old crappy Ibanez, which requires a lot more fretwork anyways, to be playable.

Then came also the idea to scallop the fretboard of that old bugger. It's something that has haunted me for years. If I mess it up, no harm done. If successful, I have probably given it a whole new life.

But yeah, I'll get back to testing the vocoder with its new companions every now and then.

In the end, I'll install the hex pickup system back to the guitar, that was used for testing it. Before that, there is some rewiring to do, to make it a bit easier.