Half a croissant, on a plate, with a sign in front of it saying '50c'
h a l f b a k e r y
If ever there was a time we needed a bowlologist, it's now.

idea: add, search, annotate, link, view, overview, recent, by name, random

meta: news, help, about, links, report a problem

account: browse anonymously, or get an account and write.

user:
pass:
register,


                                           

Q: Progressively increasing QR code

Just a quick question for other experts here on the halfbakery, to delete later
  (+1)
(+1)
  [vote for,
against]

Can anyone imagine a good way of having a QR code 'progressively' increase according to a lateral movement? I'm not sure whether it should increase in optical density (in which case it isn't really a QR code) or increase in the decoded data (making a number).

Or, if there's a better way than a QR code (most likely).

What I'm wondering is how to do a cheap low-tech potentiometer (probably slider) whereby the data value steps up and down with quite high resolution (to avoid stepping). The problem with just an increase in density sandwiched by an option-isolator type of arrangement is that if you use an LDR, they don't shut off to zero resistance (whereas a pot would do), and if you scale it and bias it or use a photodiode, you've massively increased the complexity and cost compared to a real pot (imagining a panel full of quite a few of these, not just one or two).

It occurs to me that a digital resistor ladder arrangement has an advantage in that the value can be read and stored and recalled later somehow whereas doing that to a real pot incorporates extra complexity. If the value was also represented as density, then it might be possible to achieve full to zero resistance. (And then you'd have to have some alterations in the pattern to give log/lin/antilog).

Just thinking aloud, really. It should be cheap, almost paper tech (or transparent, glass strips, or something). The only reason I was fixated on optical pot is that it might be nice to have a glow that represents the value of the pot (as well as the position doing that) without having that glow value created by additional circuitry or code. However, that's probably a distraction. The minimum product would just be a pot (not necessarily optical, not necessarily glowing) that doesn't rely on internal physical contact, like a sprung slider wiper.

Ian Tindale, Dec 16 2016

interior view of a mountainy prism bicycle reflector https://thumb7.shut...kground-2857647.jpg
[beanangel, Dec 16 2016]

[link]






       I'm not clear why you think a 'no-contact' method would be low-tech, if it involves a digital camera and hefty processing.   

       But if you were doing it that way, the obvious thing to do would be to effectively build an array of one-dimensional bar-codes along the axis of the slider. The barcode for each slider would have an invariant part encoding the slider ID, and a variable part encoding position along the slider as a Gray code.
Loris, Dec 16 2016
  

       Who mentioned a digital camera?   

       Well, one of the thing with Gray code and suchlike from rotary encoding, is that there's not a proportional relationship between the code representation and the quantity - the number 3 is denser than 4 but as dense as 5.
Ian Tindale, Dec 16 2016
  

       I do not understand this.   

       You want some sort of digital pot, yes? But with high resolution? So, what's wrong with an old fashioned analogue pot? What exactly is the application/problem?   

       In short: what??
MaxwellBuchanan, Dec 16 2016
  

       "Can anyone imagine a good way of having a QR code 'progressively' increase according to a lateral movement?"   

       viewed from above a qr code is just a bunch of squares, and they have printed a finite number of them.   

       But what if, as an optical effect similar to the extreme linear perspective travelling near giant ship (star wars "big view') you looked at the QR code gradually, sideways, at high magnification? then the QR code might actually look 3d, with the ink slightly above the page. So if you wanted to print the qr code on something up to 3 overprints high, or as few as 1 overprint high, then you could get 2^3 (8) times more data out of this transverse travel high magnification 3D qr code   

       The 3d QR code could be produced with a normal printer just overprinting the characters it already makes up to 3 times.
beanangel, Dec 16 2016
  

       If you're not wanting to use a gray code, then it's probably best to just have a gradient, or a triangular stripe. For reading it out you can get "luminosity sensors" such as the TSL2561 on breakout boards from china for about £1.30, probably a lot less in bulk. These would need to be read by a microcontroller over i2c.   

       For a full analog solution that's better than a plain LDR, there are log-scale analog light sensors available, which output a scaled voltage, but these are more expensive than the i2c sensors.   

       Non-optical methods:   

       You could have a fixed metal plate in a semicircle on the panel, and another semicircle of metal on the underside of the knob, and measure the capacitance between them.   

       You could have a magnet on the knob, and measure the change in inductance of a wire coil underneath. Or, a hall effect sensor could pick up the magnet's position. Or, a flex sensor on a piece of metal fixed at one end, attracted to the magnet.   

       Oh, here's another idea. You could have a glowing surface with a polarizing filter on it, and another polarizing filter is rotated by the knob so it changes brightness accordingly. If 90 degrees is too short a range, you could have epicyclic gears connecting the knob to the filter, which would look really cool.
mitxela, Dec 16 2016
  

       also there is another way to do it that is kind of like conways game of life. presumably kind of like a computer hash there might be hundreds of different arrangements of QR code that could actually say the same thing.   

       then have some conway game of life thing where "if it has two upper right neighbors, it grows two ears" Then after the entire actual printed QR code is conway game of lifed It has a new larger spatial area, which represents a larger QR code data space.   

       This resembles a thing called fractal compression.
beanangel, Dec 16 2016
  

       hmm. high resolution potentiometer. must have clean zero. and be cheap.   

       you could do it with a rotating prism under the potentiometer knob, then set the zero to the non-transmission of what the prism was made out of.   

       so: rotate a ( /\ ) and have a photosensor respond to the color gradient, which has as high a a/d resolution as your photosensor. Then at the zero, the uv or IR is actually outside the transmittivity of the prism.   

       I am completely clueless about why you want to use a QR code as a potentiometer. (unless you want to rotate a round QR code as if it were a rotary shaft encoder "plate" at 2d)
beanangel, Dec 16 2016
  

       The impetus behind this requirement is manifold. In recent weeks I have desoldered a significant number of the sliders in my Arp 2600 synth (which used to belong to Roger Glover, because it has his name on it). In faultfinding it, I've managed to fix the dead S&H clock (changed opamp), fix the faulty AR transient generator (changed opamp) and fix the faulty reverb amps, mixer out and final out stages (changed opamps). I also cleaned a lot of the sliders that were problematic by dismantling them, soaking the slider tracks in isopropanol, and removing the 1973 caked grease from the slider carriage and reassembling the cleaned track, carriage and cleaned spring contact together with new silicon grease. Those ARP sliders simply haven't lasted the decades, and it is a common problem (even commoner with the subsequent ARP 2800 Odyssey, which used different but worse sliders). I still have a dead VCA which I now discover is encased in solid epoxy (not a sandwich of epoxy top layer and the rest of the potting box filled with silicon, as they switched to the year after mine was built).   

       It occurs to me that if a replica of such as synth were to be built, one major improvement would be to not use that type of slider pot, or in fact, any type of slider pot. The rest of the synth actually in my opinion uses substitutable components and all the parts shouldn't really cost too much (compared to the real thing), However, an ARP 2600 with rotary pots just isn't correct. So I've been trying to think of a cheaper but more durable way of translating linear movement into information. The distraction is that it'd be nice to have the sliders lit up, where the amount of light corresponds with the value of the pot. However, that should probably be put aside. I still think it must be possible to translate a position of a thing into a value, without digitally interrogating it (although, that too, for reading it and storing it, but again that's not a prerequisite). I like the idea of a thing that gets denser the more you introduce it between a light source and an LDR, but LDRs as I say don't cut off to zero, so they're not really suitable for a drop-in pot replacement.
Ian Tindale, Dec 16 2016
  

       [ian] Thanks for the update. i think this might work, and possibly the transparent parts could be 3d printed.   

       this is a slider [ - - - - - - [] - - - - - -] depending on how you move the [] the resistance adjusts   

       Now think of the mountainy side of a bicycle reflector [link] made of little prisms. Make one of those like an optical strip the length of the [- - - -[] - - - -]   

       then put a photoemitter detector pair (digi key) at one side of the [ - - - [] - - - -] it gets different reading based on how much prismatic rereflection it gets.   

       the resolution of the optical potentiometer depends on the size of the little mountainy prisms on the bicycle reflector like strip. Also their possibly contoured angle.   

       The nifty part. if there are optically transparent 3d printers, they can print a variety of different mountainy bike relector prism sizes, shapes, and contouring.   

       then with your almost all optical, yet movable potentiometer you can make awesome synthesizer music.   

       Also this is an interesting discussion, I think you should invent something, write it at this item, change the title, and leave the original question and annotations!
beanangel, Dec 16 2016
  

       Hmm. How about de-inventing the slider pot and instead of a single linear motion, what you do is slide a handle through an angle (this isn't actually new, I used to have three 'BBC' faders that had faders connected to a series of make before break gold tracks to effect a resistor ladder internally, and the fader knob actually swept through an arc). If the arc that the handle was moved through actually shifted a light source, it could… Hmm, I wonder?
Ian Tindale, Dec 16 2016
  

       I would strongly recommend going with conventional sliders. Nothing you can build yourself would have the same feel to it.   

       But if you want an LDR to behave as a potentiometer, use two. Connect them in series, put it between the power rails and use the middle pin as the output. Design your slider window to cover one up as it reveals the other. If the gradient was printed correctly, you could have a linear output which very nearly meets the rails, probably within 0.1%.
mitxela, Dec 16 2016
  

       Yep, I was just arriving at that myself - I was thinking of two sensors, with a light source that the actuator shifts (either from one to the other, or from one to both, depending). In terms of feel, I quite like the idea of a super-duper upgraded edgewise pot (not the cheap naff sort) that presents as if it were a pipe, and you rotate a section of that pipe, and the whole travel is only about 90° or less. Sliders are actually a bit difficult to adjust accurately, and almost always involve using the rest of your hand to fulcrum the motion so that it doesn't just run away. An edge-on thumb-sized section of pipe with a quarter-turn rotation would (I'm guessing) still allow quite accurate and tactile adjustment.
Ian Tindale, Dec 16 2016
  

       //Who mentioned a digital camera?//   

       I struggled to understand what you are actually trying to do... I wrote and rewrote my comment above several times while trying to figure it out.   

       But clearly, you're talking about QR codes, and other optical signals as a replacement for analogue or high-resolution sliders. It seems obvious that you'll need some form of optical reader, QR codes being effectively 2-dimensional barcodes.
  

       I suppose you don't need a *digital* camera, but assuming that isn't your quibble (because analogue wouldn't be easier)...
I guess you don't strictly need a camera even if you're reading 2D barcodes. But then you'd need a scanning laser to go with your optical detector, and ... well, we're not approaching low-tech.
  

       //Well, one of the thing with Gray code and suchlike from rotary encoding, is that there's not a proportional relationship between the code representation and the quantity - the number 3 is denser than 4 but as dense as 5.//   

       Again I'm not clear why you care about that.
Put a human-readable indication next to it.
Or if you absolutely must combine them, and don't care too much about accuracy you could just use the stereotypical triangular 'volume' wedge underneath a strip encoding the slider ID to encode the data. Almost arbitrarily many slider controls all managed with a single 1D barcode scanner. Change the controls, or the layout? Print out a new set.
  

       You now say you're trying to fix an old synth.
Small linear potentiometers seem to cost 79 pence from Maplins, and synth ones ... well, from about 15 quid. Unless you have hundreds it's going to be easier and cheaper to just replace the broken ones. If you're wanting to build a new system to last, then you should probably worry about the longevity of your substitute components. Light sources tend not to last forever.
Loris, Dec 16 2016
  

       Well, I've fixed the broken ones. They're as good as new now.
Ian Tindale, Dec 16 2016
  

       Fractal QR code... hmmmm...
RayfordSteele, Dec 16 2016
  

       //trying to fix an old synth//   

       One option is just to remove everything and install a hi-res camera and lighting in the box. Sliders just become a slidey bit on top with a bar that can by tracked the camera underneath - would work for rotary or linear pots.
bigsleep, Dec 17 2016
  

       // removing the 1973 caked grease// As an aside, this "grease'' run through a mass spec wouldn't be half interesting, would it not? A chemical archaeology dig of Deep Purple's history.
wjt, Dec 17 2016
  

       When I desoldered the 4019 VCA module to attempt to decapsulate the potting, at the beginning of this week, I found a date code scratched into the bottom of it, as was their wont. It said 14-73, which means the VCA was made/potted in the 14th week of 1973. The synth would probably be assembled shortly afterward. Assuming it was then exported and stocked in a shop and then bought, during the rest of that year and perhaps into the next year, and looking at Wikipedia, it seems Roger left Deep Purple around that time anyway. This synth is likely the one used for The Butterfly Ball (and indeed the live performance of it later, at the Royal Albert Hall). It seemed to me that there was significant liquid damage causing electronic faults on the right hand side of the front panel. Fortunately all three VCOs are fine, and the VCF too (which uses the earlier 4012 module). Getting into the faulty/dead 4019 VCA module is the task now, and it has been soaking in acetone for a lot of this week so far, to try to help decapsulate it.
Ian Tindale, Dec 17 2016
  

       The current QR code is very limited in how much information it can contain. I'm guessing it's limited by having battle against the lowest common denominator of (A) a cheap and nasty smartphone camera held by (B) someone with shaky hands.   

       If (A) and/or (B) could be improved, then there's a way forward, otherwise it's stuck in a rut of not being able to hold a lot of information, maxing out about 2.1kb on a good day.
not_morrison_rm, Dec 17 2016
  

       But all it (or whichever encoding means) needs to do is identify which 'pot' or controller this is, and specify the set value.   

       I'm wondering if there is a precedent in nature for encoding such information and allowing the quantity it transmits to be progressively increased/decreased by some physical action or placement. I've been considering how some form of conductive moire or signal-interfering moire would do this?
Ian Tindale, Dec 17 2016
  
      
[annotate]
  


 

back: main index

business  computer  culture  fashion  food  halfbakery  home  other  product  public  science  sport  vehicle