Author Topic: The 'Dial-In' jumperless Breadboard, MK14, Nominal Animal, (and buck boost cir)  (Read 1206 times)

0 Members and 1 Guest are viewing this topic.

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
   Well now;   This one has to be dedicated, to members here, doing same or similar related posting topics.
   My 'project' is maybe very casual, but nevertheless a serious venture into electronics and questions I never encountered, before.

   The nebulous idea involves serial strings of various subassembled units, like the ubiquitous Solar Garden lights, I've mentioned on EEVBLOG.   (Some was when I used the handle
'haywardcpu'.
   
   Seems I start thinking about the buck voltage doubler used in those garden LED lights, and next thing you know;  TWO other posts appear, threads concerning buck boost circuits and inductance.
    I figure 6 switches, and 6 modules, can be switched into serial orders of variety somewhere around 1300 permutations.
I have two solar light modules, a resistive CDS light sensor, and a 'Christmas Card' LED moving display.   Obviously, many many variations will turn out to be useless.
   But I'm curious.   Things done 'loosely' like this can have interesting and educational surprises.
Happened before, on various other projects.

Also, other people's threads have things that relate;  will try to include those as I go.

Photo showing two garden light circuits, as they are connected in serial form.   What that means, is the LED OUTPUT from the first module, connects directly to the Solar Cell input, of the following module.
The results are a bit puzzling, (already, lol)!

More soon...thanks


   
 

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
   I kept debating, frivolous, or useful ?
That setup I showed sat for several days, needing simply two wires stripped, for the two conductors running from first LED unit, to the next (via it's own solar cell input).  Second unit solar cell, actually, to be a separated, object in the serial switching scheme.   That way, you've got an easy natural connection as was original when purchased as complete solar light with self-contained LED.

   Right away I've gotten perplexed;   the Buck-Boost oscillator waveform coming out WAS capable of driving the next module in line (directly, not optical).

   Best I can figure, that first output has enough low or zero voltage signal to cause the following unit to think it's nighttime.   These units use zero volts as a sensor would be used, as the photovoltaic cell itself can also be used to sense ambient light levels.

The whole little test, in the bench there is giving indications that the second unit, in line, is thinking that it's dark (and so it lights up).   That second unit has its normal LED connected.

Had the scope set up wrong, actually needed to use negative slope for trigger!
Duhuh mm.
 

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
   The overall scheme relies on 6-position rotary switches;   these switches come with two separate levels, for a 2P6T or Two-pole, 6 throw type (also with cute chickaen haed knob).

   I figure two switches gives 6 X 6,
three switches gives 36 X 6 or 216 possible (with one option being a null).
Oops; I guess that comes out way higher number of possible configurations, sorry.

Some 40,000 ways.   I can always downgrade, to 5 and 5,  to have less, uh, testing to deal with.
 

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
   First decisions involve keeping things intact;  that means not shorting two outputs, meanwhile connecting inputs together freely.  So each 'stage' would be one device, driving one line on the system bus.   That'd be the Wells, so to speak, that each device can connect to.   Ironically, I got caught up in auto-debates...as question comes up:
   Do you really need to connect, like, unit2 input to the (same) unit2 OUTPUT ?   So the compromise is, that specific position would, instead, connect to a push-button.   I still do want that option,....so but it will be easy to rewire some so any of the active units (solar light modules, usually) can, yes, be connected to it's own input, later.   Ridiculous.

The learning area for me would be gaining the skill to understand that waveform, of the voltage boosting oscillator.  As shown, I highlighted the rising edge, of the passive collapse of mag field.  I say passive as the 'shorting' transistor has opened, starting the whole field collapse and voltage rise.

   What if some student wants to know specifics, about L-R time constants as you can see the LED OUTPUT voltage rise, to about
2.2 V and then a decay curve.   I'd be challenged to explain those inductor dynamics, in any academically useful manner!

As to the brief example, running the LED OUTPUT voltage direct into another unit, would be considered a 'Non-Inverting' cascade...would have preferred an 'Inverted' relation, but thats what I get.   Have considered including a couple of mid-power NPN transistor, like 2N2222.

If I scoped it properly that oscillator runs at 250 khz, but that doesn't seem right ?
 

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
   Photo shows the preliminary set that might give useful variety of configurations.
   Suppose you wanted to switch these in, keeping a serial order that matches the order of possible elements.
   The first 2P6T switch would connect to unit 6 (for circular wrap).  Second switch is placed to get it's input from first element.   Third switch is placed to get it's input from unit2.   All the way to the last switch, sixth switch, which selects element five as input.

   I've informally gathered two solar light units, one motor unit, with oscillating lever (from a greeting / birthday card), a couple of LEDs, and a speaker, if I can.
   For audio, a speaker is nice, but 8 ohm types need power.   Maybe just a piezo style speaker, that doesn't load things down.
One choice of elements would be nice to be open, undedicated, for attaching AN ACTUAL BREADBOARD....(just so as not to hurt anybody's feelings, lol).
   That comes out to six elements that can / will select which I put they want.

   One irony, the Solar Lights actually present a 'logical high'  voltage potential, when sitting passive, by way of the 1.5 V battery cell through the inductor.   So the 'binary' states would be:
     OFF.  =.  OUTPUT 1.5 VDC
     ON.   =.   OUTPUT 250 KHZ Square wave.
 

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
Ok the posting by ledtester on 'Sandwizz' has description of a software downloadable breadboard, that has various components to activate.   Obviously complex, it will implement your downloaded schematic.
  I'd strongly recommend that great piece, as those sorts of projects facilitators are an amazing glimpse of the future.   Imagine having
a huge playfield containing THOUSANDS of TTL type or MOS type gates...!

   Rogeorge has some interesting and related stuff, here in General Section, so recommend.

   Picture blurry, sorry;   it's one of those 6-way selector switches, having old school style phenolic decks.   With an additional SPDT switch, you could select to one of 12 inputs.
 

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
   If you want to view the posts by ledtester please see;
   Projects:    approx on 2nd page, around May 20 was what I found.   (Search somehow indicated May 11, but it wasn't there)

Rogeorge, also thanks, for the thread on 'Weekend Project', covering the Flickering LED products.
 

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
   The photo shows how a full solar light unit gets separated into sub-sets for varied reconnection, as each variation as experimenting progresses.

   One basic limitation becomes apparent, that while having some 40,000 ways to try out...might have more promise, functionally, the logistics of trying every combination gets very large, as the 'channel' count goes up, much beyond 6 X 6 X 6 X 6 X 6 X 6 (that being 6 modules).
   Top circuit in photo shows original solar light, often skipping any LED load limiting resistor, in some cheaper units available at price discounts (i.e. couple dollars price).
At bottom, shows basic Buck-Boost circuit output, with the 1.5 V battery exerting a 'D.C. output, a little weird for connection use, but hey...might have some advantage, (I don't know at this point).

   Has to be built and tested, before I can really definitively say (what it's good for).
Sounding more like a politician.
 

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
   For more clarity,  enclosed schematic is highlighted to show how the variable selections avoid shorting any two outputs together;   meanwhile multiple inputs may have selected to a same output, but that shouldn't have any faulty effects.

   Actually, though, if I'm going to be dealing with A.C. in the output formats, perhaps it can be useful to mix two or more outputs, each capacitor isolated.   The 'territory' here gets increasingly tentative, which calls for a pretty loose and free-wheeling attitude, on the developer activities,  and testing....lots and lots of bench tests !
 

Online MK14

  • Super Contributor
  • ***
  • Posts: 4588
  • Country: gb
   If you want to view the posts by ledtester please see;
   Projects:    approx on 2nd page, around May 20 was what I found.   (Search somehow indicated May 11, but it wasn't there)

Rogeorge, also thanks, for the thread on 'Weekend Project', covering the Flickering LED products.

Thanks for mentioning that thread.  I've seen it now (properly), and replied in that thread, here:

https://www.eevblog.com/forum/projects/sandwizz-the-next-jumperless-breadboard/msg5518816/#msg5518816

I suppose your various multi-way switches, can (as you suggest), choose from a variety of combinations.  But if those combinations, become huge, e.g. 150,000,000.  Then, without the aid of connecting it to an automated, computer system.  Which could be left alone for a week, and check all those individual combinations out, automatically.

It could take years or even decades, to try out all of those 150,000,000 combinations (if you obtain enough multi-way switches, and connect them up to your project).
 

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
MK14:
   Actually I have the switches in stock,  as that was a main motivating factor.   It's the slightly frivolous nature of the efforts, relative to a more conventional set of tasks, that slowed down starting activities...
 
The following users thanked this post: MK14

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
   Here's a newly seen dilemma I'll be involved with explaining.   Has to do with my recent statement, that one of the Solar Lights control / LED voltage boosters will go into lighted mode when fed a voltage input from oscillating second device (square wave, basically):

   First of all, I'm assuming that LEDs can effectively transmit some square wave at 250 khz.   (???)
   When put the hard-connected trial circuit aside, and try illuminating a Solar Light module by way of a second module's light, the results are that the optically driven unit DOES shut off it's own LED lamp output.
That is same as using, say, sunlight that is essentially a constant value of intensity.
But the two trial cases don't give same result, as I've said, the direct, electrically connected 'square wave' has probably enough 'low voltage' or zero voltage, as it vibrates, to be interpreted as a solar cell in darkness.   At least my speculation.

   So, maybe in the combination, of LED sending it's light to a little solar PV cell gets squashed.  (???).
   I'm going to have to work on that issue, using a spare PV cell, and trusty oscilloscope, to see how much of a square wave makes it through.
That is;  to point one of the little Solar Lights at a spare PV cell, and see if the optical transmission can relatively convey the 250 khz,...or if instead there is a 'low pass only' situation.
   Heck...it's just simply a modulated LED, right ?

(?????). Gonna have try that (Tuesday).
Thanks.
 

Online MK14

  • Super Contributor
  • ***
  • Posts: 4588
  • Country: gb
I don't know, what circuit(s) designs are in your specific solar lights.  But the following link:
https://www.talkingelectronics.com/projects/SolarLight/SolarLight.html

Seems to show two different designs.

One which uses the solar cell itself, to disable its LED light output, during daylight/sunlight hours.  It seems to rely on the solar cell producing enough voltage, to switch on a transistor, which effectively disables the boost converter.

At a guess, perhaps your LED light source, is far too weak compared to real daylight/sunlight.  So the solar cell can't produce enough voltage and hence current/power, to switch on the booster circuit, disabling transistor.

So, the 250 kHz 'square' wave, could just be a red herring, and might have nothing to do with your issues.

CdS light cells (if it uses that method of detection), tend to be relatively slow at responding, so the 250 kHz, shouldn't be an issue.  But what would be, is they would be designed to react to very strong, outside sunlight, rather than a (presumably) relatively weak LED light source, you are using.
 

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
   I'm currently thinking along same lines;
Trying to go back to searching Internet for info, on the SIP (4 pin on-line package) controller, and what process to shut off light output.
Thought I had previously read, that with sufficient ambient light, sure, a transistor gets turned on, but also that activates setting a flip-flop, essentially getting into the 'daytime' mode, thus shutting down the LED illumination.
Maybe that function is a bit sluggish ?

   But certainly I have confidence in using a single Solar Light to shut down a second Light module.   As picture shows, positioning needs to be close, sure, but no worries about one LED having enough brightness, (right up close).
Little experiment shown also blocking ambient light, from getting to the second unit.
   So, left alone, the second unit WILL light up, as the PV cell is covered up.   To do the other state, I've covered up the first unit, causing it to light....thus shutting down the second one.
This is the classic 'inverting' of device states, device to device, when chained by way of positioning that way (shown).

   No red herrings;   but things don't make consistent sense, (thus my interest).  More testing info later today.   Perhaps, also, it's the LOW  value of light, coming from ambient, that sets up a 'Nighttime' state, setting a flip-flop to start up running the output LED, via the square wave output from buck boost at 250 khz.

Experiment trials today might have better clarity.

I need to get a personal license plate, reading:
   DISCOUNT STORE MAVEN
 
The following users thanked this post: MK14

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
Oops, here is the package to package 'serial' optical setup, one sending light to second unit's PV sell.
 

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
   Meantime, more brainstorming:
   While, as MK14 mentions, spending 10 years evaluating 'mutated' variations of hardware interaction maybe seems impractical, but the whole DNA, mutate / test a generation or two hundred generations took nature to where we are today, seems to work, haphazardly.
I might be more situated here to create one or two interesting 'features', rather than a full-fledged 'intellegent artificial being'.

   For example:
   Not nice (engineering) to just randomly connect inputs, from same unit's output, but might be of use here, to connect an LED controlling unit, output, to same unit's input...

   Supposing I put an additional capacitor, an output DC blocking capacitor on the (AC) square wave that the buck booster puts out.
This makes it more likely useful, to MIX two or more signals, directly, before putting into a next stage.   That starts to look, logically, like a logical OR, circuit.   And think about the nature of outputs here, that being an AC waveform...
Then, why not create a little stage, feeding back to itself, that will HOLD the pattern, the AC pattern of output that indicates the (logical) 'BINARY ON STATE'.
   ET Voila:   I've invented a One-Bit digital memory!   Those sorts of simple outcomes might / probably be what extent this whole project turns out!
    People don't go around criticizing anyone playing with a crossword puzzle, with daily newspaper, so should tolerate some minor play with 'unintended' uses of cheap devices.
   I like to call it:
   'SDDA, or Smart Device, Dumb Application'
 
The following users thanked this post: MK14

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
   Here is photo of a couple waveforms, to help clarify yesterday's scope work.   Landlord shares house / let's me use the
 TEKTRONIX 465. Scope...it must be, like a 1979 model Tek 2-channel very near-old- school.

   Notice that 1.5 DC level, maybe has minor current leak, into the RED led.   I suppose that's OK since sunlight is coming in and charging the 1.5 nicad battery.

   A die-hard assembly language type guy, I had trouble with using proper language, for the two binary states;
   If it's, um err, if it's 'OFF', or 'zero'....
...NO !    NO NO NO !
    If it's, err; Buck-Booster 'ON' (Oscillating), then it's 'dark' or,...NO; If it's oscillator 'ON', then the PV Cell is, uh, dark or 'Off', uh, but LED is...

Blah blah blah bla.  See what I mean?

   Best to use terminology like:
   'DAYTIME Mode' = no LED light

   'NIGHTTIME' =. unit's LED is emitting light

   So, photo has the clearest language, I decided.
 
The following users thanked this post: MK14

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
  NO TRACE.
   No visible A.C. present in the little PV Cell I tested, as single component.   The Solar Unit used to drive the LED puts out the 250 khz square wave, mentioned already, but no trace except for a pure / smooth D.C. level from the PV Cell output (approx 2 volts with open circuit, not driving anything).

   I'd expect to see at least some A.C. ripple or something, but apparently between the Red LED and the PV Cell any fluctuating signal is completely absent....(Played around w. Scope a bit, changed trigger slope, etc. to no avail...).

     Anybody see a RED MOON come up, today ?
Shit
 

Online MK14

  • Super Contributor
  • ***
  • Posts: 4588
  • Country: gb
Maybe it needs complete or nearly complete darkness, to get the (presumably) faint signal from the RED LED, across the solar cell.

If you do it in sunlight (which it seems you did), that would be a huge amount of light, and the solar cell, would have great difficulty, detecting the relatively faint LED's, light output.

LUX levels of sunlight and LEDs:
https://www.saving-light-bulbs.co.uk/blog/your-ultimate-guide-to-lux-lumens-and-watts/

So, daylight/sunlight could be 10,000 LUX, yet the small RED LED could be a tiny, tiny fraction of the sunlight value.

I.e. The human eyes are extremely logarithmic in sensitivity, which can cause confusion as to the true magnitudes of differing light intensity levels.

Here is an article about it:
https://www.telescope-optics.net/eye_intensity_response.htm
 

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
   Naw,  test bench and scope are set up inside, and any red LED set up always succeeds, in shutting down PV driven unit (putting it into 'daytime Mode' due to providing sufficient light).
Actually I wanted to try red first, before some white / blue LED.   Of course you are correct...Sun IS bright, out there.  But with red LED right up against the surface of the open circuit PV Cell, get plenty of voltage (2V).
The other state, you get, the 'Nighttime' response, simply by cutting the red light, or blocking it....That proves the room was dim enough, to be interpreted as 'night', thus lighting up it's own output.

    Pre-conclusion is that a 'dark' or zero amplitude never gets into the following unit's circuit, while a hard-wired square wave DOES get the low or zero value light amplitude signal, as a low portion of each square wave cycle.

   Yesterday, I did search on response speeds, or 'modulation detection speed limits.   However, any searches kept wanting to discuss Terra-hertz light waves, (even with my mis-worded search terms).
I'd type in 'modulation response' and get back search on 'frequency vs visible colors', at thz values, not a few hundred hz modulation.

   I'm even skipping some of the OTHER confusing / irratic test bench observations.
But I suppose you'd need a PIN DIODE to track some fast modulation.
 
The following users thanked this post: MK14

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
An additional curious test, perhaps, would be to build a little motor driven 'strobe wheel', cutting out some sunlight, at a few hz, by placing two holes in the spinning disk.   Maybe PV cells are sluggish, compared to a fast PIN DIODE used in IR receivers, etc.

---------------------------_----------



Maybe I'll start up a TEST LAB, staffed by PUPPIES....Yippee !
 

Online MK14

  • Super Contributor
  • ***
  • Posts: 4588
  • Country: gb
An additional curious test, perhaps, would be to build a little motor driven 'strobe wheel', cutting out some sunlight, at a few hz, by placing two holes in the spinning disk.   Maybe PV cells are sluggish, compared to a fast PIN DIODE used in IR receivers, etc.

---------------------------_----------



Maybe I'll start up a TEST LAB, staffed by PUPPIES....Yippee !

This article:
https://publications.jrc.ec.europa.eu/repository/handle/JRC100582

Seems to suggest that it can easily take a number of milliseconds, or longer, for solar cells to respond, to light level increases.
 
The following users thanked this post: RJSV

Offline RJSVTopic starter

  • Super Contributor
  • ***
  • Posts: 2242
  • Country: us
    One slightly sad consequence, then, (besides the fact I've expended valuable bench time on a impossibility), would be that, now that I think about use of a PV to detect FAST stuff, is that a person could have used a fast-responding PV cell to have large area for detecting incoming faint light signals, vs. a small diameter cds or pin diode detector.   I'd suppose a lens and reflector would get you that increase, for really faint incoming signals.

   Oh well, but at least THAT item is explained.
Knowledge gained, at little cost, THANKS EEVBLOG people!

    That explains that a hard connection, with wires can bring in the lower or zero signal levels which then cause response (go to nighttime mode).   Electrically, the little controller might or might not track that fast signal.
   But, clearly the response is as a non-inverting follower, less of interest than an inverting logic.

   Going with that, can't simply use, say, a 2n2222 NPN BJtransistor,  unless design a higher turn-on threshold, higher than the usual 0.7 base to emitter.
    Tempting to put an inverter in the path, but I fear that wouldn't be enough, if you've viewed the diagram, a couple posts back, one state is 1.5 VDC from battery, and other is 'vibrating'.

   Had a dream, a while back, that the human brain stores things in that format;  either vibrating or silent, as coding for the two states.

   FCB ?      That is, Frequency Encoded Binary...

Better is;    VCB. or 'vibrationally coded binary'

    Not necessarily so strange, as you could build up a system using those rules, assuming proper resources.
   Code1 + 0 =. Code1,
   
   Code1 + Code1 = Code0 + carry

That part is easy.
 
The following users thanked this post: MK14


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf