Suggestions for tri color LED patterns/uses

[MrAl]

Hello folks,


My question is about a tri color high intensity LED, with blue, red, and green.

As you know, these base colors will produce any color just about,
including white if mixed properly.

The question is, what kind of patterns might be interesting for an LED
like this? For an example, back when i was working with a Z80
controller board i had used a bi color LED for testing. This LED
only had two colors, red and green, which of course also means
it could do amber by turning them both on at the same time.
The pattern was a sort of random but gradually changing intensity
of the green and red, so that red might be 20 percent while green
was 50 percent, then red would step to 21 percent while green
stepped to 49 percent, but that's only one possibility.

The visual effect however is not really describable in terms of
duty cycle, but rather in terms of previous human experience.
That is, when the human eyes see this somewhat random pattern
if it is done right it looks just like a candle flickering. The degree
of flicker can be varied too of course, by changing the timing.
The small but significant difference in the mounting positions of
the red and green dies mean the flicker actually appears to move
as the two intensities vary over time, which really enhances the
flicker effect.

This 'candle flicker' test was a good test for the controller, and
it was interesting to do, but now that i have a tri color LED
the possibilities are even greater for some kind of visual effects.
I am thus asking about what visual effects the folks here might
come up with, such as candle flicker, model police car lights,
etc.

The second part of the question is based on more practical uses.
For example: color matching. Possibly adjust the color (three
push buttons for stepping the duty cycle to each LED color) to
match the output to some reference color, such as some paint color
or whatever.
I guess another practical use would be some sort of sophisticated
indicator that needed to indicate something that had several dimensions.
The color could lead the human mind into almost instant comprehension
of something that normally would require basic interpretation and then
comprehension (with a little practice of course).


So that's two key inquiries:
1. General visual effects
2. Practical uses

Thanks much...

[Dave Dixon]

Hi Mr. Al,
I doubt this is of any use, but years back I used one of those old led's on the "lock" output of a PLL. Darn, I don't remember the number now! It showed red when it was totally out of lock, then gradually turned from amber to green as it was locking in on the frequency. That was so much better than just a "green/off" indication... much cheaper than having a dedicated freq. counter as well.

[MrAl]

Hi Dave,


Actually that is an excellent idea. It could show a complex state
that doesnt need to be numerical but yet is good to be able to see
in between states without requiring a meter or something.
I'll have to think about this some more.

Ah yeah, a refrigerator or freezer indicator, where color is
related to temperature. Dont need a temperature meter, just
a tri color LED (maybe even just a bi color) to show high,
low, on target, and various color mixes to indicate 'nearly on target
but a little low', 'nearly on target but a little high', etc.
I think i might actually use it for this because i am tired of changing
batteries in my wireless temperature meter for the fridge.
A simple uC project with that LED for the output should do it i think.

Thanks for the idea.

[Dave Dixon]

GREAT! I was having one of those days where my ideas weren't panning out here at work. Now I feel like I made a contribution! Now I have to go to a meeting <sigh> I am sure nothing will get done for me the rest of the day
BTW I like the refrigerator idea. I just might do something like that in the near future.
I like my (domestic)beer just above freezing!

[MrAl]

Hi Dave,

Oh domestic? I go for import mostly now, but there are some
made in the US i still like. Actually i think Guiness is made in the
US now too.

If i can i am going to start on that refrigerator monitor as early as
tomorrow morn some time. I used my last two crummy Rayovac
AAA cells in that thermometer so i am motivated to build one that
doesnt need that tiny cell size with its limited capacity.
If i build it soon i wont have to buy AAA cells for a while
Anyway, i was also putting it off because i couldnt decide on what
kind of display to use, and i didnt feel like going to full blown seven
segment or LCD but i also didnt want simple on or off target, so
this new idea will really pay off because i dont have to order anything
new to do this. The algo will be simple too: take reading, display some
combo of red/blue/green that reflects the temperature. I guess
blue for the colder temps, red for the higher temps, and green for
right on target. This will mean blue green for too cold and amber for
too warm. I might have to play around with the color vs temperature
sensitivity for a while (a few experiments) but it should be quite
interesting.
Now that i think about it i have an old thermometer around somewhere
that displays digits and has temperature sensitive crystals that
change color, so this will be similar just with no digits.
To add to the sensitivity because the eye/brain is a little too relative
when it comes to the physiological perception of color, i might just
add another pure green LED right next to the tri color. This will
provide the eye with a good reference color for what is "right on"
the target temperature.
I'll try to post some pics too.

[jwax]

Are you going to make one of these?
http://www.chinavasion.com/product_info ... cet-light/

[MrAl]

Hi John,


Thanks very much for posting that link. That's interesting to check out
too.

I had started the design now, and so have thought about it a little bit
more. I am thinking now that it would be best NOT to change the colors
in gradual steps (like 99 percent blue, 1 percent red, then later
98 percent blue, 2 percent red) because the eye/brain will have
much difficulty discerning what temperature is what when there are
small changes involved, even though i intend to drastically increase
the sensitivity. What would work better for this app is to use a blinking
scheme in tandem with the coloring scheme (and with blue being the 'on
target' color). In other words, for 'temperature on target' blink blue
once, and for say 'temperature one degree lower' blink blue once as
before, but then right after that blink red once too. Then for two
degrees lower blink blue once and then red twice. Alternately,
blink blue for a shorter time or even not at all when the temp is off
target. Green would blink for temperature too cold (not as likely
however with a refrigerator i realized so tri color may be an overkill).
This would provide for a pretty accurate display without having to
use 7 seg displays, which would be an overkill for a meter with
such a narrow range (i think +/- two or three degrees will do it).

Now that i have started this project i am open to new ideas about
what would be the best display technique, given the tri color LED.

I also realized that tri color might be an overkill because it could
display many more states than i really need. For a good example,
say the unit puts out nearly pure white light for 'on target'.
Well, that would leave room for three different tints: red, blue, green,
and all colors in between like violet, blue green, amber, and various
shades and colors in between them like blue violet, etc.
Might be better suited for describing airplane angles

[ptribbey]

Probably old news, but I was thinking, if you could build a suit covered with them, and so many cameras, you could blend in to your background and become nearly invisible. My .02
Paul

[MrAl]

Hi there,


That's a very interesting idea really. Would be cool to try but it would take a huge amount of LEDs unfortunately.

Anyway, here is an example of how color can be used to show the state of three variables.
In this case this is actually part of a study of how the resistor R2 affect linearity and the color
represents the output in degrees F. Since a tri color LED can make just about any color,
the state of the three variables could be shown with one LED.



This pic doesnt always show up in the forum for some reason:

Plot1.gif (63.36 KiB) Viewed 4563 times

[MrAl]

Hi again,

I got to build the prototype today and got it up and running. Looks pretty cool. Perhaps a project for N & V ?

I was a little disappointed by the current draw however. It turned out to be 1ma when i had predicted
about 500ua (half of that which it is). I have to wonder now if this is the result of using Timer1 in the
PIC chip rather than NOT using that timer for the algorithm basic timing? Anyone have experience
with measuring the current draw as the result of using Timer1 (ie not keeping it 'off') ?

[MrAl]

Hi again,

Just a little update...

I decided to change the algo to something that would use much less power, because after changing batteries
in my 'regular' store bought wireless thermometer (for the fridge) every friggin two months (yikes) i was
after a 2 year life span using two alkaline cells and 1ma average current draw wasnt going to get me
anywhere near that length of time.
I decided it would be best to power down for the time when the display does not have to be shown,
and this saves a lot of power. The current draw is now down to 0.150ma (150ua) average, so i think
that's good enough for now. I intend to get this down to less than 20ua average in the future however,
which will extend battery life enough to go back to using two AAA cells.

The tri color LED is just a little disappointing too however, in that when all three colors are on
(i changed the algo to also show 'white' as a signal) it looks more like a Chinese Fire Cracker than
a white LED It shows red blue and green in a bunch but they dont mix well. Still, it's
attention getting so i guess i'll keep it for now

[jwax]

I'd like to see it MrAl, but I get the dreaded

Plot1.gif
You do not have the required permissions to view the files attached to this post.

I was signed in to the forum too.
Could you provide a different link to it?
If it's a video, perhaps put it on youtube?

[MrAl]

Hi John,

Geeze sorry about that, im not sure how that happened again.
Here is the pic i was posting:

[jwax]

OK, got it now MrAl!
Lots of changes going on in this forum lately! Cool!

[MrAl]

Hi again,


Just another little update...


Changing some parts i was able to get the average current draw down to 66.2ua, so this
thing should run for at least 2 years with two AA cells. Even i dont mind that much
battery changing I feel that this design is good enough now.
Almost time to get rid of that wireless battery eater thermometer.

It's also of some interest that the display 'system' uses an average current draw
of close to 68na (that's 68 nano amps). That comes from keeping the display
off during operation when it is not needed. I think this is important for battery
operated equipment where we need some indication of what is going on but
dont want to have to wear the batteries down too fast just to get that info.