Airwick Black Edition Candles

[Grayfox]

I do not know if anyone has tried out these candles but they a rather cool in what they do.


From what i see the board is about 1"wide and has the following
1 3mm Red LED
1 3mm Blue LED
1 3mm Green LED.
1 IR Photodiode
a 121 and a 124 SMD resistor
Some kind of chip(no clue as it is a small black dome(which cycles each LED))
And all powered by 2 CR2032s

I make candles as a hobby and would like to implement a very basic version in one.

Would i just need a battery, IR photo diode, transistor and a LED

Cant find anything on google about a IR activated LED circuit.

[Corporate666]

So what does it do? 

[Stonent]

So what does it do? 

[Grayfox]

When it detects IR light(from the flame(even a remote control or an old incandescent light)) it does a Blue, Blue-Green, Green, Green-Red, Red, Red-blue.
cycle

For those who do not get this thing.
[youtube]258Xux9o-n4[/youtube]

[David_AVD]

For youtube, just post the bare link like so:

[Grayfox]

Ah thanks.

[TerminalJack505]

Here's what I'm guessing they are doing.  (See attached schematic.)  The photo diode is probably a photo Darlington (= high sensitivity.)  Hard to say what wavelength it works at. 

The MCU might actually be a small custom ASIC.  The input from the sensor is likely just a level-triggered interrupt pin that the MCU can be woke up from.  For example, INT0 on most AVR MCUs or a pin change interrupt.

The MCU is powered-down until the sensor wakes it up.  When it wakes up it just goes through a sequence of lighting the LEDs.  While awake it would monitor the sensor pin to sense when the IR is no longer present.  In which case it powers down.

To make your own, you could use something like an ATtiny 4/5/9/10/11/12 or 13.  You'd want to drop the voltage from 6V to <= 5.5V with a diode, though.  And, since you aren't building millions of them, you could use separate resistors for the LEDs.  This will make matching the brightness levels of each LED much easier.  They might be using PWM to do that.

[Grayfox]

It detects Infra Red since it worked under incandescent bulbs, worked when i pointed a remote at it.

Found this
Could i replace the photo resistor for a photo diode without changing the parts all that much

[ivan747]

If you want to make it very small, you could look for a microcontroller with an integrated comparator. Put the MCU on sleep, the comparator interrupt wakes it up, and from that point on it is just software PWM (at least for low end PICs as the have no more than 2 channels). It could be done with an 8 pin device. Look for something newer than what is used in typical hobby projects, there are lots of new low power parts, that is the main advantage of newer parts. Plus it's more likely they are integrating comparators in the newer 8 pin MCUs.

[Corporate666]

Some companies actually make LED's with R, G and B dies already in them, and a tiny chip that cycles through the colors, just like that one does.

In that case, all you need is to use the photocell/transitor to switch power to the LED's. 

Here, I found some of those LED's on Sparkfun

https://www.sparkfun.com/products/11452

I'm sure you can get them on eBay even cheaper from China.

Of course, using an MCU would be a more fun way to do it, but if you are just looking for the simplest solution, that's it.

[TerminalJack505]

It detects Infra Red since it worked under incandescent bulbs, worked when i pointed a remote at it.

Found this
Could i replace the photo resistor for a photo diode without changing the parts all that much

Something close to that might work.  You'd have to build it up and play around with it.  Photo diodes will give very little current so you'd need a transistor with a large beta and the trimmer might need to be bigger.

The attached circuit was something I played around with in the simulator.  The current source is simulating the photo diode, which would be connected at that point (in reverse.) 

Note that I used a darlington configuration to get a high beta.  That might not be necessary if you can either get more current from the photo diode and/or use a transistor with a really high beta.

[Grayfox]

Will this one work?

[ResR]

Nope, sorry, mate. I tried this schematic long time ago. The phototransistor just gives too low current to operate the LED or something it doesn't like, but I had to stick the phototransistor just against the bulb to make the led glow slightly.

[TerminalJack505]

Will this one work?

Use a photo Darlington instead of the photo transistor and you might have half a chance.

The real problem with circuits like that is that they will have states where they are neither on or off but somewhere in between.  Ideally, you want to detect a particular threshold and switch (completely) on or off.  Someone mentioned a comparator earlier in the thread.  That's the sort of functionality you really want.

You can get cheap voltage detector ICs (for example) that draw very little current and have open drain/collector outputs capable of driving an LED. 

Something like the attached schematic might be worth looking into.  If you're really clever then you might be able to power the IC completely from one or more photo diodes.  This would make for an extremely efficient design since the battery would be used only when powering the LED.

[Grayfox]

I was planning on using a photodiode.

Kicad didn't have a photodiode type symbol, so i used something that looks like it.
Planning on using about 3 red LEDs max.

[TerminalJack505]

So far as the LED(s) go, I liked Corporate666's idea of using one of those "rainbow" LEDs.  You just turn it on and it does the rest.  I've got a couple of them and their pretty cool.

Regarding the photodiode, the thing you don't seem to realize about them is that they generate their own current.  They can be thought of as tiny little solar cells (with sensitivity to a particular wavelength of light.)  The current they generate in the case of a candle flame will likely be very small.  A photodiode won't work in your circuit in that configuration.

You might Google "photodiode equivalent circuit" to get an idea about how they work.  They are basically a current source in parallel with a shunt resistance (of megaohms) and a series resistance of several ohms.  Check out page 5 of this document, for example.  There is also a diode in parallel that limits the maximum voltage that can be generated.

So, if you're so set on using a photodiode then you need to design around its characteristics.  Really, no matter which sensor you use, you will need to start the whole design with the sensor in mind.  I'd order a bunch of photodiodes and see how much short circuit current they generate when you put them on the far side of a candle flame through how ever much candle wax you intend to bury them under.  That current will give you an idea of what you need to do to convert it into a signal to power an LED.

[Grayfox]

Saw that jaycar has a kit that does this.

Got the kit but failed to fully finish it due to my soldering iron pretty much dieing
 But this is the schematic that is in the kit

[TerminalJack505]

That circuit might not work for a candle flame.  The IR source has to pulse on and off for that circuit to work.  It looks like it is intended to indicate when a signal from a remote control is detected. 

[Corporate666]

That Jaycar kit is more complicated than this problem needs to be, IMO.

Is there a reason you seem reluctant to use TerminalJack's recommended solution of a transistor (or darlington) to switch the LED?  It's very simple and should work and will allow you to tune it to get the result you need.

I don't think it will get any simpler than a transistor, light sensor, couple of resistors and a trimmer with color-changing LED's. 

[Grayfox]

Cant get any BU42xx at jaycar which is the local electronics store.
Ebay may not just sell me 1 or 2.

[TerminalJack505]

Hook that photodiode up to your multimeter and see how much current it generates when you expose it to a candle flame.  Bury it in some candle wax to simulate exactly what you would do for the final project.

Once you know how much current it generates you can design a proper circuit.  You might start with something like the attached.  (Note the current source, diode, 10M resistor and 10ohm resistor are simulating the photo diode.)

Throw that in a simulator and replace the 10uA with the current that the photodiode actually generates.  If you can find a datasheet for the photodiode then you can plug in the proper parameters for the shunt and series resistance as well. 

Once you do that it should just be a matter of playing around with the values of P1 and R3.  These will be very dependent on the transistors' gains, however, so be sure to use the proper  transistor part in your simulation.  Even
then the actual transistor betas will vary so watch out for that.  The values of P1 and R3 will also depend on the amount of current generated by the photodiode.

Also remember that, as your batteries are used, their voltage will droop so you'll want to simulate with a voltage of, say, 4V to make sure the circuit still behaves.  If you are using coin cells then it might be wise to factor-in their series resistance, which is fairly high.  (Something like 20ohms, if I remember correctly.)

[TerminalJack505]

I bought one of these the other day.  Something the OP didn't mention--but is very important to how it functions--is the fact that there is some kind of light pipe next to the wick.

I'm not sure what it is made of.  It's pretty hard so it isn't just clear wax.  It obviously has to melt with the wax and not give off toxic fumes.  So, to make a homemade version of these, you'll have to solve that problem as well.

[SeanB]

Probably polyethylene. It will burn completely in the wax flame without giving off an odour.