Project idea: solar charged dark-sensing LED

[Meomaxy]

Not a novel idea, I know.

I'm interested in building a project to add glowing indices to a clock that turn on in the dark and run off of a capacitor that is charged by a small solar cell. It will be like a clock with glow-in-the-dark markers except substitute LEDs for phosphorescent paint.

The LEDs therefore only need to turn on in the dark and glow just brightly enough to be seen when it's too dark to read the clock directly. My hope is that with an efficient circuit and small underdriven LEDs it won't need a big solar cell or capacitor.

So, I'm looking for pointers for a starting point for a solar-charged capacitor-driven dark-sensing LED driving circuit for low-power high-efficiency operation.

Thanks everyone!

 

[audioguru]

A huge capacitor might drive a few LEDs for a couple of minutes.
Use a rechargeable battery (like in solar garden lights) to see a few LEDs glow all night long. But then the solar panel must be in full sunlight all day to recharge the battery.

You are thinking of a calculator with an LCD display. An LCD uses an extremely low amount of power because it does not produce light. It can be powered from a capacitor.

 

[kpatz]

If you're driving the LEDs to a brightness level compared to phosphorescent paint you would be drawing very little power. Probably drawing just a few microamps per LED. If you use high-efficiency high-brightness LEDs you could draw mere microamps.

What color LEDs do you plan on using?

Will the clock be in a location that receives sunlight during the day? Or can you place the solar panel in a window or outside?

If the power level to the LEDs is low enough you don't need to worry about switching them; the solar panel could charge the batteries with the LEDs on all the time.

 

[Meomaxy]

If you're driving the LEDs to a brightness level compared to phosphorescent paint you would be drawing very little power. Probably drawing just a few microamps per LED. If you use high-efficiency high-brightness LEDs you could draw mere microamps.

That's my hope.

What color LEDs do you plan on using?

Whatever uses the least power. Red, or maybe orange/green.

Will the clock be in a location that receives sunlight during the day? Or can you place the solar panel in a window or outside?

I'm assuming normal indoor light with the solar panel on top of the clock.

If the power level to the LEDs is low enough you don't need to worry about switching them; the solar panel could charge the batteries with the LEDs on all the time.

It would be nice to be able to control the threshhold for when the LEDs turn on so that they don't start using the charge until it's really dark enough to need them.

 

[Meomaxy]

Super-Luminova spec (among the best material for luminous markers on watches) claims 0.3 mCd/m2 (30 ncd/mm2) for more than 8 hours. I just have a hunch that LEDs can do as well or better.

**broken link removed**

 

[audioguru]

Is it dark because the clock is on a clock-radio at night and you are camping without any electricity?

 

[Meomaxy]

It is dark because the room lights are off, and it isn't a clock radio, it's an analog wall clock. The clock runs on a AA battery, but I don't want to tap into that because (a) it's okay if the luminous indicators run down in the middle of the night, but not okay for the clock to stop and (b) I'm really curious how well this can work and if it would even be theoretically possible to get adequate performance from components that could plausibly fit into a watch.

 

[Mr RB]

I would forget the big ugly solar panel, and stick 2 more AA batteries hidden behind the clock, with a LDR and a little FET to switch the LEDs on at dark. The batteries should last many months, and you have to eventually change the other battery anyway.

And how were you going to make the hands light up?? Or will the LEDs around the face make enough light to see the hands??

 

[mneary]

I became intrigued, and decided to come up with some numbers. I hooked up a ROHM green 0603 LED part no. SML-310-MT and tried various currents in a darkened room. I only gave myself about 60 seconds to get "night vision".

Up close, I could 'find' the LED at 5µA if I already knew exactly where to look, but I could only focus on it at >20µA.

So, twelve of these babies plus three more for the hands are gonna need somewhere around 750 microwatts. So I guess 750µJ/s, or 2.7 Joules/hour.

So then let's see what a good capacitor can do. Sparkfun sells a 10F 2.5V capacitor for about $5; how much energy is that? I found an energy calculator which gives what seems to be a sensible result of 31.25 Joules.

If we can somehow fully charge two 10F capacitors in series to a total of 5V, they could theoretically run all 15 LEDs at 20µA each all night.

If we allow for half the charge (charge to 2.5V, discharge to 1.25V) and some losses, we might design with two or four of these caps. Each one is similar in size to an AA alkaline.

Incidentally, two AA alkalines would last about a year for a little over $1 but it wouldn't be as much fun.

Recharging requires finding a solar array that produces 5V at 2-5mA in room light. I think this could be the deal breaker.

 

[Meomaxy]

I would forget the big ugly solar panel, and stick 2 more AA batteries hidden behind the clock, with a LDR and a little FET to switch the LEDs on at dark. The batteries should last many months, and you have to eventually change the other battery anyway.

I have to agree with you, for this wall clock, that would be the right solution, but I'm interested in seeing how small I can go. I definitely am realizing that a rechargeable battery is a better idea than a capacitor.

In a theoretical watch-based version, the battery would have a capacity in the single digit mAh and support a current draw of about 0.2mA.

And how were you going to make the hands light up?? Or will the LEDs around the face make enough light to see the hands??

Good question. I don't know yet whether the surrounding LEDs will be bright enough.

 

[Boncuk]

How about using two different sets of batteries?

The AA-batteries won't be used for the LEDs.

Using an LI-ion battery of 3.6V and 500mAh the LEDs could be on all night long if you give the battery the chance to charge during day.

Use low current LEDs (full brightness at 2mA) and dim them to draw 0.5mA.

Since a voltage drop of the battery below 3.6V cannot supply two LEDs connected in series they must be connected parallel.

The suggested circuit has a current flow of 32.7µA at daylight, increasing to 5.93mA at full darkness (using an LDR03 as specified).

Even in periods of no sunlight the battery should supply the circuit for 84hours and 18minutes with a fully charged battery.

Boncuk

P.S. For the LED placement use a position underneath the hour hand. That way you can at least see the full hour.

 

[Meomaxy]

I was just looking around for parts and whatnot, and I came across this:

**broken link removed**

It has the solar panel recharging a 1200mAh NiCd, is dark-sensing, and has a string of 60 LEDs already wired up. Seems like I could save myself a lot of work starting with this and chopping off 12 LEDs (more sets left over to play with later) and changing the resistors.

As far as lighting the hands, I've got a few thoughts, none of which seem easy or ideal:

1. Make the LEDs bright enought to illuminate the hands - seems like it will draw too much power and turn this clock into some kind of room night light.
2. Mount LEDS on the hands - I don't think this is even feasible. You can't run wires onto the clock hands because the hands have to spin around, and even if you could, the weight of the LED would be a problem.
3. Light pipe - Plastic or fiber optics can be used to pipe light from the center out to clock hands with no electrical connection needed. Seems like it would be hard to find such a thing or to build it.
4. UV - coat the hands with a fluorescent paint and shine a UV LED in that general direction. Would probably look cool, but might again use too much power. If I was going to shine a UV LED over the whole watch face, why bother with LEDs at the indices? Just put fluorescent paint there too.
5. Use something else for the hands - like Luminova or tritium tubes.

 

[stugray]

Just as a data point:

I have a Citizen Eco-Drive watch that I bought over 12 years ago.

It has a solar cell front bezel ( you cant tell, it is gold/cream colored ) and it charges an internal LiIon battery. This watch has been running all this time & I have never replaced the battery.

I rarely wear it outside in the sunlight, so it has been charging that battery from indoor ambient lights.

Stu

 

[audioguru]

The watch with a solar cell does not use power-hungry LEDs. It uses an extremely low power LCD display that does not make any light. The watch might have a dim light to see it in the dark but the light is turned on by a momentary pushbutton.

 

[kpatz]

Let's say that your 15 LEDs will draw a total of 500µA (giving you 33µA per LED).

Assuming the room will be in total darkness for 8 hours at a time, that's 4 mAH.

A small solar panel I saw at RadioShack produces 6V @ 50mA (probably in bright sunlight). In ambient room light, let's say it produces 4V @ 10mA. Assuming the room gets 8 hours of daylight, that gives you 80 mAH of charging power per day, well more than the amount the LEDs draw at night, and it should be more than enough to compensate for efficiency losses and charge leakage.

If for whatever reason you needed more power, just use a larger panel, connect two in parallel (or series if you need more voltage), or place the panel where the sun hits it during the day.

Pre-charge the batteries so they start out with a full charge and the circuit should be able to run for the total lifespan of the batteries (several years).

 

[Meomaxy]

The watch with a solar cell does not use power-hungry LEDs. It uses an extremely low power LCD display that does not make any light. The watch might have a dim light to see it in the dark but the light is turned on by a momentary pushbutton.

Actually, generally the Citizen Eco-Drive watches don't have LCD displays, they have analog hands. The titanium lithium ion rechargeable battery they use has enough capacity to run the watch in darkness for months, which is extended even further by the special power-saving mode where the hands will stop moving after a certain amount of time in darkness. The watch just uses enough power to keep the internal quartz clock running. When exposed to light, the watch will wake up and rapidly advance the clock hands to the correct time.

According to what I could find on the web, a typical Eco-Drive movement uses an MT920 (1.5V, 5 mAh) battery, which can power a watch in power save mode (i.e. not moving the hands) for up to 4 years.

 

[Meomaxy]

Here's an example of a LED-illuminated watch face:

**broken link removed**
multi-color led watch for your many moods

This is a cheapo watch that can be bought for $12 or less on eBay. What's interesting is that the watch face appears to be lit by two LEDs, and the light is distributed around the backface by light pipe action (clear plastic, I assume).

I wonder how much can the power be reduced in this configuration and still have the indices visible?

 

[audioguru]

Nobody says how many seconds the cheapo watch can light its LEDs.
It looks bright but a camera can be adjusted to make a very dim object look very bright.

 

[Meomaxy]

Nobody says how many seconds the cheapo watch can light its LEDs.
It looks bright but a camera can be adjusted to make a very dim object look very bright.

You can estimate the camera exposure time by looking at the ghosting of the second hand.

You're right, of course. Being bright is not hard if you're willing to sacrifice battery life, and in a cheapo watch, why would I expect anything different? I

The LED placement is not too different from Casio's Super Illuminator:
**broken link removed**

...except the watch is lit from beneath (cooler effect, light spills out the side of the acrylic case, but harder to read the hands) whereas Casio lights above the face.