I’ve built a better circuit for controlling LED lighting. This time, it is based on a joule thief. A potentiometer is added in series with the base resistor to act as a dimmer. The real magic is done by a photocell placed between the emitter and base of the transistor. Check it out:
Here’s a small emergency light I built from a 1 Watt LED, some phone wire, a small signal transistor, and a rewound toroid from a computer power supply:
The circuit is housed in a 35mm film canister. By using a joule thief circuit, I am able to power the LED to near full brightness off a single AAA battery. It is a fairly stable setup, which allows you to set it down and use it as an emergency lantern.
You might notice the black wire. That is the switch. By placing the bare end into a small hole drilled into the top of the case, you complete the circuit, turning the light on. This made the overall build lighter in weight and cheaper than putting in a regular switch.
Here’s a look at how it fits in the canister:
In the center of the picture, you have the 1 watt LED. Right above that is the AAA battery that powers the circuit. The yellow, red, and green thing in the bottom is the hand wound toroid inductor. You might see the small signal transistor in there as well. Everything is held in place by the mild spring action of the solid core copper wire.
I’m using NiMH rechargeable AAA batteries to power the light. The transistor gets a bit warm during operation, but has held up fine so far. For the circuit used, check out this page. With a bit of tweaking, you can modify this circuit to be more efficient and add the ability to dim or even flash the light. In the near future, I’ll publish another joule thief light that I’ve built that does just that.
With the super joule ringer 3.0 circuit, the best efficiency I measured was 67%. While that is quite good for such a simple circuit, I wanted to see if I could do better. Because of that, I set out to design a simple circuit made from easily sourced components. As you saw in my last post, driving the transformer with a MOSFET pulsed by a 555 worked, but how efficiently?
The total power going into the circuit is .721A*12.6V=9.08W. The power used to run the night light was measured at .0454A*164V=7.45W. This gives 82% overall efficiency with the first build for running incandescent lights. Due to the lack of heat on the MOSFET, I’m betting a lot of the losses are in the transformer. Not bad compared to many cheap inverters. Some further developments might improve this number.
Another important test I need to perform is to measure the light output per watt. The night light runs noticeably brighter on the inverter than it does on grid power, so there might be further benefits to running this system.
I’ve had some limited success running both the light an AC adapters at the same time. Hopefully some simple modifications to the circuit will allow both to run at the same time flawlessly.
Keep an eye on my blog for a schematic. I’ll be posting it soon.
When my camera batteries finish charging, I’ll shoot a video detailing some of the finer points of the circuit. If you haven’t done so yet, check out my youtube channel. Be sure to subscribe to keep up to date with my latest experiments.
After playing around with different transistor / transformer combinations, I’ve finally found a pair that will run lighting at full brightness with no flickering:
From this series of experiments, it seems that these simple inverter circuits have to be made tailored to the load they are going to drive. They are not suitable for plug and play use. Efficiency on these circuits is around 67%, not as good as a commercially made one, but quite good considering they are made with two components that can be serviced by pretty much everyone.
In the video, I pointed out how connecting a long cord affects the operation of the circuit. I think there is a pretty neat application for this effect, so keep a watch on my blog as well as my youtube channel for updates.
Some of you might be familiar with the joule ringer series of circuits. If not, I suggest checking out lasersaber’s website and youtube channel. In a nutshell, they take low voltage DC and convert it to high voltage AC for running lighting. They seem to hold a lot of promise for off the grid lighting, but I wanted to see if you could run a small AC adapter off one for charging cell phones or running weather radios.
I used an air core instead of ferrite core transformer to cut on costs. The air core seems to work quite well, though it took a bit of experimentation to get it where it is now.
Here’s a video of my test:
So we see it works. Like mentioned in the video, there are better ways, but in a pinch it does the trick. My transformer doesn’t have enough output to run the light and adapter at the same time. The light will glow on and off when both are connected. A more powerful transformer might be able to run both at the same time.
This video also illustrates the value in old electronics. Heat sinks and components can often be scavenged from them, so before tossing them out, see what you can find on them to be used in other projects.
We also see the alum battery still running strong after over 6 months of intensive use and abuse.
I’ll be doing more experiments with these joule ringer circuits in the future so be sure you are subscribed to the feed on this blog to keep up.
I shot a video of my driver in action. Yesterday was a perfect day due to all the moving clouds. This created variable ambient brightness within the room and you can watch the circuit adjust itself to compensate:
There remains a bit of work still to be done, I’m hoping to have a schematic posted in a week or so.
Obviously, the little LEDs I’ve been experimenting with are not sufficient for lighting an entire room. The idea is to produce a “bulb” from them for a much lower cost than off the shelf LED bulbs. Here is what I call the skeleton bulb:
You can see a homemade heat sink on one of the LED units, but I have since realized these are unnecessary when running the LED on the driver circuit. Because the voltage and current is fed to the bulb in short pulses, very little heat output is observed. The cost of this bulb is under $3.50 and puts out 1080 lumens. Running at full power, it would consume 12 watts. The cost of a premade bulb with this kind of output is around $25, and that is if you buy a 10 pack.
The lighting put out is in the warm range, and is very even. The only concern I have with this light at the time is my children looking directly at the bulb. For their safety, I will have to find a inexpensive way of diffusing the light so they can not look directly at the LEDs. Looking directly at the LEDs can produce permanent eye damage due to the high intensity of the light produced.
In my last post, I demonstrated a cheap homemade light meter made from duct tape, a photoresistor, a paper towel roll, some test leads, and a multimeter. After doing a bit of studying on photoresistors and data sheets, I’ve figured out how to calculate the lux output of my lights based off of the resistance measured. This allows me to get a better idea of actual lighting efficiency. Running through the driver, the calculated light output is 1793 lux. Running directly from a voltage source through the appropriate current limiting resistor, I calculated 2204 lux.
Now for calculating my bang per buck. I’ll spare the gory details here, but if you are curious as to how this works, check out this page. You’ll need the data sheet on your photoresistor as well if you wish to build your own light meter. I get 1793 lux out of .447W of power, making 4011 lux per watt through the driver. Running directly, I get 2204 lux out of a .986W of power, making 2235 lux per watt. A lux is equal to 1 lumen(lm) over a square meter(m^2). Lighting efficiency is calculated as lm/w, so over a given area, you can tell from these numbers that the LED driver gives you much better lighting efficiency.
I’ll be posting a schematic of the driver when I’m satisfied with its performance. If you haven’t done so already, subscribe to my feed to keep up with my progress on this and other projects.