In a previous article, we described the adaptation of LED Christmas lights as emergency solar lighting powered directly from an off-grid battery array. This approach provides many advantages versus running them from an inverter, including useful lighting while trying to fix your inverter. In that article we made some recommendations about string lengths, current and light levels. Immediately after publishing that article, we discovered that some strings burned out although they should not have, and some strings did not burn out when they should.
To figure out what is really going on, we built a test fixture using materials accessible to most people, and tested batches of cool white, warm white and red LED lights. The results are enlightening, and the whole project makes a great homeschool science lesson. Plus, by using this test fixture with your own lights, you can create light strings, emergency or otherwise, which are more reliable, consistent and long-lasting.
Materials and Tools
The materials shown below were used in this project in addition to the tools listed after.
• One LED Christmas light socket, cut from a full string (center).
• One pair of solderless banana plugs; red and black shown.
• Two lengths of two-wire security system cable.
• One 220 ohm resistor, 1/4 watt or better.
• One USB cable suitable for cannibalizing (cut end shown from bottom).
• USB charging adapter (not shown).
• Heat-shrink tubing or hot melt glue (not shown).
• Soldering iron and solder (not shown).
For convenience, we’ll use 5 volt USB power for our test fixture. You can use any USB charging source for this fixture. However, there is something poetic about using a cigarette lighter adapter, attached to our off-grid 24 volt array, to power our test fixture to characterize what will become our 24 volt light strings.
Before assembly, homeschoolers may wish to use an ohmmeter to verify the resistor value and discuss the color codes. The resistor we used is actually a 219 ohm resistor, which is well within the 5% tolerance. This value will be useful in a science project spreadsheet, mentioned later.
Our materials are to be assembled to create the following circuit:
Note that the socket has a little tab on it, both in the diagram and in real life. This tabbed side represents the anode, or positive side of the light, and is attached to the red lead wire in our fixture. For the resistor, we are using the same resistor value as in our field phone Morse decoder project, so if you have to buy some (noted later) you can use the same value on both projects.
The first step is to prepare the USB cable. We started with an old USB A-B cable, the kind with the chunky home-plate shaped end that almost nothing uses anymore but always seems to be the first one you find when looking for a mini- or micro-USB cable. Almost everyone has one of these lying around unused, or knows someone who does. Cut about two feet from the flat end, the A side that goes into the computer or charger adapter. Then, strip about two inches of the outer insulation and peel back the shield and inner foil wrapper. The cable will now look as shown below:
You will notice four wires, shown here as red, black, white and green. Red and black carry the five volts that we’ll use in our tester. Now, carefully and neatly trim all the shield wire back, as well as all the foil. Then, trim off the green and white wires to unequal lengths to prevent them from shorting against each other, the red or black wires, or the shield wires or foil. In the photo below, we have also stripped the ends of the red and black wires and tinned them with solder:
The USB connector supplies power, the red and black plugs attach to the voltage and common terminals of a voltmeter, respectively, and the socket accepts the LED being tested. The resistor and most of the connections are hidden beneath heat-shrink tubing. Underneath that outer heat-shrink is more, smaller heat-shrink tubing around the individual connections. The main job of that blob on the end, and the long wires for the meter, USB and LED socket, is to keep mechanical stress off of the resistor. You can accomplish the same goal by hot-melt gluing that end of the assembly to some rigid stick, like an ice-cream stick or a KNEX. Even if the LED socket is directly shorted, the resistor can never experience more than an eighth of a watt, so we don’t have to worry about thermal stress either.
As noted in the diagram, all the black wires from each cable are connected together in the blob at the end. The USB red wire is connected to one end of the resistor, and the other two reds are connected at the other end of the resistor. Make sure that these three connections don’t short against each other, and it will work fine.
Finally, the ends of the meter wires are inserted into the holes of the solderless banana plugs, and the screws tightened. Whether stackable plugs, as used here, or the more widely available and less-expensive straight-through kind are used, is unimportant. We have tinned the ends of the stranded lead wires first to give the plug screws something to bite into.
Using The Test Fixture
To use the test fixture, prepare it by plugging the banana jacks into a voltmeter, or a multi-meter set to measure DC voltage. Plug the USB connector into a USB charging source. We used a SoftBaugh USB charging adapter, attached to a NOCO GC017 cigarette lighter adapter with battery clamps, and then to our 24 volt battery array. Homeschoolers will wish to measure and record the open-circuit voltage which results for later use in a spreadsheet for the project. In our case, this was 5.23 volts. USB charging voltage is allowed to vary between 4.75 to 5.25 volts. Although most chargers will be on the upper end of this range, it is important to check. This voltage, along with the resistor value, will affect the results you measure with your test fixture.
Next, remove bulbs from their original sockets. This is easily done by first lifting the tab, and then inserting a thin, flat-bladed screwdriver in the gap between the socket and the LED housing. Twist slightly and the housing will jack out of the socket, making it easy to remove. It is then a simple matter to press the LED housing into the test socket, keeping the tabs aligned, until the LED lights. Record the voltage which results.
Note that it is not necessary to seat the housing all the way into the socket. Just making contact is enough. Also, some of the LEDs, especially on the white strings, will be backwards in the sockets. These units came from half of the three-wire bridge sockets. By putting these LEDs in backwards, the manufacturer could simplify inventory. If desired for these 1-in-25 LEDs, pull the candle-shaped clear plastic shield off, remove the LED, flip it around and put it all back together again. You may need to trim the leads, and you will probably break the leads. It may not be worth the effort to salvage these units.
Some LED varieties fall into two ranges, a lower voltage representing no internal resistor, and a higher voltage representing a built-in resistor. Other varieties appear to have a consistently small embedded resistor. We tested cool white, warm white and red LED Christmas lights, and discovered the following mix of test voltages among the white varieties. Red and other multicolor LEDs have different behavior, and will be addressed in a future article.
|LED Type||# Per
|Cool White||25||3.2 (+/- .1)||3 in 4||4.2 (+/- .1)||1 in 4|
|Warm White||25||3.1 (+/- .1)||2 in 3||4.1 (+/- .1)||1 in 3|
As shown above, both white LEDs had a large group centered around a low voltage, indicating little or no internal resistance, and a smaller group centered around a higher voltage, indicating a moderate internal resistance. The warm white LEDs were approximately one-tenth of a volt lower than their cool white counterparts. The warm white LEDs also contained more high range units, 1 in 3 versus 1 in 4 for cool white, probably to make up for the lower average voltage. For both white styles, each string of 50 units was actually configured as two 25-unit strings in parallel.
These results confirm online reviews which claim that one in four strings is dead-on-arrival, or fail almost immediately after their first use. We suspect that the manufacturer creates a blend of no-resistor and high-resistor units in the above proportions, and then relies on probability to insert more or less the correct proportion in each string. Sometimes, however, the tails of the curve strike, and strings are either too dim or burn out quickly, depending on how many high-range resistor units have been used in that particular string.
We also analyzed our four-string and six-string that both used about 11 milliamps, as noted in that previous article. The six-string had three each of high- and low-range LEDs, while the four-string was all high-range LEDs. These strings were cut from the larger string at random, and just happened to work out correctly, which further supports the hypothesis that the manufacturing process is random rather than a predetermined pattern. Additional sampling shows no discernible pattern on fresh strings.
A homeschool science project may wish to record the exact values encountered from a variety of strings, plot these, and analyze.
Based on these results, we now recommend constructing the following off-grid emergency lighting strings for 24 volt arrays. We’re recommending a variety here, to account for the more or less random mix of LEDs you might encounter. Because of granularity, 12 volt options are limited (half of the 4+2 options), but 48 volt users can simply double the numbers shown.
|LED Type||# Per
|Cool White||6||3||3||9 mA||Long Life|
|Cool White||7||5||2||7 mA||Long Life|
|Cool White||6||4||2||11 mA||Brighter|
|Warm White||6||3||3||11 mA||Long Life|
|Warm White||7||5||2||10 mA||Long Life|
|Warm White||6||4||2||15 mA||Brighter|
It is not necessary to sort or cull beyond identifying which range a particular LED occupies. The variations within each range are small enough to be ignored. Although warm white uses a little more current, we recommend those over the cool white as the light quality is better. Even the 15 mA option uses less than a half-watt.
Sources For Materials
While we plundered our sponsor SoftBaugh’s R&D stock for materials, readers may wish to purchase components from the following sources:
• USB cables can probably be scrounged, as mentioned before. Or, buy a multi-pack of new mini-USB cables from Amazon for about ten dollars, then cut up some of your older worn cables for this and other experiments.
• Resistors can be purchased in five-packs locally from RadioShack for about $1.49, or you can get a pack of 50 from Amazon for about that same price. Don’t worry about having that many extra lying around, these things are always handy to have in various values. Even better would be a kit with 25 each of 16 popular values for around $10 from Amazon. That kit contains values that are also useful for long, custom wiring with the field phone remote bundles.
• Banana plugs can also be purchased locally from Radio Shack for between $3 and $10 for a red and black pair. The least expensive you can find will be fine, but we prefer the solderless set-screw variety as these are easier to reuse among projects. The fancy side-entry Pomonas used in this project were overkill. Or, you can get a set of 20, ten each red and black, straight-through plugs from Amazon here for about $8. As with resistors, having more of these lying around is a good thing.
• We used a SoftBaugh charger adapter, as noted above, but any will be fine, or for that matter, any charging source. Just to be safe, we recommend a source that is not your PC, and even the weakest charging adapter will be fine for this project, which, worst case, uses a maximum of 25 milliamps.
In this article we discussed building a tester for Christmas LED lights so that they can be used for reliable off-grid emergency lighting, powered directly from a battery array. In a future article in this series we will use this same tester to analyze the non-white varieties, as well as construct longer strands of more easily employed lights.