In previous articles, we have been describing the use of LED Christmas lights as emergency off-grid lighting powered directly from a 24 volt battery array. The first article described using strings of cool white lights. Subsequent articles described building a test fixture for these and warm white lights to overcome problems with untested lights, and then constructing arrays with red LED lights, including the use of an external resistor. In this article, we finish the series by describing the use of multi-color lights, some of which are manufactured with a separate resistor embedded in the string. We’ll also describe a useful design procedure for creating off-grid LED strings using any color combination you wish.
You may recall from the previous article that the red LEDs were manufactured as 50-light strings, while the white LEDs were made with only 25-lights per string. All strings tested so far featured many LEDs without integrated resistors, while some LEDs had an integrated resistor to allow the string to be used with varying AC voltage without burning up on the peaks. With one multicolor string, the pattern of the red strings were repeated, with some yellow LEDs having an integrated resistor. However, another multi-color string used separate resistors, which appear as a cylinder inline with the string wiring. Our 150-light multicolor string featured three such resistors, one each for the three 50-unit arrays wired in parallel on this string. We wish to develop a common design procedure that allows lights from any of these various strings to be used with confidence. To develop this procedure, we start by testing the properties of the lights in more detail.
Using our test fixture from several articles ago, we tested lights from strings without these external resistors, and lights from a string with series resistors. First, the results without external resistors:
|Blue||Low||3.06 v||0.06 v|
|Green||Low||3.11 v||0.08 v|
|Red||Low||2.08 v||0.04 v|
|Yellow||Low||2.04 v||0.06 v|
|Yellow||High||4.04 v||0.02 v|
|Yellow Orange||Low||2.03 v||0.02 v|
As you can see, the yellow LEDs were the only ones which contained an integrated resistor. Unlike the all-red LEDs we tested previously, which contained four integrated resistors per string of 50, only two yellow LEDs per string of 50 contained an integrated resistor. Presumably, this is to make up for the higher voltages used by the blue and green LEDs.
Note also that the LEDs (high-range yellows with internal resistors aside) fall into two main categories: blue and green at around 3.1 volts, and red, yellow and yellow-orange at around 2.0 volts.
Now, the results from the strings with external resistor cylinders:
|Blue||2.97 v||0.04 v|
|Green||2.99 v||0.06 v|
|Violet||3.05 v||0.05 v|
|Orange||2.05 v||0.04 v|
|Red-Orange||2.03 v||0.03 v|
|Yellow||2.11 v||0.02 v|
Since there are external resistors in these strings, none of these LEDs have internal resistors, and so there is no distinction between low and high range. Also, although the averages are a little different than for the previous string, this could just be sampling error.
It is clear that the LEDs still fall into two groups: blue, green and violet at about 3.0 volts, and orange, red-orange and yellow at about 2.0 volts. In the procedure which follows, we’ll call these two groups the blue group and the red group.
The final test we ran was to discover at what voltage each LED starts to produce useful light, and at what voltages the LEDs conduct 5, 10 and 15 milliamps, respectively. To conduct this test, we used a string of four LEDs of similar type, and drove this string through a 220 ohm resistor using a variable lab supply to make the voltage adjustments more controllable. For each data point, we measured the current through the string and the voltage across the string, resistor not included. Then, we calculated the average voltage for the given LED type by dividing by four.
This information from this experiment is summarized in the table below. Cool white and warm white LEDs are also tabulated as a reference. Not shown are the results for the LEDs which came in the strings with the external resistors, but you can expect that the results would be similar for the blue versus red groups. None of the LEDs tested had an internal resistor, as determined by the single-socket test fixture; these can be tested, and our results here confirmed, as a nice homeschool science project.
|Color||On||5 mA||10 mA||15 mA|
|Cool White||2.4 v||2.93 v||3.07 v||3.15 v|
|Warm White||2.4 v||2.93 v||3.06 v||3.15 v|
|Blue||2.3 v||2.92 v||3.05 v||3.12 v|
|Green||2.1 v||3.01 v||3.10 v||3.17 v|
|Red||1.7 v||1.94 v||2.02 v||2.06 v|
|Yellow||1.8 v||1.95 v||1.99 v||2.01 v|
|Yellow-Orange||1.8 v||1.95 v||2.01 v||2.04 v|
The green appears to need a little more voltage, but this could be the result of an outlier in the sample set. The On voltage in the table above is highly subjective, and represents the voltage at which the LED appears to be producing the minimal useful light.
Now that we understand the working voltages of the individual light colors, we can now design our 24 volt strings using a common procedure. We might divide this procedure into two options: one to have more consistent illumination over a wide range of battery voltages at the expense of efficient use of power, and the other to optimize power consumption at the expense of noticeably reduced light at low battery voltages. However, to keep the math simple we’re just going to give a single procedure which gives fairly good results. You may wish to fine-tune your resistor choices with a multi-meter.
We will work through an example of constructing a string which reasonably approximates white light at a distance using multi-color lights, although other combinations will work just as well as far as the procedure is concerned. In all cases, we will try to limit the maximum current to 15 milliamps at the maximum 28.8 volts the batteries might experience. With the LEDs we tested, the blue group gets brighter faster at low currents, so mixing white might need more of the red group to balance out the usual red+green+blue used by computer monitors.
To make our procedure general, we will use a different voltage for each group (blue and red) than our tester produces. Because the tester is a little hot with the red group, and a little cool with the blue group, and the LEDs have a strong non-linear response, we’ll use different voltages for each than those previous tables show.
So, for LEDs in the blue group (blue, green or violet), we’ll use 3.05 volts for our calculations. For LEDs in the red group (red, red-orange, orange, yellow, or yellow-orange), we’ll use 2.00 volts for our calculations. We’ll call these voltages our working voltages. You will note that these working voltages correspond to roughly the 10 milliamp measurement in the previous table.
The first step in the design procedure is to select a number of LEDs using these working voltages to lie somewhere between 24 and 24.5 volts, or perhaps a few tenths of a volt higher if necessary (12 volt designs will halve these voltages, 48 volt designs will double them). Looking at the first table, let’s choose two blue, two green, two red, two yellow-orange and two low range yellow LEDs. This adds up as follows:
2x blue + 2x green: 4 x 3.05 = 12.2 volts
2x red + 2x yellow-orange + 2x yellow: 6 x 2.00 = 12.0 volts
Total: 12.2 volts + 12.0 volts = 24.2 volts
Next, calculate the ideal series resistor to limit the current to 10 milliamps at 28.8 volts (12 volt designs will use 14.4 volts here while 48 volt designs will use 57.6 volts):
Resistor = (28.8 volts – 24.2 volts) / 0.010 amps = 460 ohms.
Choose an actual resistor value slightly above this, if possible, although our 10 milliamp design point allows us to choose a value slightly less and still be safe. In this example, we could use any of the following resistors:
• A standard 470 ohm resistor.
• The 500 ohm resistor cylinder that came with some light strings.
• Two 1000 ohm resistor in parallel for an effective 500 ohm resistor.
• Two 220 ohm resistors in series for an effective 440 ohm resistor.
To validate this example, we constructed this string and used a standard 470 ohm resistor in series with our ten LEDs. With this arrangement, the string appears to be producing useful light at 20 to 21 volts, even more so at the normal 24 volt and higher operating range. The actual currents measured at our typical battery array voltages are shown in the table below:
|24.0||4.6 mA||110 mW||10 mW||9 %|
|25.0||6.1 mA||153 mW||17 mW||11 %|
|26.2||8.0 mA||210 mW||30 mW||14 %|
|28.8||12.5 mA||360 mW||73 mW||20 %|
We are satisfied with this result, as it allowed us to design a suitable off-grid LED array more or less blind, without having to tweak the resistor or array elements afterward. And, if we didn’t have the correct resistor, being off a little one way or another is fine. Also, the power wasted by the resistor, while one-fifth of the power at the charging peak, is acceptable at normal operating voltages, and comes nowhere near the 1/4 watt rating of the resistor under any conditions.
In this article series, we have worked through an increasingly scientific process of analyzing inexpensive Christmas LED lights, adapting them as emergency off-grid lighting. Following the design process developed here for any similar LED light in any desired color combination, it should be a simple process to design a useful and long-lived off-grid emergency lighting solution, no inverter required.