Gadget_Guru
Newly Enlightened
Howdy all,
I picked up an LED tent light in Walmart's camping section a week or so ago. It has 24 5mm LEDs, which are much shorter and blunter than typical 5mm emitters. This shape causes them to emit a very wide-angle flood. It's not bad, stock, but a simple, inexpensive mod makes it much better, in my opinion.
Shown here is the innards of the light, stock. These batts are the cheap and nasty carbon-zinc AA cells that are packaged with the light (Why do they even bother?):
As I recall, current to the string of 24 5mm LEDs with the four included "fresh" carbon-zinc AAs was about 500mA. This means the average to each LED would be right about 20mA, not overdriven at all. But the current will immediately start to drop as the batts weaken. This gives very long run time, probably tens of hours. But I think gradually dimming battery-powered lights are annoying and crude.
To limit current from the nominally 6V batteries, the stock light uses eight 30-Ohm SMD resistors. Although all 24 LEDs are wired in parallel, there are eight groups of three LEDs, each with it's own resistor. Quite a good design as resistor lights go, I suppose.
To make this light have a constant brightness for most of the battery life, I installed a small linear regulator in place of the resistors. The resistors all came off quite easily with a tiny screwdriver and a soldering iron. You can just see them in the next photo, in the bottom left corner. They're the little black rectangles with numbers printed on them:
After the resistors were gone, I had to bridge the gaps they occupied to carry the current. I used some fine solid copper wire, tinned then soldered across the gaps to complete the circuit:
Next, the AMC7135 regulator board went in:
It's an easy job, since these boards have easy to solder pads on them. My board has just one regulator IC on it, giving a measured 337mA through most of the life of the batteries. This means that each LED is seeing about 14mA. They are extra efficient at that low current, and will last for a LONG time. Forward voltage of the string of 24 LEDs in parallel is 2.95V, quite low. The total LED power is just a fraction under one watt.
I removed the reverse-voltage diode protection from the board, since it is a bit wasteful, dropping input voltage a half volt or more. With the diode removed, I have to be certain to install the batteries with their polarity correct, or the regulator board will get fried. That small risk is more than made up for by the extra time in regulation gained by this simple mod. The battery voltage must now drop to less than 200mV over Vf before the driver board goes into direct-drive.
I have not yet tested the run time with new cells, but I did a quick test with a set of four Duracell AA alkalines that had a 2003 exp. date on them. They came in a used camera flash I recently bought on eBay, and had been used enough to be pretty weak. Here's the graph showing the very flat regulation, followed by a fairly steep decline, then a shallower curve that would likely go for hours in "moon mode":
Another option is to run three cells, with a dummy cell in the fourth battery slot. This would give shorter time in regulation, followed by a longer, shallower "moon mode" decline. Overall efficiency would be higher with three cells, as well, since the fourth cell's extra voltage is being burned off by the AMC7135 during all of the initial part of battery discharge.
I picked up an LED tent light in Walmart's camping section a week or so ago. It has 24 5mm LEDs, which are much shorter and blunter than typical 5mm emitters. This shape causes them to emit a very wide-angle flood. It's not bad, stock, but a simple, inexpensive mod makes it much better, in my opinion.
Shown here is the innards of the light, stock. These batts are the cheap and nasty carbon-zinc AA cells that are packaged with the light (Why do they even bother?):
As I recall, current to the string of 24 5mm LEDs with the four included "fresh" carbon-zinc AAs was about 500mA. This means the average to each LED would be right about 20mA, not overdriven at all. But the current will immediately start to drop as the batts weaken. This gives very long run time, probably tens of hours. But I think gradually dimming battery-powered lights are annoying and crude.
To limit current from the nominally 6V batteries, the stock light uses eight 30-Ohm SMD resistors. Although all 24 LEDs are wired in parallel, there are eight groups of three LEDs, each with it's own resistor. Quite a good design as resistor lights go, I suppose.
To make this light have a constant brightness for most of the battery life, I installed a small linear regulator in place of the resistors. The resistors all came off quite easily with a tiny screwdriver and a soldering iron. You can just see them in the next photo, in the bottom left corner. They're the little black rectangles with numbers printed on them:
After the resistors were gone, I had to bridge the gaps they occupied to carry the current. I used some fine solid copper wire, tinned then soldered across the gaps to complete the circuit:
Next, the AMC7135 regulator board went in:
It's an easy job, since these boards have easy to solder pads on them. My board has just one regulator IC on it, giving a measured 337mA through most of the life of the batteries. This means that each LED is seeing about 14mA. They are extra efficient at that low current, and will last for a LONG time. Forward voltage of the string of 24 LEDs in parallel is 2.95V, quite low. The total LED power is just a fraction under one watt.
I removed the reverse-voltage diode protection from the board, since it is a bit wasteful, dropping input voltage a half volt or more. With the diode removed, I have to be certain to install the batteries with their polarity correct, or the regulator board will get fried. That small risk is more than made up for by the extra time in regulation gained by this simple mod. The battery voltage must now drop to less than 200mV over Vf before the driver board goes into direct-drive.
I have not yet tested the run time with new cells, but I did a quick test with a set of four Duracell AA alkalines that had a 2003 exp. date on them. They came in a used camera flash I recently bought on eBay, and had been used enough to be pretty weak. Here's the graph showing the very flat regulation, followed by a fairly steep decline, then a shallower curve that would likely go for hours in "moon mode":
Another option is to run three cells, with a dummy cell in the fourth battery slot. This would give shorter time in regulation, followed by a longer, shallower "moon mode" decline. Overall efficiency would be higher with three cells, as well, since the fourth cell's extra voltage is being burned off by the AMC7135 during all of the initial part of battery discharge.