Something a little different- A worklight mod

I have several of these lights. They're made by a automotive tool manufacturer called Uview. They're very well made and rugged, but somewhat anemic. Uview has a newer version that replaces the 30 LED panel with 3 100lm Luxeon units.

I'd like to modify the one's I have and get a bit more performance out of them- I would appreciate any ideas or suggestions.

The tube bodies are 1" ID polycarbonate which measures out to 25.5mm. OAL is 11.5". Current battery is a 12-pack of AAA NiMH marked 14.4v/600 mAh. In the base there's a barrel plug for charger attachment and a micro toggle switch rated at 5A@125VAC. I have approx. 3/16" extra space in the base and over an inch above the LED array so there's plenty of room for a little longer battery and a driver.

On the battery side, there's room (I think) to fit a 25650 battery inside. If necessary I can open the tube up a little. An 18650 would be an option, but I hate to sacrifice the additional runtime. I was considering using a copper strip arrangement to get current up to the LED's from the switch in the base if I can get the 25mm battery to work.

I understand that the polycarbonate is going to be a limiting factor thermally- plus being a work light, I'd like to retain some level of practicality which means a reasonable amount of battery life.

My thinking is to use a piece of 1" dia. aluminum bar stock and mill a flat/channel in the face and mount the LED's there. There's about 3" of open tube where the existing array resides. I'm thinking either three or maybe six LED's in that space arranged in a line. If I could, I'd like to get around 600lm out of it. I'm thinking the tube's inability to conduct heat will become an issue above that point- or maybe even less.

Would six units be better than three for even distribution of light?

What are the opinions on the thermal issue?

What suggestions on LED's to use? I'd prefer a warm white with a high CRI if possible.

I'm open to any suggestions. What's written above are only my beginning thoughts. I'm not married to any of my ideas.

Thanks for any and all assistance!

Depending on the LEDs they use and the power at which they are driven, those LED arrays could actually match the high power LEDs in efficiency.

The Nichia 219 high CRI or Seoul P4 high CRI could both work. Neutral white and high CRI LED, but lower efficiency compared to normal cool white LEDs.

Your idea with the channel should work, you don't even need a deep channel for it to work. I would carve some other channels to act kind like fins. How much runtime would you like to keep? The 26650 batteries might be slightly too wide to fit, and the protected ones are even wider.

As for output, a Nichia 219 high CRI is 90-110 lumen at 350 mA (~3.1V), so 6 of them will give you ~600 lumen at the LED. With 1 26650 battery, you'll get anywhere from 1.5 hours to 2 hours of runtime depending on battery capacity and driver efficiency.

The 219 was attractive to me. What would be the most efficient way to drive them? I have a very basic understanding of how an LED driver works but by no means have any real knowledge. In my reading I have come across references to linear drivers. My impression is that they tend to be more efficient which I assume would lead to lower temperature. levels.

At 350 mA, that's driving the diode conservatively. Am I correct in that as a diode is driven harder (beyond datasheet currents) they become less efficient and thus generate more heat? I recently picked up a Surefire Fury @500lm. After some time it does get warm but nowhere near difficult to hold onto. With that diode driven to 500lm at fairly conservative current levels, would I be correct to assume I would be seeing similar thermal output from the 6X219's?

On runtime, they would only be used for 20-30 minutes max. For extended periods I use a corded light. These are nice for wedging into very tight spaces to complete shorter procedures. At the same time, if it were needed two or three times, it would be nice not to have to recharge it. I suppose an 18650 would be a reasonable option. I get the impression that the better units are closer to the on-paper specs than most other sizes.

I have once considered a mod of something like this on a 12V florescent that had a ballast fail in a matter of months, but I eventually gave up because its difficult to dissipate heat when the heatsink is enclosed. The upside is with one or two 3W LEDs can compensate for the 30 little ones you got in there. A beefy bar of aluminum and a short on-time should be fine otherwise

A linear driver is essentially an adapting resistor, so they can be pretty efficient if your source voltage is slightly above your load voltage. If the source voltage is far above your load voltage, then a linear driver can be very inefficient. If you are using one Li-ion battery to drive a/parallel white LEDs, then a linear driver should be fine.

At higher powers, the LEDs heat up more, and are less efficient. As to why this occurs, I only have a vague idea (heat depend characteristics of the LED and the phosphor).

Due to the lower efficiency of the Nichia 219 high CRI, you'll probably get similar amounts of heat. The Fury uses an cool white XM-L if I recall (correct me if I'm wrong), so you use the Cree XM-L spec sheet and estimate the efficiency from there. Assume ~700 lumen from the LED, since Surefire rates out the front lumens.

When you are holding a flashlight, your hand also acts as part of the heatsink, so you'll have to take that into account as well. I'm assuming you won't be holding the Al portion of the worklight while you are using it.