From one extreme to the other (just posted about batteries in freezing temps)…
We all know what turbo timeouts and heat sinking is good for, but what are the actual limits for various LEDs before they burn out? Has anyone measured the temp of the MCPCB? How hot can it get before becoming a problem? Also, providing that wiring and soldering is good, is the LED the weakest link? Is it the most likely part to fail if the light is run too hot for too long? Any concerns for the driver overheating if tightly fitted into a small pill?
LEDs don’t usually burn out unless severely overpowered or poorly heatsinked or both. LEDs tend to degrade over time that is, lumens per watt decreases. I believe LEDs are run for 1000 hrs and estimated life is extrapolated. For 2nd Gen LEDs, the die is maintained at 85deg C. The die, not the star. Personally, I expect even lights running at 3A to degrade faster than cree’s estimates but in my reality, changing an LED is usually not a problem and lights will likely be upgraded or given away long before much lumens loss. In short, ‘heat becoming a problem ’ is dependent on what you call a problem.
I have run an LED to the point where it is no longer working. If it has reflowed itself or burned out I can not say, but it was done in a test environment without a host and the pill was extremely hot.
I’m asking because I have programmed a crude sort of temperature monitoring that actually works (tested in a light) but need to get an IR thermometer in order to see how accurate it is and adjust it if it is accurate enough… So I’m more interested in actual measurements than “too hot to hold” but I guess it hasn’t been done. I’ll have to take a measurement when the pill is too hot to hold and work with that. The idea is to replace the turbo timeout with temperature monitoring.
There is an approach that may work without measuring temperature. The way I test my heat sinks is to put a lux meter somewhere in front on the light I'm testing. I also hook up a DMM (set to measure current) between the driver and the emitter. I then turn the light on high and watch for light output drop in the intended high/turbo run time. If I have significant drop in lumens but not in current to the emitter, I either beef up the heat sink more or decrease current for high/turbo. If the lumen output is accompanied by current decrease, I check out the cells and driver I'm using for issues.
You could do the same, but instead of adjusting thermal path or current, you adjust your FW step down threshold.
Best wishes. Really interested in hearing more details when you get to a point you'd like to share.
Mike, I had an XML with <10hrs use using the normal convoy 5/50/100% driver die instantly when I hooked it up to a Qlite driver. Not even a flash.Determining death temperature for an LED is NOT an exact science.
i abuse my lights, and i’ve had LED’s slide right off the PCB. (And believe you can just set the LED back in it’s place, turn it back on high, and solder it right back in place the same way lol) Short of that, no issues. Light still bright, cannot even hold it, Dreamcharger says cell still has same internal resistance after the abuse…
I cannot find heat to be a significant problem.
I haven’t popped the plastic wrap off a cell tho, so i guess I don’t abuse them tooo much (18sixfifty’s wife did that once)