Adventures in Thermal Overload
I’m a great believer in hobby projects. In addition to giving me something to do in my spare time (that’s a little joke right there), there’s no better way to learn something than to have some sort of task or function you want to perform, after which you work out how to achieve your goal.
Another thing about working on hobby projects is that you tend to learn things you weren’t expecting to learn — oftentimes things you really would rather have not learned, if the truth be told. In my previous column, Adventures in ESD, for example, I described how working on my Inamorata Prognostication Engine (don’t ask) taught me several things. These included not neglecting electrostatic discharge (ESD) precautions and not assembling things in such a way that they would be hard to take apart and fix when my carefree attitude to ESD decided to bite me in the nether regions.
It was while working on my Prognostication Engine that I discovered something else that will doubtless be of use in the years to come, but that I wish I had learned by reading it in a book, or in an article (like you are now).
This all came about because the Prognostication Engine’s main control panel has five potentiometers, each of which is accompanied by a 16-element NeoPixel Ring from Adafruit. As an aside, these potentiometers are motorized, so if anyone who is unauthorized attempts to make any changes, the Prognostication Engine can return them to their original settings when the interloper steps away (that will surprise them LOL).
I was attempting to attach the first ring to the back of the brass control panel using my hot glue gun (this turned out to be a dismal failure because the brass is both too smooth and acts like a massive heatsink, but that’s a story for another day). Since the 16 elements needed to be precisely aligned with the holes in the control panel, and since I hadn’t really thought this out in advance (another lesson learned), I left the ring powered up and illuminated while I applied the hot glue, only to watch the ring go out in an awfully ominous way.
It was as I feared. When I extracted the ring and started to test it, I discovered that the first LED — the one that had found itself bathed in hot glue — was no more; it was an “ex-LED” (to paraphrase Monty Python’s Dead Parrot sketch).
My first thought was that the hot glue might be conductive in its liquid state — perhaps my cheap-and-cheerful Hobby Lobby glue sticks contained contaminants. So, I started off by performing a raft of experiments, such as powering up the hot glue gun and using my multimeter to measure the resistance of the glue itself, and also measuring the resistance from the gun’s metal tip to the Live and Neutral pins on its power cord. In all cases, the resistance was so high as to register infinite (or open circuit) on my multimeter. To be honest, the main result from these experiments was to cover large portions of my multimeter, the kitchen table, and myself in hot glue.
Since ESD was much on my mind following the disaster with my LED-powered furnace, my next idea was that squeezing the hot glue gun’s “feed trigger” had somehow generated enough static electricity to “blow” the LED. By this time, however, everything was shielded or sitting on a grounded electrostatic pad, including myself, so this seemed unlikely.
One of my chums, Aubrey Kagan, was the first to raise the issue of thermal overload. He noted that the melt temperature of the glue is about 120°C, that the maximum operating temperature of a NeoPixel is only around 80°C, and that the storage (unpowered) temperature rating for the NeoPixel is a lot higher than this. I’ve come to the conclusion that Aubrey hit the LED on the head, as it were, because I’ve soldered hundreds of these little scamps — both before and after this incident — with no power and no problem.
Postscript
As a point of interest, another friend, Rick Curl, took this a little further. Rick told me that he’d run into something similar at the small company where he works, with tricolor LEDs failing all over the place on the boards they were building.
After reviewing their anti-static handling procedures and ruling our ESD as the cause, it finally dawned on Rick that the labels on the LED packages that say “bake if the humidity is above 40%” might be there for a reason. Rick says that, from what he understands, moisture can seep into the LED package and — when the device is heated (through soldering, for example) — the moisture can turn to steam and damage the chip. Rick’s solution was to modify a toaster oven as illustrated below.
Rick says that he took a regular toaster oven, bypassed the thermostat and timer, and connected the heating elements to a 16-volt transformer that happened to be lying around. The oven ended up consuming about 15 watts, which produced a temperature around 100°F. Rick tells me that they’ve actually started to store the reels of LEDs in the oven, they only pull them out just before soldering, and that all of the failures have ceased.
This isn’t applicable to me, but it might be of interest to you if you work for a small company and you are using tricolor LEDs (and other affected components) in a production environment. Speaking of which, what do you think about all of this? Do you have any thermal-related stories you’d care to share in the comments below?
