A few months back, I came across an LED-based desk lamp queued up to go out to the trash. When I asked my wife about it, she said (or at least my recollection is that she said) that it had gone dim, so she’d replaced it with another one. But the device didn’t include any sort of “dimmer” functionality, and I wasn’t (at the time, at least) aware that LED lighting’s inherent intensity could fade over time, only that it would inevitably flat-out fail at some point.
My curiosity sufficiently piqued, especially since I’d intercepted it on the way to the landfill anyway, I decided to take it apart first. It’s Hampton Bay’s 15.5 in. Black Indoor LED Desk Lamp, originally sold by Home Depot and currently “out of stock” both in-store and online; I assume it’s no longer available for purchase. Here are some stock shots of what it looks like, to start:
See: no dimmer. Just a simple on/off toggle:
I don’t remember when we bought it or what we paid for it; it had previously resided on my wife’s sewing table. The Internet Archive has four “snapshots” of the page, ranging from the end of June 2020, when it was apparently on sale for $14.71 versus the $29.97 MSRP (I hope we snagged it then!), through early December of last year. My wife took up sewing during the COVID-19 lockdown, so a 2020-era acquisition sounds about right.
Here’s what it looks like in “action” (if you can call it that) in my furnace room, striving (and effectively failing) to differentiate its “augmentation” of the baseline overhead lighting:
Turn off the room light, and the lamp’s standalone luminary capabilities still aren’t impressive:
And here’s a close-up of the light source in “action”, if you can call it that, in my office:
Scan through the reviews on the product page and, unless I overlooked something, you won’t find anyone complaining that it’s not bright enough. Several of the positive reviews go so far as to specifically note that it’s very bright. And ironically, one of the (few) negative ones indicates that it’s too bright. The specs claim that it has a 3W output (no explicit lumens rating, however, far from a color temperature), which roughly translates to a 30W incandescent equivalent.
Time to dive in. Let’s begin with the underside, where a label is attached to a felt “foot”:
A Google search on “Arcadia AL40165” reveals nothing meaningful results-wise aside from the Home Depot product page. “Intertek 4000145” isn’t any more helpful. And, regardless of when we actually bought it, this particular unit was apparently manufactured in December 2016.
Peeling back the felt “foot”, I was initially confused by the three closed-end crimp connectors revealed underneath:
until I peeled it away the rest of the way and…oh yeah, the on/off switch:
Note the wiring colors. Typically, in my experience, the “+” DC feed corresponds to the white wire, with the “-“ return segment handled by the black wire, and the “+” portion of the circuit is what’s switched. This implementation seems opposite of that convention. Hold that thought.
Now for the light source. With the lamp switched off, you can see what appears to be a central LED surrounded by several others in circumference. Conceptually, this matches the arrangement I’ve seen before with LED light bulbs, so my initial working theory was that whatever circuitry was driving the LEDs in the perimeter had failed, leaving only the central one still operational. Why there would be such a two-stage arrangement at all wasn’t obvious, although I postulated that this same hardware might find use in another lamp with a three-position (bright/dim/off) toggle switch.
Removing the diffuser:
unfortunately dashed that theory; there was only a single LED in the center:
Here’s what it looks like illuminated, this time absent the diffuser:
A brief aside at this point: what’s with the second “right?” in the title? Well, when I mentioned to my wife the other day that I’d completed the teardown but hadn’t definitively figured out why the lamp had dimmed over time, she now said that to the best of her recollection, it had always been dim. Hmmm. If indeed I’d previously misunderstood her (and of course, my default stance is to always assume my wife is right…right?), then what we have is a faulty LED from the get-go. But just for grins, let’s pretend my dimmer-over-time recollection is correct and proceed.
One other root cause possibility is that the power supply feeding the LED is in the process of failing, thereby outputting under-spec voltage and/or current. Revisiting the earlier white-vs-black wire discussion, when I initially probed the solder connections with my multimeter using standard polarity conventions, I got a negative voltage reading:
The LED theoretically could have been operating in reverse-bias breakdown (albeit presumably not for long). But more likely, in conjunction with the earlier-mentioned switch location in the circuit, the wire colors were just reversed. Yep, that’s more like it:
Note that their connections to the LED might still be reversed, however. Or perhaps the lamp’s power supply was current output-compromised. To test both of these suppositions, I probe-connected and fueled the LED with my inexpensive-and-passable power supply instead:
With the connections using standard white vs. black conventions, I got…nothing. Reversed, the LED light output weakly matched that delivered when driven by the lamp’s own power supply. And my standalone power supply also informed me that the lamp pulls 180 mA at 10 V.
About that “lamp’s own power supply”, by the way (as-usual accompanied by a 0.75″/19.1 mm diameter U.S. penny for size comparison purposes):
The label refers to it as an “LED Driver,” but I’m guessing that it’s just a normal “wall wart”, absent a plug on the output end. And a Google search of “JG-LED1-5UPPL” (that’s the number 5, not an S, by the way) further bolsters that hypothesis (“Intertek 4002637” conversely wasn’t helpful at all, aside from suggesting that this power supply unit (PSU) was originally intended for a different lamp model). But I’m still baffled by the “DC5-10V MAX” notation in the labeled output specs…???
And removing two more screws, here’s what the plate the LED is mounted to looks like when separated from the “heatsink” behind it (note the trivial dab of thermal paste between them):
All leaving me with the same question I had at the start: what caused the LED-based desk lamp’s light output to dim, either over time or from the very beginning (depending on which spouse’s story you’re going with)? The most likely remaining culprit, I’m postulating, is the phosphor layer above the LED. I’ve already noted the scant-at-best heat-transfer interface between the LED and the metal plate behind it. More generally, as this device definitely exemplifies, my research suggests that inexpensive designs skimp on the number of LEDs to keep the BOM cost down, compensating by overdriving the one(s) that remain. The resulting thermal stress prematurely breaks down the phosphor, resulting in color temperature shifts and reduced light output, along with eventual complete component failure.
That’s my take; what’s yours? Let me know your thoughts in the comments!
—Brian Dipert is the Editor-in-Chief of the Edge AI and Vision Alliance, and a Senior Analyst at BDTI and Editor-in-Chief of InsideDSP, the company’s online newsletter.
Related Content
- LDR PC Desk Lamp
- Constant-current wall warts streamline LED driver design for lamps, cabinet lights
- Magic carpets come alive with LEDs
- Can GE’s new LED bulbs help you get to sleep?
- Six LED challenges that still remain
- LED lamp cycles on and off, why?
The post Disassembling a LED-based light that’s not acting quite right…right? appeared first on EDN.
