A quick perusal of the instruction sheet educated me that the CO alarm was the “end of life signal” operating state, and it would need to be replaced. The word “teardown” near-immediately popped into my still-groggy brain.
As regular readers likely already realize, I’m no fan of obsolescence by design. However, I do have exceptions to the general distaste rule, specifically in cases where continued use beyond reasonable operating lifetime might cause harm. Does a printer, whose manufacturer unilaterally decides to terminate its functional existence once a (convenient-to-manufacturer) unreplaceable part is deemed so ink-soaked that its continued use would adversely affect print quality, count? Uhh…no. But, say, a carbon monoxide (CO) sensor? That’d qualify for an exemption, although as you’ll see, the manufacturer seemingly still took (slight) liberty with it.
We have two CO alarms in our house, one on each level. They’re fairly elementary First Alert CO600 units, powered solely by AC; more exotic variants from both this and other manufacturers are battery-powered (either as backup or solely), add smoke detection capabilities, etc.
Last fall, in the middle of the night (of course), the one downstairs started loudly chirping, five rapid-cadence iterations each minute. Speaking of smoke alarms, at first I thought that the sound was coming from the ancient hardwired Statitrol SmokeGard 907B2 smoke detector hanging above it and from the ceiling; I leave it to creative readers to imagine the content and intensity of my utterances once I’d unhooked the smoke detector, still heard the next sequence of still-present chirps and realized that I’d need to re-wire and re-attach it before proceeding.
Once I discerned the true sonic source, I saw that the red LED embedded inside (normally steady) was also blinking. A quick perusal of the instruction sheet (PDF) educated me that this was the “end of life signal” operating state, and that the CO alarm would need to be replaced. The word “teardown” near-immediately popped into my still-groggy brain. Here’s our victim, as usual accompanied by a 0.75″ (19.1 mm) diameter U.S. penny for size comparison purposes:
The large front-side button serves two primary purposes:
In scanning the instruction manual while prepping for this writeup, I discovered a third function, one which would have come in handy last fall in the middle of the night:
The LED, which normally glows red steadily, is visible through the hole at the bottom of the button. And perhaps obviously, the alarm speaker is peeking through the vent at the left.
I must say, in turning the unit over, that the back-panel sticker has perhaps the highest text-per-square-inch density that I’ve yet encountered in a product:
But wait! Rotate the unit so its bottom edge is visible, and even more bold type appears!
Time to dive inside. If you look again at the backside overview shot, you’ll see tiny holes in each corner. These provide convenient access to spring the latches holding the two sides of the case together. This convenience is something I’ve not encountered before, notably with a consumer device (AC-powered, to boot) for which poking around inside is especially verboten:
At first glance of the PCB, I was immediately confused. Wasn’t this supposed to be an AC-only powered CO alarm? And if so, what was that brown rechargeable backup battery doing there?
The answer, of course, is that it isn’t a battery. A quick Google search of the manufacturer (Figaro) and model number (TGS5042) enlightened me that this was the CO sensor:
Specifically, it’s an electrochemical-type gas sensor. As Wikipedia explains:
The electrochemical detector uses the principle of a fuel cell to generate an electrical current when the gas to be detected undergoes a chemical reaction. The generated current is precisely related to the amount of carbon monoxide in the immediate environment close to the sensor. Essentially, the electrochemical cell consists of a container, two electrodes, connection wires and an electrolyte, typically sulfuric acid. Carbon monoxide is oxidized at one electrode to carbon dioxide while oxygen is consumed at the other electrode.
For carbon monoxide detection, the electrochemical cell has advantages over other technologies in that it has a highly accurate and linear output to carbon monoxide concentration, requires minimal power as it is operated at room temperature, and has a long lifetime, which typically is five years to ten years. This technology has become the dominant technology in the United States and in Europe. Test buttons only indicate the operational effectiveness of the battery, circuit and buzzer. The only way to fully test the operation of a CO alarm using an electrochemical cell is with a known source of calibrated test gas delivered in a shroud to maintain the concentration level for the test period.
Next step: Get the PCB out of the remaining case half-shell. A flat head screwdriver-as-lever was all I needed to separate the AC prongs from the circuit board; no desoldering was necessary:
Highlights (IMHO) in this front-side overview included the curiously unpopulated capacitor and fuse sites in the lower right corner, and the switch and LED in the center. The IC to the right of the switch is an enigma to me; here’s what its topside mark says:
Google searching on I870263002 and I870263002-V/SS produces several for-purchase pages, albeit none with meaningful information on what the 20-lead SSOP-packaged IC actually is or does, aside from claims that Microchip Technology is the manufacturer (which the arcane logo stamped on the IC doesn’t allow me to confirm). Curiously, however, one other Google-regurgitated search results leads to the PDF datasheet for a series of Microchip 24 MHz, 2.5 mA rail-to-rail output op amps…whose product names are in no way reminiscent of I870263002, nor is the I870263002 anywhere mentioned in the datasheet. I assume that this IC is somehow involved in the AC-to-DC conversion process, but beyond that I’m at a loss: any ideas, readers?
Also, check out this closeup of the test points:
And of course, there’s the transducer (speaker) at left, now visible in full view…its topside, at least. First Alert’s specifications claim that it meets or exceeds current UL standards of 85 dB at 10 feet (3 meters). While I didn’t test the claim with a SPL meter, I can concur that it’s loud.
Peering underneath the transducer, I see more ICs underneath:
So of course, I felt compelled to proceed with the investigation. The transducer was seemingly solidly soldered to the PCB at three solder-mounting points, but remembering my earlier AC plug pry success, I decided to press my luck with the screwdriver. Once again, I was rewarded:
Here are a couple of closeups of the transducer’s underside, showing the PCB connection sites:
And here’s the now-exposed additional circuitry (speaker-removal scratch marks were from yours truly), including the system’s “brains”, a Microchip Technology (I can definitively confirm the manufacturer this time!) PIC16LF1827 microcontroller with embedded flash memory:
For grins, here’s a closeup of the other half of the PCB front side, complete with the mystery IC:
And wrapping up, here’s a look at the as-usual more boring PCB backside (unless you’re into solder points and traces, that is):
In closing, let’s circle back to my earlier comment that First Alert may have taken liberty with my obsolescence-by-design exemption. We moved into the house in mid-2014 and I bought the original set of CO alarms at that time. That means that this one lasted a bit more than seven years before it began making end-of-life utterances. But if you revisit Figaro’s product page, you’ll see that the manufacturer rates the CO sensor for 10 years of usable life.
On the one hand, a 25% operating-life subtraction seems excessive. But then again, the CO600 is currently selling for only around $25 on Amazon as I type this. Wikipedia’s lifetime claim for electrochemical gas detectors is more conservative: “typically is five years to ten years.” And we are talking about saving lives through timely, reliable alarms here, after all. So, I’m still willing to give First Alert a (still-slightly reluctant) obsolescence-by-design “pass”. What do you think, readers? Sound off in the comments!
This article was originally published on EDN.
Brian Dipert is 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.