I had a Carbon Monoxide sensor keep chirping at me, it was also displaying END. Turns out that CO detectors have a mandated lifetime and are programed to stop working after 7-10 years. So rather than toss it out, why not tear it down 🙂
Model # KN-COPP-3 manufactured June 2009.
A 9V transformer is actually modular, and just clipped into the rest of the frame.
There is just one screw in the middle under the label that takes the case apart.
Here is the PCBAnd here with the buzzer and CO sensor removed. I suspect the empty sockets are for an ionization smoke detector, with the board also used for a combination smoke/CO detector.The board and the CO sensor are labelled Kidde. Is Garrison selling a re-branded Kidde sensor, or is Kidde and Garrison different brands of the same company?
There are just two chips on the board that we will look at, but first lets first take a peak at the CO sensor. Cutting off the yellow tape that is holding a desiccant pack and a plastic washer onto the top electrode.
It has a Kidde part number (800-0437) I know that it is an electrochemical type sensor. Most of the descriptions of electrochemical CO sensors discuss three terminal devices, this just has two. There is a description of how a two terminal electrochemical CO sensor works here.
In short this is basically a type of fuel cell, which in the presence of CO a chemical reaction occurs
CO + H2O → CO2＋ 2H+ + 2e–
and the protons created flow between electrodes causing a current that is proportional to the CO concentration.
Now how do I get into this thing?
After some deliberation, attack it with the diamond saw blade, starting at the bottom
Turns out there the canister is mostly empty
I drained about 4mL of colourless electrolyte out which I believe is Sulphuric Acid. I did lose some so there might be as much as 5mL in the canister. But it is certainly not full, I estimate the canister volume to be ~12 cm2 so it is less than half filled with liquid
Now turning to the top electrode. The canister is not easy to separate, and formed with quite thick metal. I tried making a couple of small cuts, as well as digging out the rubber grommet surrounding the electrode but it is still tightly held
So in the end I just went full on and hacked it apart?
There is really not much to it, two metal electrodes, both 1.4mm thick, the top one 19.5mm diameter, the bottom one 21 mm diameter. Separated with a 0.5mm thick plastic washer, and in the middle two rubber like discs. The top one which I am pretty sure is a activated charcoal filter (That increases sensitivity to CO and reduces false readings from other gases) is 7mm diameter 0.35mm thick. This is on top of a 9.6 mm diameter 0.5mm thick disc that is the active electrode and I believe is where the reaction takes place. These discs are attached to the bottom electrode in an area that has a surface coating of Platinum, over a central area, that is the catalyst for the reaction. Here is the platinum surface under the microscope.
If you look carefully you can see there is a tiny capillary hole in the top electrode. Under the microscope I measured it as 0.4mm diameter. Large enough to let air in but too small for liquid to enter/exit. On the lower electrode under the active surface there are two 0.4mm holes a couple of mm apart.
Because of the way I
did the teardown hacked it apart I could not tell if the upper electrode region is sealed from the sulphuric acid or not. However even if it isn’t, in normal operation (With the unit plugged in to a socket) the canister is sitting upright with the sulphuric acid electrolyte is sitting in the bottom. This is for several years at a time so I don’t see how the active electrode surface does not dry out, and what the purpose of the ~5ml of acid is sitting at the bottom of the canister.
Now lets look at the two chips on the board. Scraping the tax covering off to show it is a Kidde branded chip Kidde 810-2794
After decap I can see the die marking, RE280 and A&E in a diamond. I have seen the A&E and RE part number before. In the Garrison smoke detector I looked at a few months ago there was an RE46C144 that was made by Microchip. So this is almost certainly a Microchip ic branded (And perhaps designed) by Kidde.
The small (1.65mm x 1.65mm) die itself is visually very interesting to look at and pretty colourfulIt is fabricated on what I think is a BiCMOS process with 2 levels of Cu metal. Without any datasheet I really cannot figure out what its functions are, but it is nearly all analog. Down the right side you can see a large driver transistor. In addition another sizable set of transistors in the bottom left corner.
Most of the visible transistors are Bipolar like this one
Peeling the label off of the other chip on the board revealed it as a Microchip PIC MCU 16F883