The pandemic brought with it a need to maintain adequate ventilation in enclosed spaces, and thus, there’s been considerable interest in inexpensive C02 monitors. Unfortunately, there are unscrupulous actors out there that have seen this as a chance to make a quick profit.
Recently [bigclivedotcom] got one such low-cost CO2 sensor on his bench for a teardown, and confirms that it’s a fake. But in doing so he reveals a fascinating story of design decisions good and bad, from something which could almost have been a useful product.
Behind the slick color display is a PCB with an unidentified microcontroller, power supply circuitry, a DHT11 environmental sensor, and a further small module which purports to be the CO2 sensor. He quickly demonstrates with a SodaStream that it doesn’t respond to CO2 at all, and through further tests is able to identify it as an alcohol sensor.
Beyond the alcohol sensor he analyses the PSU circuitry. It has a place for a battery protection chip but it’s not fitted, and an error in the regulator circuitry leads to a slow drain of the unprotected cell. Most oddly there’s an entire 5 volt switching regulator circuit that’s fitted but unused, being in place to support a missing infra-red module. Finally the screen is an application-specific LCD part.
It’s clear some effort went in to the design of this unit, and we can’t help wondering whether it could have started life as a design for a higher-spec genuine unit. But as [Clive] says, it’s a party detector, and of little more use than as a project case and battery.
Need more dubious instrumentation? How about a magnetic field tester?
41 thoughts on “Anatomy Of A Fake CO2 Sensor”
This makes no sense. They went through all the trouble of designing the device, procuring the components, then manufacturing it, just to mislabel it and sell it as something that its not?
Seems like such a massive waste, a color LCD screen, battery, and a couple of sensors. It could’ve been an actual product! Now its just going to end up in the bin when people notice that opening a coke can doesn’t affect the CO2 readings at all
BigClive goes through a lot of those: Scam products where they could’ve gotten away with way less effort and spending to make a product that would be just as ineffective and safer, or just gone through a few small financial steps to make an actual working (albeit unsafe) product.
Id say its actually recycling of unsold parts.
I was thinking the same, or a major manufacturing error that someone decided to sell to recover costs.
I think the manufacturer is the one that got scammed, one of their suppliers figured out they could relabel a cheap surplus of ethanol sensors as more expensive CO2 sensors (but still far below the price of an actual CO2 sensor) and the customer won’t know.
Maybe a real product where they failed to buy all components (Covid has wrecked the component market). So someone decided to recover some money by taking what they had and package and sell anyway. Missing components not to save production cost but because they didn’t have them. And a different “CO2” sensor that was cheap and could be abused in a pretend way. The original company designing it might even have folder, but someone had access to the partial BOM.
Its actually a real CO2 sensor. If you read though the comments people talk about it, one of the ways to make a CO2 meter is to use a VOC sensor, thats why it is sensitive to alcohol. Its lot cheaper than using infrared detectors.
No, not a real CO2-sensor. You can call it a eCO2 because they can be some corellation between VOC’s and CO2, but it’s not a given that those chemicals track each other. Calling it a CO2-sensor is a scam. Simple as that.
What does the e stand for?
Not to mention the potential liability that could result from the use of these devices. Imagine someone has one sitting in their car and they sit there takling or texting then get drowsy and pass out because the exhaust from the car was filling the cab of the car for some reason. Imagine this device could cost lives.
No, it is not a real CO2 sensor, as it is not measuring CO2.
It might be a eCO2 sensor (e for estimated, I guess) but even the best eCO2 sensors with multiple VOC sensors and clever algorithms are horrible at guessing the actual CO2 levels.
It looks like this product have a connector and the 5V supply for a real CO2 sensor though.
I actually looked at something on Ali a few days ago that appears to be at least the same front plate even if the PCB is different. It is meant to work with the PWM output of a NDIR CO2 sensor. I only needed a MH-Z19 Module so I didn’t get it. Here is the link if anybody is curious:
Hope this doesn’t get flagged as spam because of the link . :)
Is that a fake MH-Z19 module, maybe the “e” version?
CO2 is quite inert and does not have any “electrochemical” cells available. As you say, it has to be NDIR (non-dispersive infra red) laser excitation. A small unit like that would also have to auto-calibrate. I agree with the comment that it is reacting to some volatile organic compound as a CO2 proxy, but ppm precision is out of the question. CO2 is quite difficult to detect, like argon. Usually the oxygen concentration is measured and the CO2 calculated from that, factoring in the CO. CO2 is not dangerous to people until it is over 5000 ppm, or even 10,000 ppm (1%). Take a big breath, hold it, breathe out all the way. There you are: 10,000 ppm CO2.
Electro acoustic chipsets are available, a lot less jumpy than FLIR. So if you are using a duct mounted sensor, go FLIR, like a Veris, and slow down response. But for a handheld, the acoustic ones are the right balance. Increase living room occupancy by three adults, in 10 minutes, ppms are up 40% above setpoint, ventilation rate increases by 60CFM, and the meter sees levels drop in a minute or so.
The most spectacular fake product Clive has yet uncovered is an “ioniser” where the bristles on the front are just connected straight to the mains supply. Shorting them gives spectacular results:
Frightening… Perhaps outsourcing all our manufacturing was a bad idea.
Depending how much it costs its not a bad source of a DHT11 and an alcohol sensor, if of course you need them and don’t mind desoldering.
I think those sensors suffer quite a bit (at least don’t get better) when soldered/desoldered repeatedly…
If you had bothered to research this you would have found that there are 2 types of CO2 sensor modules.
Type 1 measures CO2 directly, and costs about $50-$100 USD for the component itself.
Type 2 measures VOC/eCO2. They don’t directly measure CO2 but they can approximate it under many conditions AFTER they self calibrate, which requires being powered on for 60-90 minutes. These can be had for pennies.
Type 2 isn’t a “fraud”. Most sensors are just a clever algorithm approximating a value from a related measurement.
A step counter isn’t actually counting steps.
A O2 saturation sensor isn’t actually measuring the O2 in your blood.
Your car’s speedometer doesn’t measure your speed.
But the approximations ARE still useful.
+1 ….. it’s not fake !!!!
It’s an interesting article, though :)
It is interesting, and leads one to wonder how many fake CO or smoke sensors are out there.
Quite a lot I would say, but I guess you get what you pay for if buying from reputable supplier. Recently, I had the choice of these cheap sensors and proper expensive ones from a reputable supplier …. and went for the expensive one and it is superb!
They have gotten cheaper. You can get the whole device for about 50€
It IS fake. If you encounter a gas with reverse effect on the cell, you won’t measure any CO2 – you will het false negative.
That’ why there are sensors like MiCS-6814, for example. This one can measure reductive gases, oxidative gases and NH3. But not even thisnsensor dares to claim it would in any way detect CO2.
Yeah, you can calibrate it to somewhat work. But once you get any fancy nonstandard fumes, it stops working – like Clive has shown.
And after all – as Clive has shown – even pure CO2 atmosphere makes almost no difference in the reading.
Sort of like some of the fake shake / Faraday lights from years ago. I have one with a real magnet, a real bridge rectifier circuit, but the coil is just a single layer of magnet wire that’s not connected. The LED runs off a coin cell. Another one I have has a similar coil but the “magnet” is just a crudely cut chunk of metal rod and there’s no components on the PCB.
I do have one that is the real deal. Its magnet is very strong and when the light is held upright the magnet levitates inside the coil. It has a bridge rectifier and a small rechargeable battery. I dunno what kind of plastic it’s made of, but it’s far stronger than the cheap, clear polystyrene the others are made of.
Hmmm, that might be a use for the LED from one of the newer Harbor Freight flashlights that use three AAA cells. The early 5mm white LED in the shake light has rather pitiful output.
Ian commented about approximations being useful.
I am wondering if that is the whole idea behind blood glucose meters. Just approximations. There are too many stories about problems with them:
— two different meters, same drop of blood, significantly different results.
— significantly different results seconds apart
— significantly different results from different fingers.
— same meter, same drop of blood, two test strips.
Significant difference >1.0 (Canadian system)
At the same time, all the medical professionals are still encouraging the use of
I think the most important question here is, can the sensor be replaced with a ndir sensor and end up with an accurate device with a nice lcd screen?
The only fraudulent part is missing the e from eCO2 sensor. I’ve got a genuine NDIR in my office and a few very unscientific cross checks show the results are pretty close. I bought it assuming that it was eCO2, as I’ve used one of these detectors as a module in the past.
Factors that may give a falsely elevated reading include smelly things, people wearing too much aftershave, perfume or deodorant. Also volatile organic chemicals in the air from other sources. Generally none of these things are good for you. So, I have absolutely no problems with the alarm going off when there is someone who is over scented in my vicinity. I like these so much I bought three!
But also, I’d happily pay an extra $40 for a true NDIR CO2 sensor in this package, if you are listening factories and designers of AliExpress!
Check out product called “Hama Safe”. It has battery operated mode (a couple of hours), USB powered mode and a real NDIR CO2 sensor. And all that for around 55 EUR for the whole unit.
I bought this exact model on Amazon a year ago or so. I was the first to comment on the Amazon review and claimed it wasn’t an actual CO2 sensor. I also used my soda stream to test this, though I put the sensor in a plastic bag and filled it with the gas.
I haven’t checked, but maybe they delisted it after I claimed how dangerous it could be. The danger of these fakes is glossed over in this video. There are a number of industries that produce a lot of CO2 and they need warning so as not to suffocate. Gourmet mushroom growers have tents full of they’re oyster mushrooms for example, and if the vent goes down (as it is prone to do if choked with spores), the grower could be walking into a room devoid of oxygen. The fact they set the PPM at a reasonable baseline for CO2 makes this even more dangerous, since people will assume it’s working.
I have one of these shown in the picture. I have tested it against some good CO2 sensors and it has given good values.
It is possible that there is (again) somebody selling fake versions. Chinese companies are famous in that the first versions of products are good and then they start making shit.
Take for example those ‘energy saving’ and LED bulbs. If you take one in pieces, you will find that the old ones were perfect and then they started selling ones, where Voltage, temperature and current limiting circuits were replaced with straight wires.
The same thing with compressors, drills, batteries and welding devices I have tested and opened.
My VT-6IN1 alarm woke me up a few days after using, of only the few days more using since battery doesn’t last hardly much longer than overnight and I got a better PM Detector with the PMS-5003 module that I leave plugged in. Anyways, I had just ripped a huge fart that woke me briefly.
I also found cleaning with 91% IPA, that similar CO2 values were pegged to the max.
The PM detector isn’t good in mine either. I was wondering about reverse engineering where I guess there might be a serial connection somewhere maybe so one can access or find during a power cycling maybe?
Shame if these weren’t documented somewhere to be hacked into a more useful item or at least how to interface with a better sensor modules.
Well, to somewhat get the ball rolling, I made a bad short video a few month back to at least get some moments to look back at the board and modules later, since didn’t find any videos regarding online. Later to start finding the CO2 detector similar designs YT videos suggested reporting as a sham
The board is in Mandarin or something, though might be some access spots to see what the firmware is going on: https://www.youtube.com/shorts/wSHn9lHC104
Are there situations now where the firmware isn’t in English? Just wondering.
This one is similar but has all 3 CO2 HCHO and TVOC. There appears to be documentation here, but I don’t know if all products that use this case and display are the same https://www.visindustry.com/product/bc125new-5-in-1-co2-meter/
Yeah, their BC126 claims to use the ” NDIR gas sensor technology” and that model looks almost exactly the same… at least in regards to the display from a quick glance. The link does show some spec data for the sensor in the last pdf listed as well.
This tear down video shows some similar information with a crappy sensor design in the test section of the video:
You will start to feel anxiety long before the CO2 levels in your environment are actually harmful to you, it is the O2 level that you need to track, and or CO if you have a combustion source nearby. If you fell for all this CO2 detector hyper (and it has been a “thing” that I noted) then you need to be less gullible.
Why measure co2? Isn’t carbon monoxide (co) more dangerous and more likely in a home from a gas boiler/ furnace or log burner/open fire
I bought an externally identical one from aliexpress a while ago and it never seemed to work right so this prompted me to open it up. the pcb is totally different and contains a discrete thermistor and humidity sensor along with what appears to be an MG812 co2 sensor. however testing with co2 from a tank vs alcohol vapor has similar results to what the author reported. there is also an unpopulated 4 pin connector for another type of sensor and the microcontroller isnt lasered off so I’m saving this for a future project
Yet another placebo device from the people’s republic I remember the mod purchased fake bomb detectors probably another placebo device from the prc
I am selling CO2 meters in the Netherlands and Belgium for 10 years now. I represent SenseAir (Sweden) and Aranet4 (Latvia), who have proven to manufacture world’s best CO2 meters. I really hate all the cheap CO2 meters from China. I warn people on my website for inferior CO2 meters as the measurments are not reliable. I made a photo collage with pictures from various Chinese CO2 meters that are offered in the Netherlands. Most similar CO2 meters are unbranded or have unknown brand names. Battery life is not more than one day… (https://www.co2-meters.nl/advies.html)
Please be kind and respectful to help make the comments section excellent. (Comment Policy)