1PCS MQ-138 MQ138 Formaldehyde HCHO Volatile Organic Compounds (VOC) Gas Sensor Module with Alarm Signal Light

My expectations were not that high considering that a photo-ionization detector (PID) would certainly be preferable, but they are hard to obtain for VOCs, datasheets are scant, and they are *expensive*.I was surprised at how well the MQ-138 works. No, you won't get a super accurate PPM reading, but if what you want is delta readings then it will do what you want. In particular I wanted to test for acetone presence and it is very reactive to acetone fumes.The datasheet for this sensor is can be found by the manufacturer (Chinese company called Winsen). They provide a manual for usage of the sensor component. Strangely, a company called Hanwei Electronics provides a somewhat more detailed datasheet. So far I've come across at least 3 PDFs for the MQ-138, all having overlapping details but some missing from one another.The boards that these sensors come on seem to have no documentation. Basically, they all use an op amp IC as a voltage comparator, require 5V to heat the internals of the sensor, provide an analog pin with a 0.5 - 5V range relative to detectable gas levels, and a "digital" pin that is pretty much only useful merely detecting presence (perhaps for setting off an alarm). The threshold can be changed by adjusting a potentiometer on the board.Long story short, the MQ-138 and probably all MQ line sensors are essentially variable resistors. The MQ-138 is supposed to have a resistance range of 20k to 200k, 200k being no detectable level of gas to 20k being the highest level of detectable gas (anywhere between 500ppm and 3000ppm depending on the vapor, 500ppm for toluene). However, this is influenced by humidity and temperature, so you may want to account for this if you are trying to get a ppm reading.This sensor has a preheat time of 48 hours, meaning it should be left powered on for that amount of time to achieve the most consistent readings. However, this does *not* necessarily mean that the sensor won't be useful as soon as you power it on, or that you have to wait 48 hours every time you power it on. When not powered on for a long period of time, the resistance provided by the sensor component simply drifts. As soon as I powered on the sensor module for the first time, it was able to detect acetone on a cotton swab placed near it, so it has no problem detecting even if it hasn't been preheated. It's just about providing consistent resistance to work off of. If you power on the sensor once a day, even if only for minutes, then you shouldn't need to worry about preheating for hours on end. The PDF manual recommends that the sensor should be warmed up for no less than 48 hours if it hasn't been powered up in over a month, and 168 hours for 6 months. While preheating my MQ-138 sensor component I measured the resistance and it stopped changing after ~24 hours. YMMV In short, don't be worried about whether you can immediately start developing with this sensor.Actual warm-up time is something I've found to be negligible. Turn it on and it immediately returns expected values. Maybe wait just a minute to be sure the heating element is at its full temperature.Since I have electronics knowledge but rarely work on my own electronics there are big gaps in my knowledge, and I expected the DOUT pin of the board to produce a digital reading I could use with a GPIO pin on my Raspberry Pi. Well, it *is* digital, but in the sense that it is either on or off if the level of gas goes above a threshold. This wasn't exactly what I was looking for. The analog pin (AOUT) is what you want if you need to get an approximation of the actual level of gas.Unfortunate for my use case, the Raspberry Pi can't read voltage levels, so if you are planning on using this or any of the other MQ gas modules with one then you probably want to get an Arduino or an MCP3008, both of which can be used to convert analog to digital. Alternatively, if you can't wait for either of those things, you can remove the sensor component itself from the module and record the amount of time it takes to discharge a capacitor (> 900nF works fine) through it; the less time it takes to discharge, the more gas is present. If you do that, just make sure you find the PDF manual for it and use the correct pins. But should you wait for an actual analog to digital hardware? Most likely because it will provide more useful output and it should be somewhat more accurate (since Raspbian is not a realtime OS).The pinout on the sensor component itself is... interesting. Each side of the sensor has 3 pins. Both of the middle pins (2 and 5) are used to provide voltage to the heating element inside. Polarity doesn't matter with those. The other pins (1 and 3, 4 and 6) are used for detection. I don't know why there are 2 detection pins on each side, but what you want to do is connect the detection pins together on each side to form 1 lead so that you end up with 2 leads that you would provide current through. Why that is, I haven't the faintest clue. Some guy on YouTube did it that way for a different MQ sensor and the PDFs seem to suggest it should be done that way, so I figure why question it if it works. Again, polarity doesn't matter since it's effectively a variable resistor. Just don't run current between detector pins on the same side or it may damage the sensor (i.e. current should run from one side to the other, not between 1 and 3 or 4 and 6).What's nice about this and other MQ modules is that you can basically use it by itself since it has some LEDs mounted on the board. Just provide VCC and GND and one of the LEDs will turn on if enough VOC vapor is present. This is adjustable with a pot on the board.Is it super accurate? That remains to be seen. Is it super useful? Potentially, yes, if what you want is to track changes to VOC levels in real-time and draw a graph, which is what I want to do, the MQ-138 works sufficiently. Or if you just want to set off an alarm at a certain gas threshold.