💬 Gas Sensor

  • Admin

    This thread contains comments for the article "Gas Sensor" posted on MySensors.org.

  • Just a note for any Domoticz users who may try this sketch - it presents in the log, but is not available under 'Devices' to be added UNTIL it gets a reading > 0. My solution to that was to go put the node over the gas hob and let the gas run (unlit) for a few seconds so that it generated a reading 😄

    Remember kids, gas is dangerous, stay safe.

  • Where can I get chep MQ-138 sensor please?

  • Mod

    Have you looked at aliexpress?

  • HI Gohan,

    Yes I did, seems that they are very expensive wherever I look. I'll wait on this one ofr now.

    Thanks for the suggestion though 🙂

  • Mod

    I believe it is on the cheap side. These price ranges are mainly for hobbyist sensor. The real calibrated ones are usually way more expensive. The cheap ones are only good to measure trends, not for the actual value.

  • Please advise what would be curve for MQ135 sensor for CO2. Thank you.

  • Mod

    @apl2017 itsn't the curve in the datasheet linked on the build page sifficient?

  • Sure, the curve is there. I am trying to make sense out of two points curve that is used in sketch for MQ2 sensor, CO gas. "point1: (lg200, 0.72), point2: (lg10000, 0.15)" while I see on the curve from datasheet (lg200, 5.2) and (lg10000, 1.3). That is why I am asking for help, sorry for not making it clear.

  • Lack of support makes you think, so I figured out the "hidden" calculations in the sketch above and proposing the code modification to allow users to port the code to a different gas sensor using data taken directly from the sensor chart. Here are modified fragments of the code showing only CO curve for MQ2 sensor:

    float CO2Curve[4] = {0.8,200,2.2,10};
    //two end points are taken from the curve with two coordinates for each point
    //in form of{Y1, X1,Y2, X2) where X is PPM axis, Y is Rs/Ro axis from the sensor chart for specifc gas
    //then equation X=(Y-b)/m is used,
    //where m is slope of the curve, b is Y axis intersept point
    float Slope = (log(pcurve(1)-log(pcurve(0))/(log(pcurve(3)-log(pcurve(2));
    float Y_Intercept=log(pcurve(1)-Slope*log(pcurve(3);

    int MQGetPercentage(float rs_ro_ratio, float *pcurve)
    { return (pow(10,((log(rs_ro_ratio)-Y_Intercept)/Slope))); }

  • Mod

    I have found this sketch for MQ7 sensors

    This code was developed by the_3d6 from Ultimate Robotics (http://ultimaterobotics.com.ua).
    License: you can use it for any purpose as long as you don't claim that you are its author and
    you don't alter License terms and formulations (lines 1-10 of this file).
    You can share modifications of the code if you have properly tested their functionality,
    including confirming correct sensor response on CO concentrations of 0-30ppm, 100-300ppm and 1000-10000ppm.
    If you can improve this code, please do so!
    You can contact me at aka.3d6@gmail.com
    //WARNING! Each sensor is different!
    //You MUST calibrate sensor manually and
    //set proper sensor_reading_clean_air value before using
    //it for any practical purpose!
    int time_scale = 8; //time scale: we altered main system timer, so now all functions like millis(), delay() etc 
    					//will think that time moves 8 times faster than it really is
    					//to correct that, we use time_scale variable:
    					//in order to make delay for 1 second, now
    					//we call delay(1000*time_scale)
    void setTimer0PWM(byte chA, byte chB) //pins D5 and D6
    	TCCR0A = 0b10110011; //OCA normal,OCB inverted, fast pwm
    	TCCR0B = 0b010; //8 prescaler - instead of system's default 64 prescaler - thus time moves 8 times faster
    	OCR0A = chA; //0..255
    	OCR0B = chB;
    void setTimer2PWM(byte chA, byte chB) //pins D11 and D3
    	TCCR2A = 0b10100011; //OCA,OCB, fast pwm
    	TCCR2B = 0b001; //no prescaler
    	OCR2A = chA; //0..255
    	OCR2B = chB;
    void setTimer1PWM(int chA, int chB) //pins D9 and D10
    	TCCR1A = 0b10100011; //OCA,OCB, fast pwm
    	TCCR1B = 0b1001; //no prescaler
    	OCR1A = chA; //0..1023
    	OCR1B = chB;
    float opt_voltage = 0;
    byte opt_width = 240; //default reasonable value
    void pwm_adjust()
    //this function tries various PWM cycle widths and prints resulting
    //voltage for each attempt, then selects best fitting one and
    //this value is used in the program later
    	float previous_v = 5.0; //voltage at previous attempt
    	float raw2v = 5.0 / 1024.0;//coefficient to convert Arduino's 
    							   //analogRead result into voltage in volts
    	for (int w = 0; w < 250; w++)
    		setTimer2PWM(0, w);
    		float avg_v = 0;
    		for (int x = 0; x < 100; x++) //measure over about 100ms to ensure stable result
    			avg_v += analogRead(A1);
    		avg_v *= 0.01;
    		avg_v *= raw2v;
    		Serial.print("adjusting PWM w=");
    		Serial.print(", V=");
    		if (avg_v < 3.6 && previous_v > 3.6) //we found optimal width
    			float dnew = 3.6 - avg_v; //now we need to find if current one
    			float dprev = previous_v - 3.6;//or previous one is better
    			if (dnew < dprev) //if new one is closer to 1.4 then return it
    				opt_voltage = avg_v;
    				opt_width = w;
    			else //else return previous one
    				opt_voltage = previous_v;
    				opt_width = w - 1;
    		previous_v = avg_v;
    float alarm_ppm_threshold = 100; //threshold CO concentration for buzzer alarm to be turned on,
    float red_threshold = 40; //threshold when green LED is turned on red turns on
    float reference_resistor_kOhm = 10.0; //fill correct resisor value if you are using not 10k reference
    float sensor_reading_clean_air = 600; //fill raw sensor value at the end of measurement phase (before heating starts) in clean air here! That is critical for proper calculation
    float sensor_reading_100_ppm_CO = -1; //if you can measure it 
    									  //using some CO meter or precisely calculated CO sample, then fill it here
    									  //otherwise leave -1, default values will be used in this case
    float sensor_100ppm_CO_resistance_kOhm; //calculated from sensor_reading_100_ppm_CO variable
    float sensor_base_resistance_kOhm; //calculated from sensor_reading_clean_air variable
    byte phase = 0; //1 - high voltage, 0 - low voltage, we start from measuring
    unsigned long prev_ms = 0; //milliseconds in previous cycle
    unsigned long sec10 = 0; //this timer is updated 10 times per second,
    						 //when it will overflow, program might freeze or behave incorrectly. 
    						 //It will happen only after ~13 years of operation. Still, 
    						 //if you'll ever use this code in industrial application,
    						 //please take care of overflow problem 
    unsigned long high_on = 0; //time when we started high temperature cycle
    unsigned long low_on = 0; //time when we started low temperature cycle
    unsigned long last_print = 0; //time when we last printed message in serial
    float sens_val = 0; //current smoothed sensor value
    float last_CO_ppm_measurement = 0; //CO concentration at the end of previous measurement cycle
    float raw_value_to_CO_ppm(float value)
    	if (value < 1) return -1; //wrong input value
    	sensor_base_resistance_kOhm = reference_resistor_kOhm * 1023 / sensor_reading_clean_air - reference_resistor_kOhm;
    	if (sensor_reading_100_ppm_CO > 0)
    		sensor_100ppm_CO_resistance_kOhm = reference_resistor_kOhm * 1023 / sensor_reading_100_ppm_CO - reference_resistor_kOhm;
    		sensor_100ppm_CO_resistance_kOhm = sensor_base_resistance_kOhm * 0.5;
    		//This seems to contradict manufacturer's datasheet, but for my sensor it 
    		//looks quite real using CO concentration produced by CH4 flame according to
    		//this paper: http://www.iafss.org/publications/fss/8/1013/view/fss_8-1013.pdf 
    		//my experiments were very rough though, so I could have overestimated CO concentration significantly
    		//if you have calibrated sensor to produce reference 100 ppm CO, then
    		//use it instead    
    	float sensor_R_kOhm = reference_resistor_kOhm * 1023 / value - reference_resistor_kOhm;
    	float R_relation = sensor_100ppm_CO_resistance_kOhm / sensor_R_kOhm;
    	float CO_ppm = 100 * (exp(R_relation) - 1.648);
    	if (CO_ppm < 0) CO_ppm = 0;
    	return CO_ppm;
    void startMeasurementPhase()
    	phase = 0;
    	low_on = sec10;
    	setTimer2PWM(0, opt_width);
    void startHeatingPhase()
    	phase = 1;
    	high_on = sec10;
    	setTimer2PWM(0, 255);
    void setLEDs(int br_green, int br_red)
    	if (br_red < 0) br_red = 0;
    	if (br_red > 100) br_red = 100;
    	if (br_green < 0) br_green = 0;
    	if (br_green > 100) br_green = 100;
    	float br = br_red;
    	br *= 0.01;
    	br = (exp(br) - 1) / 1.7183 * 1023.0;
    	float bg = br_green;
    	bg *= 0.01;
    	bg = (exp(bg) - 1) / 1.7183 * 1023.0;
    	if (br < 0) br = 0;
    	if (br > 1023) br = 1023;
    	if (bg < 0) bg = 0;
    	if (bg > 1023) bg = 1023;
    	setTimer1PWM(1023 - br, 1023 - bg);
    void buzz_on()
    	setTimer0PWM(128, 128);
    void buzz_off()
    	setTimer0PWM(255, 255);
    void buzz_beep()
    	byte sp = sec10 % 15;
    	if (sp == 0)
    	if (sp == 1)
    	if (sp == 2)
    	if (sp == 3)
    	if (sp == 4)
    	if (sp == 5)
    void setup() {
    	//WARNING! Each sensor is different!
    	//You MUST calibrate sensor manually and
    	//set proper sensor_reading_clean_air value before using
    	//it for any practical purpose!
    	pinMode(5, OUTPUT);
    	pinMode(6, OUTPUT);
    	pinMode(3, OUTPUT);
    	pinMode(9, OUTPUT);
    	pinMode(10, OUTPUT);
    	pinMode(A0, INPUT);
    	pinMode(A1, INPUT);
    	setTimer1PWM(1023, 0);
    	Serial.print("PWM result: width ");
    	Serial.print(", voltage ");
    	Serial.println("Data output: raw A0 value, heating on/off (0.1 off 1000.1 on), CO ppm from last measurement cycle");
    	//beep buzzer in the beginning to indicate that it works
    	delay(100 * time_scale);
    	delay(100 * time_scale);
    	delay(100 * time_scale);
    	delay(100 * time_scale);
    	delay(100 * time_scale);
    	delay(100 * time_scale);
    	startMeasurementPhase(); //start from measurement
    void loop()
    	//WARNING! Each sensor is different!
    	//You MUST calibrate sensor manually and
    	//set proper sensor_reading_clean_air value before using
    	//it for any practical purpose!
    	unsigned long ms = millis();
    	int dt = ms - prev_ms;
    	//this condition runs 10 times per second, even if millis() 
    	//overflows - which is required for long-term stability
    	//when millis() overflows, this condition will run after less
    	//than 0.1 seconds - but that's fine, since it happens only once
    	//per several days
    	if (dt > 100 * time_scale || dt < 0)
    		prev_ms = ms; //store previous cycle time
    		sec10++; //increase 0.1s counter
    		if (sec10 % 10 == 1) //we want LEDs to blink periodically
    			setTimer1PWM(1023, 1023); //blink LEDs once per second
    									  //use %100 to blink once per 10 seconds, %2 to blink 5 times per second
    		else //all other time we calculate LEDs and buzzer state
    			int br_red = 0, br_green = 0; //brightness from 1 to 100, setLEDs function handles converting it into timer settings
    			if (last_CO_ppm_measurement <= red_threshold) //turn green LED if we are below 30 PPM
    			{//the brighter it is, the lower concentration is
    				br_red = 0; //turn off red
    				br_green = (red_threshold - last_CO_ppm_measurement)*100.0 / red_threshold; //the more negative is concentration, the higher is value
    				if (br_green < 1) br_green = 1; //don't turn off completely
    			else //if we are above threshold, turn on red one
    				br_green = 0; //keep green off
    				br_red = (last_CO_ppm_measurement - red_threshold)*100.0 / red_threshold; //the higher is concentration, the higher is this value
    				if (br_red < 1) br_red = 1; //don't turn off completely
    			if (last_CO_ppm_measurement > alarm_ppm_threshold) //if at 50 seconds of measurement cycle we are above threshold 
    			setLEDs(br_green, br_red); //set LEDs brightness
    	if (phase == 1 && sec10 - high_on > 600) //60 seconds of heating ended?
    	if (phase == 0 && sec10 - low_on > 900) //90 seconds of measurement ended?
    		last_CO_ppm_measurement = raw_value_to_CO_ppm(sens_val);
    	float v = analogRead(A0); //reading value
    	sens_val *= 0.999; //applying exponential averaging using formula
    	sens_val += 0.001*v; // average = old_average*a + (1-a)*new_reading
    	if (sec10 - last_print > 9) //print measurement result into serial 2 times per second
    		last_print = sec10;
    		Serial.print(" ");
    		Serial.print(0.1 + phase * 1000);
    		Serial.print(" ");

    from this guide http://www.instructables.com/id/Arduino-CO-Monitor-Using-MQ-7-Sensor/
    I was wondering if it could be used on a MySensors node even with all those timers modifications

  • Does anyone know if any of these sensors can be used to detect vaporized white mineral oil?

  • Mod

    Vaporized oil is very messy, I don't think they would any good as it is not volatile

  • Does it change things if it's actually atomized and not vaporized?

  • Mod

    What I'm saying is that it is not a gas that dissolves in air, it is more particulate matter suspended in the air so an optical sensor is probably better given that the sensor could last as oil tends to stick to stuff

  • Contest Winner

    I spent the last few days trying to get a sense out of the black magic behind the calculations used for this sensor.
    I got the math now but still there is something I feel like it is not completely accurate with the current implementation.
    First of all the resulting ppm seems like a sort of normalized value; since I know e.g. CO2 ppm in clean air is around 400, the code seems not to take this into consideration. Also, configure it with a different MQ sensor doesn't seem an easy task.
    For these reasons I've tried starting from scratch, inspired by http://davidegironi.blogspot.com/2014/01/cheap-co2-meter-using-mq135-sensor-with.html#.WyLQo6qFNn5.
    Starting point is the power function y = a*x^b. The way to make the code more generic is to let the user provide the coordinates of two points (like @APL2017 was suggesting a while ago) which is an easy task and let the code solve the two equations and derive the values of a and b. Then, since ppm = a(rs/ro)^b, with a known value of ppm (e.g. the concentration in clear air of the gas, for co2 is 411), the equation can be solved for Ro by measuring Rs from the adc. Once a, b and Ro are known, the ppm comes naturally by solving again ppm = a(rs/ro)^b.
    I'm not sure this is better than the other methods but at least I get the same results from the blog above, both in terms of values of a, b and Ro as well as a real value of CO2 ppm measured with a MQ135.
    The downside of this of course is the difficulty to provide a known value of ppm for the calibration for other gas like e.g. Ch4, but if I claim I'm showing a ppm value, I want to be sure this is a real ppm 🙂
    The code is here, within the dev branch of NodeManager, any feedback would be appreciated!

  • Hi all,
    which of the above sensors (listed in https://www.mysensors.org/build/gas) is the most suitable to detect a relevant leak of gasoline inside an engine bay?

  • Mod

    I'd say the ones with benzene

Log in to reply

Suggested Topics

  • 3
  • 2
  • 10
  • 584
  • 347
  • 163