michalrz
Supreme [H]ardness
- Joined
- Jun 4, 2012
- Messages
- 4,260
* Work in progress, see post 2 *
Hello everyone,
I've been slowly making a thing for a while now. As I have some spare time, some money, etc.
Basically what I'm building is a cheap but quite robust wireless environment sensing and alarming device.
Operating principle:
1. Detect a fire by sampling multiple sensors. Try to establish the likelihood of false alarm/sensor malfunction/actual fire.
- is it getting more hot than usual?
- optical detection of an open flame via infrared?
- is there smoke?
- smoke detection by checking both the MQ-2 sensor's readingAND a simple diode emitter-phototransistor setup
-(added 26.04) (removed 12.06, see below) measure supply voltage, notify in event of insufficient power
2. Detect presence of combustible gases (only tested smoke and Buthane for now, using MQ-2)
2a. (added 26.04) MQ-2 is a tube-like sensor. It needs to warm up for a few minutes and have adequate current available.
During startup, current draw to MQ-2 will be measured to determine sensor readiness and a baseline readout will be established (calibration).
(added 01.05) - measuring the MQ-2 sensor's current draw isn't helpful for determining readiness because power supply fluctuations and humidity affect the readout more than the heating up... But a baseline will be recorded.
3. Motion detection (optional)
- motion detection via passive infrared (PIR) sensors
- multiple sensors, partially overlapping to help identify false alarms
4. Shit's on - tell b0ss
- the ESP8266 by default supports WiFi.
- on alarm, send e-mail, detail which sensors tripped
- (21st May - removed - complex, expensive, harder to obtain)send SMS message (there are GSM/GPRS modules available for Arduino, ESP8266, etc.)
- (21st May - removed - complex, expensive, harder to obtain)send SMS message via an API / SMS gateway on the interwebs
- buzzer buzzes. Will adjust signal based on how real the shit has gotten (multiple sensors tripped?)
4a. Level of paranoia (panic thresholds) can be configured.
5. And here's the kicker.
- total cost of device - around $30-$40.
- it's so cheap because all the mentioned parts are very inexpensive and can be had for peanuts.
They aren't obviously pr0 grade calibrated certified etc, but using multiple methods of sensing the environment makes up for that.
- because it's cheap, multiple devices can be used and there's no need for a compromise (which room? which corridor?)
Current parts list:
- a Wemos D1 development board, based on the ESP8266 microcontroller ($5)
- OLED 0,91" display ($5)
- MQ-2 LPG, Smoke, Alcohol, Propane, Hydrogen, Methane and Carbon Monoxide sensor ($2)
- optical fire sensing detector (760 to 1100 nm) ($3)
- AM2302 based temperature and humidity sensor ($5?)
- (removed 21st May - complex, expensive, harder to obtain)GSM module - $10 (?)
- backup power in the form of any small power bank ($8?)
- wall adapter to charge/top off the power bank
- wires and a plastic box ($5?)
- (removed 12.06-cost, can be done with an analog2digital channel, ESP8266 has just one, ESP32 has more for a similar price)INA-219 digital current & voltage sensor module ($4)
- (added 13.05) Keyes KY-006 passive buzzer (volume adjustable in software via PWM)
Hello everyone,
I've been slowly making a thing for a while now. As I have some spare time, some money, etc.
Basically what I'm building is a cheap but quite robust wireless environment sensing and alarming device.
Operating principle:
1. Detect a fire by sampling multiple sensors. Try to establish the likelihood of false alarm/sensor malfunction/actual fire.
- is it getting more hot than usual?
- optical detection of an open flame via infrared?
- is there smoke?
- smoke detection by checking both the MQ-2 sensor's reading
-
2. Detect presence of combustible gases (only tested smoke and Buthane for now, using MQ-2)
2a. (added 26.04) MQ-2 is a tube-like sensor. It needs to warm up for a few minutes and have adequate current available.
(added 01.05) - measuring the MQ-2 sensor's current draw isn't helpful for determining readiness because power supply fluctuations and humidity affect the readout more than the heating up... But a baseline will be recorded.
- motion detection via passive infrared (PIR) sensors
- multiple sensors, partially overlapping to help identify false alarms
4. Shit's on - tell b0ss
- the ESP8266 by default supports WiFi.
- on alarm, send e-mail, detail which sensors tripped
- (21st May - removed - complex, expensive, harder to obtain)
- (21st May - removed - complex, expensive, harder to obtain)
- buzzer buzzes. Will adjust signal based on how real the shit has gotten (multiple sensors tripped?)
4a. Level of paranoia (panic thresholds) can be configured.
5. And here's the kicker.
- total cost of device - around $30-$40.
- it's so cheap because all the mentioned parts are very inexpensive and can be had for peanuts.
They aren't obviously pr0 grade calibrated certified etc, but using multiple methods of sensing the environment makes up for that.
- because it's cheap, multiple devices can be used and there's no need for a compromise (which room? which corridor?)
Current parts list:
- a Wemos D1 development board, based on the ESP8266 microcontroller ($5)
- OLED 0,91" display ($5)
- MQ-2 LPG, Smoke, Alcohol, Propane, Hydrogen, Methane and Carbon Monoxide sensor ($2)
- optical fire sensing detector (760 to 1100 nm) ($3)
- AM2302 based temperature and humidity sensor ($5?)
- (removed 21st May - complex, expensive, harder to obtain)
- backup power in the form of any small power bank ($8?)
- wall adapter to charge/top off the power bank
- wires and a plastic box ($5?)
- (removed 12.06-cost, can be done with an analog2digital channel, ESP8266 has just one, ESP32 has more for a similar price)
- (added 13.05) Keyes KY-006 passive buzzer (volume adjustable in software via PWM)
Attachments
Last edited: