# Looking for an inexpensive Smoke Detection Sensor

As the title says, I am looking for a smoke detection sensor. Took a look from SparkFun to Mouser and did not find anything. Is there a part you would recommend? I would prefer something that interfaces with the ATMEGA168/ATMEGA328 via a single wire or two wire interface. Would prefer to avoid a SPI or Serial interface as my circuit already utilizes those pins.

Bumped into a few sensors on Ebay but they were $15+. Is this normal? Is there a cheaper alternative that works great? Should I just buy a cheap smoke alarm and extract the sensor that way? Thanks in advance for your help. ## 5 Answers The MQ-2 Flammable Gas/Smoke sensor might be what you are looking for. Datasheet: http://www.pololu.com/file/download/MQ2.pdf?file_id=0J309 Here is one source: http://www.pololu.com/catalog/product/1480 They also have a carrier board to make it easy to interface. It does not provide a digital interface, but provides an analog voltage that could easily be read by the ATMEGA168/ATMEGA328 I'd go for the cheap smoke alarm. The actual sensor is a black plastic chamber with a LED and a photo sensor. It is constructed so that with clean air no light reaches the sensor, it passes the center of the chamber and is absorbed on another side. With smoke particles in the air some light is reflected towards the sensor and off goes the alarm. As for the electronics, I opened up one such alarm and found an MC145010. Well, if you have some time to test, you can go for LED + Photodiode. Should be <1$ :-)

The idea is that LED should not shine directly on photodiode, air should be able to fly around, but not light.

In ultrapure air photodiode should see 0, the more smoke in the air - the more diffused light will get to diode, and you'll be able to measure it via ADC.

But this definitely require some testing.

Personally I would go for UV diode, not IR/visible one, as there is less UV light noise (except xenon flashlights from cameras, be sure to test & filter for it). This surely require some experimenting.

As a side note, I love to see air quality in beam of 100mw green lazer, but this is definitely not cost effective :-D

• Very little external light (noise) gets into photo-interruptor modules. – tyblu Jan 9 '11 at 8:23
• The problem is that there IR LED shines directly into photodiode, so difference with and without smoke is some 0.001% :-) When detecting diffused light, you need to differentiate between 0 and 0.001% which is much much easier :-) – BarsMonster Jan 9 '11 at 8:53
• That is true. – tyblu Jan 9 '11 at 9:15

If you're looking for the optical type, it looks like one can use an analog infrared photodiode/sensor and emitter module along with a controller, like this one.

• So wait, are you telling me that smoke alarms you buy off of the shelf measure light to determine how 'smoke' is in the air? Guess I am lost. Figured this would be a different type of sensor. – ThaKidd KG5ORD Jan 9 '11 at 8:36
• Check here. en.wikipedia.org/wiki/Smoke_detector Optical ones are way easier to implement. Ionization probably a bit better, but they cannot be cheap, as they need some radioactive elements. – BarsMonster Jan 9 '11 at 8:54

Microchip make the required ICs.

For photo detection

For ionisation detection