Maker Pro
Maker Pro

Choosing an IR photodiode or phototransistor for fire detection

PhiBiSoLa

Sep 17, 2012
2
Joined
Sep 17, 2012
Messages
2
Hello,

A couple of people and I are trying to build a smart fire detector. We want to use a couple of IR sensors to detect a hot stove element (with or without a fire burning on top of it) a couple of meters away (It's going to be stuck to the ceiling), while still being able to detect if a candle is lit up or not, 10 meters away (hopefully).

Our research tells us that carbon based fire irradiates IR rays at about 4.3um, so the sensor would most favorably detect rays in that range.

Any parallel information to the subject is appreciated, thanks!
Phil
 

wingnut

Aug 9, 2012
255
Joined
Aug 9, 2012
Messages
255
You could use PIR or pyroelectric infrared sensors. These only detect changing IR such as a flame flickering.

Or you could use a phototransistor. Some come with infrared filters. Or you could cover a regular phototransistor with over-exposed film negative to make a filter. The phototransistor is connected to a comparator

The TEFT4300 is black and daylight blocking working best in the 900nm to 950nm range, which is a shorter wavelength than you are looking for.
The SFH309 works best at 380-1180nm, which also misses your target. Looking up a few cheap IR phototransistors, they all seem to operate in this wavelength.

Most phototransistors have a limited angle they detect, so I doubt they could look at a stove and a candle.

Have you considered a cheap smoke detector?
 

GreenGiant

Feb 9, 2012
842
Joined
Feb 9, 2012
Messages
842
The problem you will run into with an IR sensor is you are going to have to fine tune it for each and every installation, humans emit IR, sunlight emits IR, heat emits IR

someone walking by after walking into the house from outside they are going to be a massive IR source, just out of the shower, same thing, sunlight through a window, etc

even dealing with IR change you are going to be setting it off constantly if you walk by it

you would probably be better off with either a camera that outputs the data of what the max temperatures it is reading at the time and then flag at say... 150 deg?
 

gorgon

Jun 6, 2011
603
Joined
Jun 6, 2011
Messages
603
I know that the flame detector our company make, use a small low resolution camera chip to 'see' the flames. In addition to some heavy computing on the 'picture'.


TOK ;)
 

PhiBiSoLa

Sep 17, 2012
2
Joined
Sep 17, 2012
Messages
2
Wow, thanks for the replies!

We are indeed using a camera. It is the main detector: we want the IR sensors as backups. While I will probably buy a couple of PIR and pyroelectrics to test, the reason we selected the photodiodes is because of their cost. We don't want the cost of those things escalating past a dollar, and since we have 4, we would be looking a 30cents a piece each.

The way we would be cancelling other IR sources is by having the 3 other sensors as references. They would also be calibrated to ignore smaller, less intense sources, such as humans. Again, we are unsure about all of this. We will have to test, test, test.

I will keep you posted for the other sensors if I do check them out, but as for the photodiodes and phototransistors:

We are trying to stay low power, so what would be the best option: photodiode amplified with a low power opamp, or a phototransistor that needs to be amplified less with said opamp?

Thanks a lot for all your replies!
 
Top