@vvcstemplay Vickram Crishna

Electric kitchen devices, such as microwave ovens, almost always have a beeper fitted, to let users know that the heating/cooking cycle is over.

Of course, this is trivial, because they work on a timer in the first place.

In the other hand, fossil fuel devices, such as ovens and heaters, do not have timers, and, chicken and egg, alerts. Overcooked food is the subject of many jokes.

However, untended gas cookers are a home fire hazard, apart from wasting fuel and money and creating climate change effluents unnecessarily (meaning, while not even providing the benefit of cooking).

A simple, and hopefully inexpensive, heat sensitive device, that might detect an open flame, or might detect that a flame has been burning in excess of some ‘normal’ time, could be very useful.

A set of bimetallic elements could be used to detect the heat of burners that have been lit, and this could be linked to an electronic timekeeper to sound off an alarm, which could be a buzzer or beeper. The power source could be batteries or mains power.

@jtd Jude T Dsouza

Haha. Milk. Damned liquid spills on the gas stove. Especially when you imagine you are watching it while reading the papers.

@vvcstemplay Vickram Crishna

Of course, when it is milk, almost anyone will notice fairly quickly something is wrong. The scent of suspicion will be far sharper than any mere needle.

But what if water boiled over and extinguished the flame? The odour of mercaptan is not strong enough to alert somebody in the next room, until too much gas has leaked, and might be filled in low level pockets in the kitchen and neighbouring rooms/passages.

So, a necessary addition to a good gas alert system must also detect the flame has gone out, while the gas valve is still open.

Some thought needed on how this will work. A simple heat detector is not good enough, as the metal of the burner may remain hot enough to fool it for some time, while not being hot enough to reignite the gas flame.

@jtd Jude T Dsouza

UV detection is the method used.

@tnana Anant Sujatanagarjuna

MQ-2 Gas sensors can be used with arduino (found here).
Another sensor , such as a thermal camera such as this one, could possibly detect spillage? Since when something spills, the radius of “hot stuff”, would increase to areas that are pre-designated to the camera, to be not as hot as the flame.

@vvcstemplay Vickram Crishna


is a good idea. Although spills by themselves may not be hazards, they may either lead to a hazard soon or simply be waste. Either way, a sensor set plus alarm could incorporate such a function.

If the thermal imaging sensor performs all three functions (detecting a doused flame, a non flammable spill and a flammable spill) it may be cost effective.

Only, the intelligence needs to be built in a small enough device to be clearly inexpensive, without the need to engineer a special purpose chip.

° Apart from detecting the flame parameters, the device must provide a timer and an alarm, to cover all the needs of an add-on emergency annunciator for a gas stove.
° Settings for the timer need to be accessible, for the operator to be able to walk away after putting something on to cook that needs lengthy heating with no supervision, such as a pressure cooker.
° The switching of the device can be retrofitted to the main gas supply valve near the stove, so that it only comes on when the stove is in use.

@vvcstemplay Vickram Crishna

On checking, the thermal camera
turns out to cost nearly Rs 3k. That’s a heavy expenditure for a gas stove alert.

Bimetallic switches seem to be far cheaper, less than a quarter the price, in ready-made housing. A typical stove will need 4 of these for a single alert system, so as much as a single camera.

However, if the design allows for the camera to sequentially scan each burner in turn, it pays for itself. Since it is capable of detecting the flame, a spill and a flammable spill, it is also much more versatile than a bimetallic sensor.

The total system still appears to be quite expensive, but compared to the cost of a single out of control kitchen fire, it may well be worth it.

@jtd Jude T Dsouza

Perhaps a lens mounted on the sensor would provide an angle covering all burners. Also mounting the sensor on one corner underneath should help.
Speaking of infrared, I have a few hundred infrared filter coated photodiodes used in PIRs. At Rs. 100 each we can use one per burner.

@vvcstemplay Vickram Crishna

So, clearly, it isn’t necessary to ‘look’ at all the burners all the time. Only the ones that are deliberately lit. When they are, the field of the camera that ‘sees’ the burner gets activated. The controller will have less work to do, and undoubtedly will do it better.

Can we list the services possible with an infrared diode and that by a thermal imaging camera? It seems to me that the camera is actually an array of infrared diodes, no? If 4 diodes are normally much much cheaper, by a huge factor, than the camera, the design choice becomes clear.

@jtd Jude T Dsouza

Having any mechanical movement for the camera would be failure prone and hard to implement.

Indeed. But fabricated a few \mu meters apart, alongwith some analog amplifiers on the same die. . The discrete diodes are 10 \phi mm.

They are.

The gas stove is a very splashy, sticky environment though. I am wondering about reliability and maintenance. Keeping the grime off the optical window will be quite a task.

@vvcstemplay Vickram Crishna

So, I’m thinking, the sensor device form factor will be a narrow rectangular plate mounted behind the stove. It can be mounted either flat on the wall, or attached to the stove itself, if there is no wall conveniently (reliably) close behind. It should be easily removable, for maintenance and cleaning.

Probably not more than a cm thick, so the power cord needs to be fixed with a flat connector. Instead of single diodes for each burner, probably 2 or 3. That way, if one gets obscured by grease, the controller can start taking readings off a neighbour, which might otherwise be used for sensing a different burner. If several get obscured, the alert should go off the moment one burner is lit, so that the user knows it’s time to do the maintenance.

I’m just assuming that sort of smart behaviour can be programmed into the controller, which now becomes, probably, the most expensive part of the solution. Even so, as I’m learning from this and other discussions here, fairly capable microprocessors are available at ridiculously low prices, yet powerful enough to run a dedicated purpose device like this very reliably.