Friday March 20, 2009

An indicator light for gas stovetops

Leaving a burner turned on after cooking has finished is wasteful of energy and unsafe, but it is also an easy mistake to make. A harried or forgetful cook might well turn off the lights and leave the kitchen, leaving a burner going for hours. For this reason, electric stovetops have indicator lights which glow when a burner is on.

However, gas stovetops have no such indicators. Adding one after the fact is likely to be difficult and might well be dangerous: modifying gas appliances is not a safe hobby for amateurs. Can we fashion a solution that is reliable, safe, convenient, etc?


The big problem is detection: how can we tell that a gas burner is on? Modifying the stovetop's valves to add switches is not a safe idea. Anything we mount on the stovetop's surface is going to be a nuisance to clean, etc. A pressure or flow detector might work, but a burner set on low doesn't use all that much gas. Fortunately, there are other options.

We can simplify the problem by disregarding the case where a burner is turned on, but unlit: most cooks will quickly notice the "natural gas smell" from an unlit burner. Consequently, we can assume that the burner is lit. (Actually, natural gas is odorless, but utility companies add an odorant.)

A lit burner affects its environment in various ways (eg, combustion byproducts, light). The easiest of these to detect is light: An optical sensor can detect a flame from several feet away. So, with the kind assistance of physicist Doug McNutt (The MacNauchtan Laboratory), I looked into this possibility.


My initial notion was to use an infra-red (IR) sensor. Either a burner or a heated pan will exceed 130 degrees (Fahrenheit) and an IR sensor could look for radiation in the corresponding frequency range. This is not a perfect solution, however, for a couple of reasons.

First, it takes a substantial amount of time for a pan to heat or cool; I'd prefer a system that responds immediately. More critically, both sunlight and artificial light contain energy in this frequency range. This "background noise" could confuse the sensor, reducing its sensitivity and/or producing "false positives".


Fortunately, there is a solution that avoids these problems. Gas flames generate substantial amounts of ultra-violet (UV) radiation, at frequencies that are not found in either sunlight or artificial light:

  • Raw sunlight contains a wealth of frequencies, but the Earth's upper atmosphere filters many of them out. In particular, it creates a "solar blind" gap in the UV spectrum, just where we need it.

  • Artificial light isn't a problem, either. Conventional incandescent light bulbs don't get hot enough to produce UV radiation. Fluorescent tubes produce UV internally, but the internal phosphor coating captures part of this (turning it into visible light); the surrounding glass prevents the rest from escaping.
Finally, UV reflects nicely from the kinds of surfaces (eg, glass, metal, tile) that are typically found in kitchens. If our sensor is sensitive enough to work from reflected light, it doesn't even have to "see" the flames directly.


Looking around on the web, I found a number of UV-based flame detectors. Unfortunately, most of these were designed for use in laboratories or industrial applications. So, they were far outside of the available budget, looked difficult to apply, etc.

Eventually, however, I found a very nice detector and driving circuit, made by Hamamatsu Photonics. The combination was relatively inexpensive (about $80) and no minimum quantities were required:

These are ultraviolet sensors that utilize the photoelectric effect of metal and gas multiplication effect. Flame sensors have high sensitivity and a wide field of view (directivity) that allow them to swiftly and reliably detect the weak ultraviolet rays emitted from flames (for example, flame from cigarette lighter more than 5 meters away). This feature makes them ideal as flame detectors and fire alarms.

Flame Sensor (UV TRON) R2868

UV TRON Driving Circuit C3704 Series


The testing, fabrication, and installation of the device were performed by Ed Morin. Esthetic details were supervised by Georgina Morin, whose kitchen was being used as the installation site.

The testing was very simple: hook up the sensor and electronics to a power source and indicator light, then see whether it detected a gas flame reliably. As promised by Hamamatsu, the sensor was quite sensitive: any level of gas flame could be detected from several feet away.

Packaging and assembly

A low-profile, black plastic box was used to house the detector and electronics. The box, about three inches square and an inch tall, had room for the sensor bulb (lying down) and the driver board. In order to get the sensor bulb to lie next to the driver board, the leads were extended a bit.

Note: The sensor bulb's leads should be insulated and kept as short as possible. Hamamatsu recommends a maximum length of 5 cm.

Ed wired the power source (a 10 VDC "wall wart") directly into the box. It might have been possible to find a mating socket for the wall wart's output cable, but pluggability wasn't really a concern in a semi-permanent installation. Besides, there wasn't enough room in the box for a socket.

Note: If you decide to use your own plug and socket for the DC power connection, be sure to use a polarized, female plug.

Safety Window

The sensor bulb is fragile and the driving circuit generates a few hundred volts. Clearly, these parts need to be protected from the cook (and vice versa :-). So, some sort of "safety window" seemed advisable.

Unfortunately, most "clear" plastic or glass sheets are quite opaque to UV. After looking into sapphire, pyrex, and other possibilities, I found a relatively inexpensive source of fused quartz sheet (~$15 for a 2" square, 1/16" thick) I was rather dismayed, however, to find that the (unannounced) shipping charge had added another $14 to the cost. Grumble...

Technical Glass Products, Inc.
2X2X0.062 Ground & Polished Plates
This material can probably be cut with an ordinary glass cutter, but it's thin enough that a diamond scribe or cutting wheel is really a better choice. To be on the safe side, you might want to take the sheet to a local glass shop for cutting. Once you have the sheet cut to the desired size and shape, you can glue it in place using (say) silicone caulk.


The driver board provides an "open collector" output, based on an NPN power transistor. This can "sink" (ie, connect to ground) up to 100 mA, which is quite sufficient to drive a small indicator light, relay, etc. The driving circuit is very tolerant, handling 10-30 VDC; many wall warts (eg, 10 VDC @ 100 mA) can meet this requirement.

The driver board emits a pulse whenever the sensor detects UV, but the amount of UV emitted by a gas flame is quite variable. So, when Ed performed his initial testing, the light flickered instead of staying on. Following a suggestion in the technical note, Ed added a small capacitor (0.5 mfd; 15 VDC) to the driver board. This extended the length of the pulses, resolving the problem.

Here is a summary of the wiring (all named connection points are on the driver board):

  • Connect the output of the wall wart to the driver board: wire the positive and negative leads, respectively, to the "+" and "-" connection points.

  • Connect the positive lead to one side of the indicator lamp. Connect the other side of the lamp to the open collector output (point 3).

  • Connect the sensor bulb to the driver board. The anode (wire) connects to point A; the cathode (plate) connects to point K.

  • Connect the pulse-extending capacitor to the driver board at point Cx. Be sure to wire the positive side of the capacitor to the "+" connection.

  • Leave the "background cancel level" jumper at position 3.
Here's a rough diagram of the circuit; see the technical notes for more details:



The unit is mounted under a cabinet, to the right of the vent hood:


This is a rear view of the unit, including the indicator light:

Sensor back.JPG

Note that the indicator light is quite visible in daylight:

Sensor lit.JPG

An indicator light for gas stovetops - posted at Fri, 20 Mar, 02:56 Pacific | «e» | TrackBack


I'm glad to see someone else sees the necessity of indicator lights for gas stove tops. I cannot figure out for the life of me why it is a requirement for electric stove tops, but not gas.

However, I'd have to respectfully disagree with your assumption that "most cooks will quickly notice the natural gas smell from an unlit burner". Not only am I convinced that cooks often miss this unsafe condition, but I'm most concerned about my aging parents. My interest in this is based on the fact that I now have an aging parent living with me; the "burner is turned on, but unlit" issue has quickly become one of my primary concerns. In fact, I'm considering replacing my gas stove top (which I love) with an electric version because I'm fearful of the consequences of the "burner is turned on, but unlit" condition.

I wish there was a simple way to incorporate a basic electric switch solution into the knob of each burner that would illuminate an indicator light if any of the knobs were not completely turned to the off position. I realize this would most likely require a slightly different installation for each manufacturer, but I think its the only way to address both the "burner is turned on, but unlit" issue as well as the "burner is left on" issue.

Even with an electric stovetop, the indicator light may not be very visible. A while back, I installed an auxiliary light on our stovetop. See A Signal Accomplishment for details.

Post a comment

Note: All comments are subject to approval. Spam will be deleted before anyone ever sees it. Excessive use of URLs may cause comments to be auto-junked. You have been warned.

Any posted comments will be viewable by all visitors. Please try to stay relevant ;-) If you simply want to say something to me, please send me email.