Get Tech Tips
Subscribe to free tech tips.
Flame Sensing – The Basics
Proving flame is an important part of the gas firing sequence. Without proof of flame, you risk dumping unspent gas into the heat exchanger, resulting in an explosion.
There are many ways to “prove flame,” but we are focusing on the flame sensing rod (flame rectification) method here.
Here are the facts:
Flame sensing rods, also known as flame rectifier rods or flame rectification rods, are commonplace in modern hot surface and ISI (intermittent spark ignition) gas-fired appliances.
Flame sensing rods stick out into the flame and connect back to the furnace board. Once the board sends a call to the gas valve to open, it monitors the current flow on the flame sensing rod. It generates a potential (voltage) at the flame sensing terminal; this terminal is connected to the sensor with a conductor. When no flame is present, there will be potential at the rod and no current. When a flame is present, a small microamp (μA) direct current (DC) will be present as a path is made between the rod and the ions in the flame. This small DC current signals the board that flame exists and all is well with the world. If it does not sense this microamp (μA) DC current within a few seconds, it will shut off the gas valve and try again.
The board outputs this potential (voltage) on the flame sensing terminal right at the beginning of the sequence to confirm that the path is “open” with no flame. This step ensures against false positives (sensing flame/current when there should be none), and once it goes from 0 current to the rated microamp (μA) current, the board “knows” that flame is present.
These flame sensing rods are “dumb” devices. They do not generate potential (volts) or current (amps); their predecessor, the thermocouple (seen in standing pilot systems), does generate a potential itself, which is often the source of the confusion.
A flame sensing rod is a piece of metal with a ceramic insulator that keeps it from grounding out. That is all. However, because it is conducting in the millionths of an amp (microamp), a lot can go wrong with it that a normal electrical component wouldn't have any issue with. Tolerances are tight, so small factors make a big difference.
Flame sensors fail when:
- They short out due to a cracked insulator.
- They fail open because they are broken.
- They don't conduct because they are not properly placed in the flame.
- They become coated in silica (glass) or carbon.
Before I go any further, I want to address a common question. Do flame sensors have a special coating that can be rubbed off with improper cleaning?
Well, if we are talking about a thermocouple or a thermopile, then yes, absolutely, but we aren't discussing standing pilot systems here.
I have seen many flame-sensing rods, and I have done a good deal of research; I have found no evidence that typical flame-sensing rods have a special coating on them that can be rubbed off. Now, if you have real, quantifiable proof from a manufacturer that says otherwise, PLEASE provide it to me so I can retract this statement.
I think it's more likely that issues techs see from cleaning are due to cleaning with sand cloth (Emory cloth) and leaving behind grit that can lead to a coating and poor conductance once heated. I chalk the other part up to confusing a thermocouple with a flame rod and bit to superstition.
Here are the steps to test a flame sensor:
- Ensure the furnace is properly grounded. You can do this by powering down the heater and taking an ohm reading between neutral and the burner assembly. You should read a few ohms of resistance max; the lower the ohm reading, the better grounded it is.
- Make sure your polarity is correct: incoming hot connected to hot, neutral to neutral.
- Ensure that the rod is positioned so that it will be covered in flame.
- Get a meter that reads in the microamp scale with a 0.10 resolution minimum. Use a good QUALITY meter for this, and make sure your leads are in the correct locations.
- Connect your leads in SERIES. Disconnect the lead from the rod, connect one lead to the rod, and connect the other to the terminal to the board WITH THE CONNECTOR UNHOOKED FROM THE ROD.
- When the flame lights, you should read between 0.5 and 10 microamps (μA), depending on the furnace. Readings between 2 and 6 are common.
If you do not have a proper microamp (μA) reading, you can confirm the following:
- The flame rod is not open—ohm from the tip to the terminal on the rod. If the rod is open, it is failed.
- Check the insulator and make sure it isn't cracked or grounded
- Check for proper burner grounding and incoming power polarity (as mentioned)
- Clean the rod. Now, this is a controversial one. I suggest using very fine steel wool or an abrasive pad (magic erasers often work). Remove and clean the rod and ensure you wipe it clean of any particles left over from cleaning. Handle it very gently. Once complete, perform an ohm test from tip to terminal again to ensure that you haven't damaged anything during cleaning. If you want to be really crazy, use some electrical contact cleaner on it after cleaning to help remove any residue—just nowhere near the flame, unless you don't want eyebrows.
Once you have established all of the above and are still not getting the required microamps, your only remaining option is to replace the board.
Word of warning:
Test your tools regularly. If you trust your meter and aren't 100% sure your meter is working and set up properly, you may end up with a misdiagnosis. Test and calibrate your tools regularly.
Do every possible test before replacing a board. Many techs advocate just replacing a flame sensor if they suspect it isn't conducting well. I am cool with that so long as:
- You wouldn't charge the customer for it if there were nothing wrong with it.
- Your company is OK eating the cost of unneeded rods.
Or, you just install a new one long enough to test. That is all fine and good if you have extra flame rods in your truck. Many techs do not have that luxury.
If flame rods are getting dirty/coated often, you will want to find out why. There is something in the environment or the combustion that is causing it. Many techs notice that furnaces bringing their combustion air from laundry rooms or basements with cleaners nearby often get dirtier quicker.
In summary, flame rods should be:
- In the flame
- Not open
- Not shorted
Now is the part where the furnace techs from all over the world tear me apart.
You’re awesome!! I don’t know where you find the time to put these articles together, but they are dead on. I hope everyone reads them because we need more highly skilled technicians out there and they are hard to find. The articles are like free gold. I hope more people read them and retain what you’re teaching!! Kudos to you!
Thank you Scott. You know I respect you and you your business, that means a lot coming from you.
Here are the steps to test a flame sensor –
Ensure the furnace is properly grounded. You can do this by powering down the heater and taking an ohm reading between neutral and the burner assembly. You should read a few ohms of resistance max, the lower the ohm reading the better grounded it is.
Great article…when you test from neutral to the burner assembly, IF you have neutral hooked up to unit and test to the burners… wouldn’t you be reading back on the ground that is hooked up on the unit and would give a false low OHMS?
Hello Greg. The intent is to confirm that the burners are electrically equivalent to Neutral (good path). By checking from the burners to neutral you are simultaneously confirming that the burners are grounded, the unit is grounded AND connected to neutral. You kill a few birds with one stone… not that I’m into killing birds ?
Just did a Johnstone furnace basics class this week, thought I might pick up on something that would help when my cooks cry that something “ain’t right” with a fryer or flat top grill.
Was amazed that the dozen or so “techs” from a large local HVACR company did not know if there meters read uA or how to read uA, or why?
So what sorts of things make the flame sensor dirty (besides cracked heat exchangers) ?
Anything that causes a dirty flame. Improper gas pressure, poor combustion air, contaminants in the combustion air etc..
Lint from dryers, litter boxes, those plug in glade air fresheners to name a few…
Mouse urine, very common, also a water leak.
Thanks for the info, really look forward to any information I can get my hands on. Need to get a new meter with microamp capabilities.
Nicely Done. I enjoy these kind of helpful articles.
Thanks David. I’m glad you are here.
Well written and well said. Too few techs understand this! And don’t forget to also clean the face of the burner that is being used as the ground for the flame rectification!
Great article Bryan! Flame rectification has seemed to always be a sort of questionable topic. But yours was spot on. Thank you.
I appreciate that Kyle
Great article Bryan – a suggested clarification (in caps):
“It does this by generating an AC potential (ALTERNATING CURRENT voltage) at the flame sensing terminal, this terminal is connected to the sensor with a conductor. When no flame is present there will be potential at the rod and no current, when a flame is present a small microamp DC current will be present as a path is made between the rod and the ions in the flame . IONS IN THE FLAME ONLY ALLOW AC CURRENT GENERATED BY THE BOARD TO FLOW IN ONE DIRECTION, SO WHILE AC VOLTAGE IS GENERATED AND APPLIED FROM THE FLAME SENSOR TO GROUND, THE BOARD ACTUALLY MEASURES DIRECT CURRENT IN MICROAMPS. IN THE WORLD OF ELECTRONICS, THEY CALL THIS PROCESS OF “CUTTING” AN ALTERNATING CURRENT SIGNAL TO A DIRECT CURRENT SIGNAL, “RECTIFICATION.”
Why does it get rectified?
In the flame, there are heavy positive ions and light negative electrons. They typically apply 110VAC 60Hz to the flame sensor rod. It is changing from pos to neg very fast. The electrons are far more mobile than the heavy ions. When the flame sensor rod is positive, the negative electrons in the flame are absorbed by it, creating a vary small current in in the wire. The electrons are quickly swept up from the flame. When the flame sensor rod is negative, it will attract the positive ions and repel electrons, which are far more massive, so they don’t have time to reach the rod before it becomes positive again. There is no current when the flame sensor is negative. The current in the flame sensor rod is indeed half-wave rectified. On caution I didn’t see in your great explanation is that quite often they apply 110VAC 60Hz to the flame senor rod. If you happen to touch it, you will likely be shocked, and I’m not referring to the emotional kind of shock.
Incidentally, I would often tease working HVAC guys with a question of how a flame sensor works, and they all invariably tell me it is a thermocouple. I then show them one I carry with me in my pocket, and it has only one wire coming from it, and they say, “yeah, so? ..the thermocouple is in the metal probe.” (In my younger days, I would RF diode sputter deposit, and remarkably, the flame sensor rectifies in a similar way to sputtering.)
Thanks for the rich article
You’re welcome 🙂
I noticed you mentioned how to clean the flame sensor rod, is it also important to make sure the burner surfaces and chamber are also cleared of any oxidation, soot, or other causes of high impedance that might impair the current from the rectified voltage?
My flame sensor signal comes from the intermittent spark module, not the main control board (Janitrol GUN1004B). So if I verify no signal, how do I know which to replace, the board, or the module? Also, why is this wire so doggone fat? Is it because it is shielded to protect against emf interference because of the puny current?
Regarding “coatings” — a flame-detection rod (pin, needle, wire, electrode, etc.) may have a surface treatment (plating or etc.) intended to deter any buildup of deposits (corrosion, carbon, silica, whatever) which might reduce conductivity, i.e., lower microamps. In that case, abrasive cleaning might improve performance in the short term (you’re a hero!) but the benefit might deteriorate in the long term (why does the failure return?)
Thank you. I now see a much better article on the common [rectifying] flame sensors used in residences in the USA. Because the misconception that the metal rod detects heat, your final comment is probably well founded (about the techs). There are sensors that detect heat (thermocouples or RTDs) and optical sensors, but I believe the common flame sensor (simple metal rod) is far more robust (because of simplicity) and less likely to fail or give false positives (reporting a flame when there isn’t one … letting gas fill the room). Another factor is cost, a board can cost the better part of $100, but a new flame sensor seems to cost similar to one or two trips to Starbucks … or Tim Hortons. Simply swapping them out and/or using them to diagnose other problems (like carbon buildup) could be routine. Another great thing about the flame sensor as an interlock is that it is very hard to defeat… a thermocouple may be tricked into turning on with a lit match or an optical sensor may be tricked with a lamp. Simply shorting the flame sensor will shut off the gas, and they can’t be tricked with a match or a lamp.
Very nice fix to article. Thank you.
Have you ever seen a flame rod that gives 5 Volts at the honey well but does not have a flame in the burner?
I was giving a igniter that was for a dryer, and seems to work fine, have you heard of this? This igniter’s seem to be rated the same , so is there any difference other than being round and the dryer one a different shape?
How do we check flame current on a flame sensor that is also a spark electrode?
Your are spot on with your flame sense tech tip. I am an engineer who has been designing and producing gas fired restaurant equipment for over 25 years. Getting the flame sense right always takes a lot of time and testing. We run into the basic issues you described on every product.
Flame rods have a coating to prevent corrosion.
Do flame sensors have a life expectancy?
Great post, thanks. I’ve got a flame sensor that is sporadically giving false positives which causes the blower to cycle even when the furnace is firmly off. New controller board. Works great when there actually is a flame. Are these things subject to RF interference? Any thoughts what else could cause false positives? Could lead length be an issue?
The only thing I do differently is where you mentioned connecting the meter to the rod and board. I connect it to the rod, and the wire I just removed from the rod….. which goes to the board.
This Article is Correct. 26 years in the hvac trade, Lead Service Technician & i Still love what I do. On a Simple Fall Heat Maintenance 3 DAYS AGO Rheem 95% Furnace Displaying Error code (11) IGNITION FAILURE. Before performing heat maintenance Furnace locked out after 4 IGNITION Sequence FAILURES. ERROR CODE 12 STATES (LOW FLAME SENSE.) I knew it wasn’t a Dirty flame sensor. Powdered furnace off. With the flame sensor connected using meter from flame sensor to ground, 52 Ohms. Traced out F.S. wire to harness plug on board and disconnected harness. Tested Pin that energizes F.S. at Circuit board and sure enough 52 Ohms. Bad Circuit board. Client was upset saying it worked be4. Clearly blaming me for Heat FAILURE Until I Showed her why Heat Failed. Furnace located in garage where there is moisture. De humidifier isn’t big enough. Back of board was oxidized. In the past she has replaced two inducer Motors third circuit board nobody knew why it was MOISTURE, HUMIDITY in her Garage ALL along.. this explains why she went through so many components. Replace circuit board OEM backup and running. Loss of flame & IGNITION FAILURE are 2 different Error codes. Hope everyone reads this and leans if they never came across to this Situation.