What caused the change in a pilot light at my workplace?

[fleetingmoment]

This has been bugging me for a few months now, and I'm hoping that someone can provide me with an answer.

A while ago, my colleague was soldering with an oxy-acetylene torch and paused for a few moments in between jobs, leaving only the pilot light on top of the unit burning. Sitting about 4-5 metres away, I happened to use my air pistol to clean a mechanical component, when he noticed that the pilot light had suddenly increased in length. He shouted and called my attention to it, which led to us speculating as to what might be causing it. The flame wasn't affected laterally - increasing only in length, so we ruled out air disturbance or pressure from sound waves. I also tried banging the bottles, humming at various pitches, shouting and clapping - all to no avail,

Eventually I walked across the hall and tried with another air gun about 20 metres away and the result was the same as the original. Noticing that the flame on the pilot light turned from blue to a sooty yellow colour during the moments when the air gun was activated (indicating incomplete combustion), I reasoned that the sound of the escaping air was at a frequency which resonated with the acetylene, causing a disturbance in the mixture of the two gasses.

I could be completely (even laughably) wrong, but it'd be great if I could get an answer as to what was causing it and finally set my mind at rest.

 

[sophiecentaur]

This sort of phenomenon is always good fun. You seem to have had some fun, already, trying to isolate the problem.
You are suggesting that the flame was being modulated by a sharp pulse of sound? I could believe that in principle but the distance involved makes it a bit difficult to accept. (i.e. more sensitive than I'd intuitively expect) If you could repeat the experiment with a range of distances for the airgun. Or different guns?
An Oxy Acetylene torch has a very small mixing tube, if I remember right so I don't think the effect could be due to the burning with the air. I could suggest that the pressure regulator is a bit sensitive at low flow rates and that it was being a bit microphonic. If you have the time and opportunity, you could try a loudspeaker with a variable tone source (smart phone App) and see how a continuous tone affects the flame. Did you actually tap the regulators on the bottles? I don't know the design of such regulators but Scuba regulators are designed to balance out disturbances so your air gun pulse may just be getting past 'this common mode rejection' mechanism.

 

[fleetingmoment]

This sort of phenomenon is always good fun. You seem to have had some fun, already, trying to isolate the problem.
You are suggesting that the flame was being modulated by a sharp pulse of sound? I could believe that in principle but the distance involved makes it a bit difficult to accept. (i.e. more sensitive than I'd intuitively expect) If you could repeat the experiment with a range of distances for the airgun. Or different guns?
An Oxy Acetylene torch has a very small mixing tube, if I remember right so I don't think the effect could be due to the burning with the air. I could suggest that the pressure regulator is a bit sensitive at low flow rates and that it was being a bit microphonic. If you have the time and opportunity, you could try a loudspeaker with a variable tone source (smart phone App) and see how a continuous tone affects the flame. Did you actually tap the regulators on the bottles? I don't know the design of such regulators but Scuba regulators are designed to balance out disturbances so your air gun pulse may just be getting past 'this common mode rejection' mechanism.

Thanks for your suggestions, sophiecentaur.
I'll give them a try when I get the chance, though it may take a while, due to me having been moved to a different working area. I can't manage different air guns (they're all of one type), though perhaps I could try with different feed pressures as well as different distances. The main feed from the compressor is (officiallyI) 8 bar, though I've often found it to go as high as ten. I suppose I could fit an old pressure regulator onto a feed line and experiment with various pressures at different distances to see if there's any change.

I didn't tap the regulator on the bottles, though these are quite stiff, being levers that are pulled up (perhaps to open a ball valve inside) in order to open the flow, so I can't imagine them being affected. The regulator for the pilot light could, however, be another matter. It's a screw-type valve and certainly easier to operate, requiring only the slightest turn to change the size of the pilot light, though that said, I've never know it to be affected by anything other than a deliberate turning of the knob - despite long periods, during which all sorts of loud work takes place nearby.

 

[sophiecentaur]

I didn't tap the regulator on the bottles, though these are quite stiff, being levers that are pulled up (perhaps to open a ball valve inside) in order to open the flow, so I can't imagine them being affected.

The sort of regulator I am used to, for Qxy Acetylene welding has a pressure gauge on it and a diaphragm to maintain the pressure to the desired value (one for Oxy and one for Acetylene). What sort of regulator do you use? Also, you refer to 'soldering'. Does that need oxygen? I am familiar to Acetylene torches, which use air (like a bunsen burner or sometimes blown). The gas pressure is less critical for those.
What are you actually describing?

 

[fleetingmoment]

The sort of regulator I am used to, for Qxy Acetylene welding has a pressure gauge on it and a diaphragm to maintain the pressure to the desired value (one for Oxy and one for Acetylene). What sort of regulator do you use? Also, you refer to 'soldering'. Does that need oxygen? I am familiar to Acetylene torches, which use air (like a bunsen burner or sometimes blown). The gas pressure is less critical for those.
What are you actually describing?

Sorry, I meant tap the cylinder valve, not regulator. I can't find any pictures of these. The best way I can describe these cylinder valves is that they have curved levers for handles, which you pull up towards you to open the flow and push down away from you to shut off the flow. The actual regulators themselves are typical regulators with two pressure gauges. I'm not that familiar with the set-up (or oxy-acetylene in general) and only use it occasionally, but from memory there are two (for want of a better term) lever cylinder valves, which open the flow from the bottles, two needle valves on the torch and a needle valve for the pilot light.

With regard to soldering, I actually meant braising (in this case braising with silver). My English isn't what it used to be after many years living abroad and I translated directly from the Danish lodde, which means solder. So yes, it's braising with oxy-acetylene, though with a gentle flame. My apologies for the confusion.

 

[sophiecentaur]

The actual regulators themselves are typical regulators with two pressure gauges.

They would be the most sensitive things in the system so I would go for them.
I get the brazing thing. Much more understandable.

 

[Asymptotic]

I suppose I could fit an old pressure regulator onto a feed line and experiment with various pressures at different distances to see if there's any change.

I like this idea. Another isolation step may be to cut a hand hole in a cardboard box, and actuate the air gun inside of it to eliminate the (remote) possibility that disturbances in ambient air near the pilot flame are involved.

I've used oxy-acetylene cutting and welding torches, and air-acetylene rigs for soldering, but none were equipped with integral pilots, and I'm having a hard time visualizing what such a torch looks like. Could you take a camera with you when you run the above experiment, and post overview and detail photos?

typical regulators with two pressure gauges

Two gauges (inlet and outlet pressure) indicate a two-stage regulator, which provides more precise control and better quality pressure regulation than is possible with a single stage regulator.

The regulator for the pilot light could, however, be another matter. It's a screw-type valve and certainly easier to operate, requiring only the slightest turn to change the size of the pilot light ...

...there are two (for want of a better term) lever cylinder valves, which open the flow from the bottles, two needle valves on the torch and a needle valve for the pilot light.

Is this pilot needle valve on the torch body? Is this what was referred to earlier as "the regulator for the pilot light"?
What colors are the gas cylinders?
Is a compressed air line connected to this rig?

I've never seen one, but Linde, National, and probably other manufacturers offer brazing torches designed to use acetylene and compressed air. These develop hotter flames than air-acetylene torches, but are far cooler than oxy-acetylene. I'm prone to discount this possibility (although it fits with pilot flame disturbance) because you mention having two cylinders. However, unless the torch can be fitted with a cutting head (cutting requires oxygen; compressed air won't do to cut steel) there remains a remote possibility the second tank contains compressed air.

Color coding is far from standardized, in fact, it's a mess. In the US, acetylene cylinders have traditionally been deep red and oxygen bottles green, but lately suppliers have been moving toward the ISO standard of marking the radiused part of the bottle near the top with red to indicate a flammable gas, or bright green to indicate it is inert (although I've yet to see a white oxygen cylinder with a green neck, which is the closest thing to an international standard). A black cylinder body usually denotes nitrogen, but black (or black and white) has also been used to indicate compressed air.

The limb I'm standing on is very much more a twig, but is it possible this rig uses a compressed air/acetylene torch, and was plumbed to use plant compressed air when available, or bottled air when it is not?

 

[sophiecentaur]

Two gauges (inlet and outlet pressure) indicate a two-stage regulator, which provides more precise control and better quality pressure regulation than is possible with a single stage regulator.

I thought the first gauge is to indicate the state of charge of the cylinder. Afair, you just opened the inlet tap to maximum and used the knob on the regulator to produce the pressure appropriate to the torch.
For brazing, I remember just using an acetylene / air torch (like a bunsen). Its heat output Power was much the same - just a much lower flame temperature.
I still think the regulator is the only component that would be as sensitive as necessary, if the 'signal' is getting through the room air. If the torch and gun were both getting an air supply from a common source then that would explain everything - an extra load on the air supply would affect the flame colour.
Could the OP please give us a precise description of what was actually connected to what? We are all on tenterhooks about it.

 

[Asymptotic]

I thought the first gauge is to indicate the state of charge of the cylinder. ... We are all on tenterhooks about it.

Yes, on both counts. :)

I think of a two-stage regulator as a pair of single stage regulators connected in series to form a coarse/fine adjustment. The first regulator and it's gauge sees full cylinder pressure (for instance, 2200 PSI/70°F is at standard full charge for a US oxygen cylinder) and regulates this down to a preset intermediate value (say, 200 PSI).

The second regulator stage reduces this intermediate pressure to whatever the user adjusts it to, and is what the low pressure gauge shows. Adjusting a two-stage is less finicky - regulator screw adjustment changes outlet pressure by less (0-200 PSI full scale versus 0-2500 PSI or so). The other advantage, at least at high flow rate O² (for instance, as may be required to cut 1 inch and thicker plate) are fewer problems with the regulator freezing up.

I still think the regulator is the only component that would be as sensitive as necessary, if the 'signal' is getting through the room air.

Seems that way to me, too, once the possibility of a compressed air torch has been eliminated.

 

[sophiecentaur]

I think of a two-stage regulator as a pair of single stage regulators connected in series

Same as a scuba set. The intermediate pressures not adjustable and has a very chunky diaphragm and spring. This gives the second stage a chance to be 'sensitive'.
I wonder about your mention of "cutting". Isn't it usual to use a different kind of torch? (one with a trigger)
Perhaps the OP could send us a picture of what we are talking about. I am prepared to be totally surprised because this problem just doesn't make sense to me any more.

 

[Asymptotic]

Same as a scuba set. The intermediate pressures not adjustable and has a very chunky diaphragm and spring. This gives the second stage a chance to be 'sensitive'

Exactly so!

Over the years I've used Victor, Miller, and Linde gear. No preference, really - it boiled down to whatever manufacturer the local gas & welding supply dealer carried at the time, but the downside is that while hose connections are standard, none of their torch parts are interchangeable. Pictured below is a small Victor torch kit.

The main body has two needle valves at the gas connections prior to a mixing chamber. Welding tips (at the bottom of the image) are available in a range of bore sizes to accommodate different task requirements, and one would be mounted onto the torch body for welding, or brazing service.

For cutting, the bullet-shaped tip at the lower right (again, available in a range of sizes) is fitted into the cutting head shown at the top. Preheating is done by feeding an oxy-acetylene mix determined by regulator settings in conjunction with torch body needle valve setup (most often adjusted to provide a neutral flame) through the lower tube, along the tip periphery and through a rosette of tiny holes around the tip face. Once the piece is hot enough, the cutting lever is depressed to send a gout of O² (adjusted by the needle valve adjacent to the lever) through the upper tube, and into a large hole bored into the tip's center.

Color me perplexed, too.

 

[fleetingmoment]

I like this idea. Another isolation step may be to cut a hand hole in a cardboard box, and actuate the air gun inside of it to eliminate the (remote) possibility that disturbances in ambient air near the pilot flame are involved.

I've used oxy-acetylene cutting and welding torches, and air-acetylene rigs for soldering, but none were equipped with integral pilots, and I'm having a hard time visualizing what such a torch looks like. Could you take a camera with you when you run the above experiment, and post overview and detail photos?Two gauges (inlet and outlet pressure) indicate a two-stage regulator, which provides more precise control and better quality pressure regulation than is possible with a single stage regulator.

Is this pilot needle valve on the torch body? Is this what was referred to earlier as "the regulator for the pilot light"?
What colors are the gas cylinders?
Is a compressed air line connected to this rig?

I've never seen one, but Linde, National, and probably other manufacturers offer brazing torches designed to use acetylene and compressed air. These develop hotter flames than air-acetylene torches, but are far cooler than oxy-acetylene. I'm prone to discount this possibility (although it fits with pilot flame disturbance) because you mention having two cylinders. However, unless the torch can be fitted with a cutting head (cutting requires oxygen; compressed air won't do to cut steel) there remains a remote possibility the second tank contains compressed air.

Color coding is far from standardized, in fact, it's a mess. In the US, acetylene cylinders have traditionally been deep red and oxygen bottles green, but lately suppliers have been moving toward the ISO standard of marking the radiused part of the bottle near the top with red to indicate a flammable gas, or bright green to indicate it is inert (although I've yet to see a white oxygen cylinder with a green neck, which is the closest thing to an international standard). A black cylinder body usually denotes nitrogen, but black (or black and white) has also been used to indicate compressed air.

The limb I'm standing on is very much more a twig, but is it possible this rig uses a compressed air/acetylene torch, and was plumbed to use plant compressed air when available, or bottled air when it is not?

Hi Asymptotic,
Yes, I'll try and take a camera with me at some point and incorporate your suggestion with the cardboard box. I just have to make sure I do it after my colleagues and supervisor have gone home, as it's forbidden to take photos where I work. Not that I haven't done it before in connection with work-related matters, but I'd have a hard time explaining this one to my boss, were I discovered. In the meantime, the pilot is situated on top of a separate feed that runs to a tubular metal structure on top of the rig which also contains (for want of a better description) a hook-type lever, which cuts off the flow to the torch when it's not being used - the torch being hung by its tip on the hook. The needle valve for the pilot light sits on the same structure - i.e. not on the torch.

I can't remember the colours of the bottles off hand. I'm not a regular user and it's been a few months since I last used them. Also, I could be wrong on the regulators, so don't quote me on those - especially as there are two different oxy-acetylene rigs. No, there is no compressed air line to the rig. In addition, given that we have yearly ISO inspections, I imagine they fall under ISO standards.

 

[fleetingmoment]

I thought the first gauge is to indicate the state of charge of the cylinder. Afair, you just opened the inlet tap to maximum and used the knob on the regulator to produce the pressure appropriate to the torch.
For brazing, I remember just using an acetylene / air torch (like a bunsen). Its heat output Power was much the same - just a much lower flame temperature.
I still think the regulator is the only component that would be as sensitive as necessary, if the 'signal' is getting through the room air. If the torch and gun were both getting an air supply from a common source then that would explain everything - an extra load on the air supply would affect the flame colour.
Could the OP please give us a precise description of what was actually connected to what? We are all on tenterhooks about it.

Hi sophiecentaur,
As I mentioned to Asymptotic, I'll try and get some photos done soon to clear up any confusion.

 

[Asymptotic]

... the pilot is situated on top of a separate feed that runs to a tubular metal structure on top of the rig which also contains (for want of a better description) a hook-type lever, which cuts off the flow to the torch when it's not being used - the torch being hung by its tip on the hook. The needle valve for the pilot light sits on the same structure - i.e. not on the torch.

Does it looks something like this?

 

[fleetingmoment]

Does it looks something like this?
View attachment 204864

Yes, that's pretty close to it.

 

[Asymptotic]

Yes, that's pretty close to it.

The gizmo is called an 'economizer' - this one is from a German company (Zinser). The hook toggles against a pair of normally open valves, and closes them when the torch is put into place. Fuel gas for the pilot is tapped off prior to the shut-off valve, and mixed with ambient air.

Don't know how much closer this gets us, but at least variation in the the oxygen circuit is out of the picture.

 

[Pratyeka]

This is very intriguing. Does the change in the pilot flame start and stop instantly with the activation of the air gun? How big is the room where this takes place? Is the room closed and air tight or open with ventilation?

 

[fleetingmoment]

This is very intriguing. Does the change in the pilot flame start and stop instantly with the activation of the air gun? How big is the room where this takes place? Is the room closed and air tight or open with ventilation?

Hi, Pratyeka,
The change in the pilot light starts instantly with activation of the air gun and also stops instantly with deactivation of the air gun. I've no idea of the size of the room. It's a fairly large hall. I'll try and get back with a rough idea of how much area it covers tomorrow. It's fairly well ventilated with air being pumped in through a ventilation unit suspended from the ceiling which runs most of the length of the hall, individual fume extraction points at work benches, as well as port doors at one end, which are usually open and lead to warehouse areas - so it's not small or airtight at all.

 

[fleetingmoment]

The gizmo is called an 'economizer' - this one is from a German company (Zinser). The hook toggles against a pair of normally open valves, and closes them when the torch is put into place. Fuel gas for the pilot is tapped off prior to the shut-off valve, and mixed with ambient air.

Don't know how much closer this gets us, but at least variation in the the oxygen circuit is out of the picture.

That's interesting. I'll try my best to get something done this week, but as I said, no longer being in that department means I'll have to wait until everyone else has gone home.

 

[CWatters]

Is it definitely getting longer or is it that the flame is just becoming more visible so it looks like it's longer?

 

[Pratyeka]

I think the color change is more significant than the length of the flame, because the length would be dependent on the pressure, while the color would be related to the chemical reaction. Turning from blue to yellow seem to indicate a reduction in the amount of oxygen available to the flame. So with all these facts presented above (large room, distance not affecting the effect, instantaneous reaction between air gun activation and flame behavior, no physical link between source of gas and air gun compressed air line...) what can possibly be happening here? There is more than one witness, so the cause is not related to the observer.
There has to be some environmental change around the flame. It's kind of difficult to test ideas without being able to reproduce the phenomenon.

 

[Asymptotic]

When compressed air passes through an orifice significant levels of ultrasound are produced in addition to an audible hissing noise, and this is what forms the basis of ultrasonic detectors like the UE Systems stuff I've used in tracking down air leaks. After this thought struck me I started nosing around, and learned of an earlier form of ultrasonic detection called the "sensitive flame" method.

A narrow sensitive flame is moved along the medium. At the positions of antinodes, the flame is steady. At the positions of nodes, the flame flickers because there is a change in pressure. In this way, positions of nodes and antinodes can be found out in the medium. The average distance between the two adjacent nodes is equal to half the wavelength. If the value of the frequency of ultrasonic wave is known, the velocity of ultrasonic wave propagated through the medium can be calculated.

I'm not claiming ultrasonics are involved in our mystery, but nevertheless the possibility exists.

 

[fleetingmoment]

I moved the rig to another location in order to get a bit more peace. Even so, I decided that I won't be conducting any more tests, because people where I work have a habit of asking lots of questions and are generally suspicious, which means that I was only able to do a couple of tests with 1 fixed distance between the air gun and pilot flame and 1 fixed air supply pressure (assumed to be about 8 bar) for the air gun.

Anyway, the area of the new location is about 250 m² and the volume approximately 950 m³. The rig looks as follows:

I set the rig at distance of 2.9 metres from the air gun. I used 2 different flame sizes, though in my haste only managed to record one of them: 20.77 mm. I noticed that the initial flame wasn't as blue as I remembered and the affected flame not as sooty, which is most likely due to my less that trustworthy memory.

For the first test, the air gun was open to the surrounding air. I pointed the air gun in the direction opposite to the flame and downwards. Strangely, there seemed to be quite a long delay before the flame was affected:


For the second test I increased the length of the pilot flame (length unrecorded). This flame seemed to increase in length more than the first flame of 20.77 mm, which seemed to increase in intensity:


For the third test I activated the air gun in a closed cardboard box with a volume of approximately 0.026 m³ : This seemed to make the flame contract.


For the fourth test, I filled the box with cloths in order to muffle the sound and activated the air gun. This seemed to have no effect on the flame.


Sorry about the poor quality recordings.

 

[Asymptotic]

This is a different color-coding scheme than any I'm familiar with, but the gray tank with white neck suggests oxygen.

I'm getting "This video is unavailable." messages on all four of the embedded clips. Tried copying a URL, and playing in Youtube directly, but had the same results.
However, from your commentary it seemed that altering the sound produced by the air gun affected flame shape.

 

[fleetingmoment]

Sorry, I had the videos set to private on YouTube. I've just tried them in another browser and they should load now.

 

[JBA]

Somewhere far back in my memory it seems I previously seen a case where a candle flame was similarly affected by sound from a string instrument (maybe a violin) but that is all I can recall at this point. A bit of searching turned up nothing on that subject; but, some combustion instability references led me to consider the possibility of an a harmonic in the gas flow through the flame tube of the burner that could be affected by an external sound source. Just a thought.

 

[Pratyeka]

excellent videos, it appears to be caused by the sound of the air gun, the last test with the stuffed box shows that with all other factors being the same the reduction of the sound led to no change in the flame.
There must be some research data available on the effect of sound (ultrasound) wave on gas pilot light.
Very good of you to perform and document these test despite the work environment. Thank you.

 

[Pratyeka]

found it: https://en.wikipedia.org/wiki/Sensitive_flame

 

[fleetingmoment]

Many thanks, Asymptotic and Pratyeka for your answers and information regarding the sensitive flame. It's always nice to be able to able to clear up a mystery. I'll be sure to show the answer to my colleague who noticed the phenomenon.

 

[Asymptotic]

Excellent videos. The first one is particularly interesting in that at first the flame seems to be pushed sideways (~1.54 seconds after air flow becomes audible), then flares into a more orangish, more reducing flame at ~ 1.87 second. In the first three tests, flame appearance changes shortly before sound is no longer heard. I'm guessing the sound persists due to echo for a short time after the air gun is turned off.

Two observations about economizer mounting - the distance between the last support point and the economizer is very nearly the height of the cart itself, the aluminum right angle bracket it is bolted to appears rather "springy", and has a fairly large angle away from horizontal. For the first and second point, I'd bet it becomes resonant at a number of audible and ultrasonic frequencies. For the third point, the flame front won't sit evenly around the annulus of the burner tube, and I suppose this would increase the sensitivity of the flame to disruptions.