Guide
Backfire sustained backfire and flashback: what each one is, what causes it, and what to close first
A backfire is a pop as the flame enters the nozzle; a sustained backfire keeps burning inside the blowpipe with a hiss or squeal; a flashback runs back into the hose or regulator. Close both blowpipe valves at once, in the maker's order: the C.P.I. 1968 code put acetylene first, some makers oxygen.
- Normal outcome of a backfire
- A flame that backfires into the neck or body of a blowpipe normally becomes extinguished there
- Shut-down order
- Close both blowpipe valves at once
- Root causes recorded
- Overheated nozzle; obstructed nozzle letting higher-pressure oxygen climb the acetylene duct; a leak at the handle-to-injector joint
- Quenching diameter
- A stoichiometric oxyacetylene flame passes any opening larger than about 0.12 to 0.16 mm at 1 atm
- Arrestor test
- Each arrestor must stop five flashbacks of a static 32 to 35 % acetylene-in-oxygen mixture at maximum operating pressure
Three different events get called "flashback" on a shop floor, and treating them as one is how the dangerous one is missed. A backfire is a pop at the nozzle. A sustained backfire is a flame that has moved inside the blowpipe and stays there. A flashback is a flame that has left the blowpipe altogether and is running back up the hose. They share causes, they escalate into one another, and they are stopped by the same discipline: shut the fuel off, find the cause, and never relight on a hunch.
The three events side by side
| Event | What you hear and see | Where the flame is | Immediate risk | What stops it |
|---|---|---|---|---|
| Backfire | A sharp pop; flame goes out or relights at the tip | Momentarily inside the nozzle orifice | Spatter, a cold weld, a warning | Usually self-extinguishing |
| Sustained backfire | A continuous hiss or squeal; flame gone from the tip; neck or handle heating fast | Burning in the mixer or neck | Blowpipe destroyed in seconds; ignition can travel on | Shutting the blowpipe valves at once |
| Flashback | A bang in the hose or regulator; soot at fittings; hose may burst or burn | In the hose, regulator, possibly the supply | Hose fire, regulator rupture, flame into cylinder or pipeline | A flashback arrestor; closing the cylinder valves |
The 1965 welding literature, as compiled by S. A. Miller, put the normal case plainly: a flame that backfires into the neck or body of a torch normally becomes extinguished there. The abnormal case is what the arrestor industry was built on. The same passage records that the oxygen hose may ignite, possibly through hot carbon particles carried down the line, and that a hose-protector non-return valve will not prevent an explosion travelling back if a backfire occurs when there is already a mixture of oxygen and acetylene in one of the hoses. Patents of the early 1950s for hose-mounted arrestors (Schnautz, GP 814,588, 1951) already combined a sintered element with non-return and blow-off valves.
Rules that matter
A flame sits steady on a nozzle only while the gas leaves faster than the flame can burn back into it. The flame-stability work of Gaydon and Wolfhard (1960) describes this as a wedge bounded by blow-off, where the gas is too fast and the flame lifts away, and strike-back, where the gas is too slow and the flame walks back into the passage. Acetylene burns faster than any other common fuel gas; for the equimolar oxyacetylene mixture the calculated burning velocity is about 6.2 m/s at 15 °C, roughly four times its laminar value in air. Its wedge is therefore narrow, and anything that slows the flow at the tip or heats the nozzle pushes it across the line.
The C.P.I. 1968 safety code recorded three mechanisms shown by experience:
Why a backfire happens at all
- An overheated nozzle The code's conclusion was that torch detonation is avoided most reliably by keeping the nozzle at a low temperature, by frequent cooling or by heat-lagged designs. A hot tip raises the burning velocity of the gas inside it and ignites it before it leaves.
- An obstructed nozzle When spatter or slag partly blocks the orifice, the mixture cannot leave freely, and oxygen, being at the higher pressure, climbs back up the acetylene duct beyond the mixing chamber. The code's remedy was to clean the nozzle frequently with a proper tool.
- A leak at the handle On an interchangeable-head injector torch, a poorly tightened joint between the tube and the handle lets oxygen into the acetylene passage. The resulting mixture inside the torch flashes back on lighting.
Every one of these ends the same way: a combustible mixture inside a passage meant to carry only one gas. Touching the tip on the work, running a large nozzle at too low a flow, and lighting with the oxygen already open produce the same condition. The detailed list of causes is on what causes flashback.
How fast it escalates
Once a premixed flame enters a tube it accelerates. Martin's high-speed photography of 1:1 oxyacetylene in a 5 cm tube (Physics of Fluids, 1958) showed the flame reaching 800 m/s within the first 30 cm and becoming a detonation within about a metre at reduced pressure; Bollinger and Edse measured run-up distances of only a few centimetres in a 15 mm steel tube at atmospheric pressure for the most sensitive compositions. Oxyacetylene detonation velocities run from about 1.9 to 2.8 km/s. A hose full of mixed gas is, physically, a fuse.
That is why the arrestor exists and why it is tested hard. IS 11006:2011 requires each arrestor to stop five flashbacks, one after another, of a static mixture of 32 to 35 % acetylene in oxygen at its maximum operating pressure (cl 11.6, Table 2). The quenching element works because an oxyacetylene flame cannot pass through passages narrower than its quenching diameter, extrapolated to about 0.12 to 0.16 mm at 1 atm from the work of Anagnostou and Potter (1959) and of Gaydon and Wolfhard. Egerton, Everett and Moore (1953) cautioned that even sintered metal plugs do not always operate for oxyacetylene, which is why the arrestor is type-tested rather than designed on paper. The mechanism is on how a flashback arrestor works.
What to do, in order
The C.P.I. 1968 code gives the order for ignition inside the torch: close at once, first of all the acetylene valve, then the oxygen valve. Its logic is that the fuel is what the internal flame is consuming. Sources do not all agree on this order: some current blowpipe manufacturers' instructions specify oxygen first for a sustained backfire. What every source agrees on is that both blowpipe valves are closed immediately, then the cylinder valves. The order of the two blowpipe valves is the blowpipe maker's instruction; the C.P.I. order is given below for reference only.
- Close both blowpipe valves at once, in the order the blowpipe maker gives (C.P.I. 1968: acetylene, then oxygen; some makers: oxygen first).
- Close both cylinder valves with the key left on the acetylene spindle. The C.P.I. code requires the cylinder valve to be closed at once in the event of the gas catching fire or a flashback in the pipe leading from it, which is why the key must be present.
- Do not relight. Let the blowpipe cool; if it is too hot to hold, it has had a sustained backfire.
- Find the cause. Check the nozzle for spatter and damage, the nozzle seat and head joint for tightness, the gas pressures against the nozzle chart, and the hoses for soot or swelling at the fittings.
- Inspect the safety devices. An arrestor whose temperature-sensitive cut-off has operated cannot be reset under IS 11006 cl 10.5 and must be replaced; one that has stopped a flashback is examined or replaced as its maker directs. A hose that has carried a flashback is scrapped.
- Purge before relighting. The C.P.I. code's rule for any oxygen blow-back is that a quantity of mixture remains in the line and the torch must never be relit until it has been eliminated; let each gas run for a few moments, separately, before lighting.
Where the protection belongs
A modern oxy-acetylene set carries two levels of protection on each gas: a torch-end non-return valve, and a flashback arrestor with non-return valve and thermal cut-off at the regulator outlet. In 1965 the literature recorded that reliance was usually placed on arrestors in the piping system and that hose-mounted arrestors were only proposed; practice has since moved the protection down to the regulator and the torch, and IS 11006 and ISO 5175-1 now test devices for exactly those positions. The difference between the two devices is explained on non-return valve vs flashback arrestor, and when to retire one on flashback arrestor testing and replacement.
Rules that matter
Rules to apply
- Treat every pop as a warning: cool or clean the nozzle before carrying on.
- A hiss or squeal from the handle is a sustained backfire: both blowpipe valves off at once, in the order the blowpipe maker specifies (C.P.I. 1968: acetylene, then oxygen; some makers: oxygen first).
- A bang in the hose is a flashback: blowpipe valves off, cylinder valves off, nothing relit until the cause is found and the devices checked.
- Fit arrestors to IS 11006 on both gases at the regulators, and non-return valves at the torch.
- Keep nozzles clean and cool, joints tight, and pressures to the nozzle chart.
KIGL manufactures flashback arrestors, and the standards behind them are summarised on the flashback arrestor standards page.
Backfire — frequently asked questions
Backfire, flashback or sustained backfire: how do I tell them apart?
A backfire is the flame momentarily burning back into the nozzle tip with a sharp pop; it either goes out or relights at the tip. A flashback is the flame travelling back through the mixer into the hose, the regulator or beyond. The first is a nuisance and a warning; the second is a fire inside your gas supply and is the event a flashback arrestor to IS 11006:2011 exists to stop.
What is a sustained backfire?
A sustained backfire is a backfire that does not go out: the flame keeps burning inside the mixing chamber or neck of the blowpipe, usually with a hiss or a high squeal, and the handle heats quickly. It is burning on the premixed gas inside the torch and will damage it within seconds. Shut both blowpipe valves at once, in the order your blowpipe maker gives. Codes differ: the C.P.I. 1968 code put acetylene first; some current makers give oxygen first.
Why do backfires happen?
The C.P.I. 1968 code records three mechanisms from experience: an overheated nozzle; an obstructed nozzle, which lets oxygen, at higher pressure, climb back up the acetylene duct past the mixing chamber; and a leak at the joint between the handle and an interchangeable head or injector. Touching the tip on the work, spatter in the orifice, and too low a gas flow for the nozzle size all produce the same conditions.
Which valve should be closed first after a flashback?
The one the blowpipe maker's instructions name, with both valves closed at once. No single order is settled: the C.P.I. 1968 code put the acetylene valve first, reasoning that the fuel is what the flame is burning, while some current makers give oxygen first. Then close both cylinder valves, let everything cool, and find the cause before relighting.
Do I have to replace equipment after a flashback?
Inspect it before reuse, and replace what the flame reached. An arrestor that has stopped a flashback may have a tripped thermal cut-off, which IS 11006:2011 requires to be non-resettable, or a sooted element; replace it as its maker instructs. A hose that has carried a flashback is scrapped. The regulator is examined, and the torch and nozzle are cleaned and checked for the leak or blockage that started it.
Does a non-return valve protect against backfire and flashback?
Not reliably. The C.P.I. 1968 code says dry non-return valves do not necessarily stop flashbacks, and the 1965 literature compiled by S. A. Miller records that a hose-protector check valve will not prevent an explosion travelling back once oxygen and acetylene are mixed in a hose. A non-return valve stops reverse flow; only a flame-quenching element stops the flame. Fit both.
Can a backfire ignite the oxygen hose?
It can. The 1965 welding literature records that the dangerous case after a backfire is ignition of the oxygen hose, possibly through hot carbon particles carried down the line. That is why the hose lining specification then required no ignition in oxygen at 360 °C, and why an arrestor on the oxygen side matters as much as on the fuel side.
References
- Bureau of Indian Standards. (2011). Flash Back Arrestor (Flame Arrestor) — Specification (IS 11006:2011, First Revision).
- C.P.I. (1968). Safety in the production and use of acetylene. Paris.
- European Industrial Gases Association. (n.d.). Code of Practice: Acetylene (EIGA Doc 123).
- International Organization for Standardization. (2017). Gas welding equipment — Safety devices, Part 1: devices incorporating a flame arrestor (ISO 5175-1:2017).
Enquiries
Request a quotation
State the gas or equipment, the quantity or capacity, and your location. KIGL will confirm the grade, the cylinder or plant size, and whether the location can be served from Noonmati, or through the agent network.