Chapter 5 of 10 · 10 min read
The root pass and purging
Make the root that decides a pipe weld's soundness: open roots, backing rings and inserts, the keyhole, common root defects, and how much gas and time a purge needs inside stainless pipe.
The root is the first pass of a pipe weld and the one that matters most. It must fuse both root faces and penetrate to the inside of the pipe, leaving a smooth, slightly convex bead there, all the way round, in every position from flat to overhead, while the welder can see only the outside of the joint. Everything laid on top of it depends on it, and a defect in it sits on the inside surface where the pressure, the fluid and the fatigue loads all act. Pipe welders are judged, and qualified, largely on their roots.
This chapter covers the ways a root is made, the defects that afflict roots and their causes, and, for stainless and alloy pipe, the purging of the inside of the pipe with inert gas so that the root is not ruined by air. It calculates how much gas a purge needs, and shows why small purge dams save time and money.
Ways to make the root
There are several ways to make a full-penetration root, each with its place.
| How it works | Where it suits | Watch for | |
|---|---|---|---|
| Open root, covered electrode | electrode run through the gap, melting both root faces | carbon steel pipe and pipelines; field work | penetration varies with gap; slag; skill |
| Open root, gas tungsten arc | torch melts root faces with filler added by hand | stainless, alloy and quality carbon steel; small bore | slow; needs purge on stainless; wind |
| Open root, short-circuit or controlled-arc wire | wire process with controlled metal transfer | carbon steel in shop and field | lack of fusion if parameters wrong |
| Backing ring | ring fitted inside the joint, fused into the root | heavier pipe where an open root is difficult | crevice inside; flow disturbance; not hygienic |
| Consumable insert | preformed filler ring set in the gap, melted by the arc | high-quality roots, often mechanised | accurate fit-up needed |
The open root, made without any backing, is the commonest. On carbon steel pipelines, covered electrodes designed for the purpose, often cellulosic types that penetrate deeply and freeze fast, are run downhill through the gap at high speed. On process piping, a gas tungsten arc root, slower but controllable, gives the smoothest, most consistent inside bead, and is the usual choice for stainless, alloy steel and any line where the inside surface matters. Modified short-circuit wire processes, with power sources that control each drop of metal, make open roots faster than the tungsten arc and are widely used on carbon steel.
Backing rings make an open root easier but leave a ring inside the pipe, with a crevice between it and the wall, which collects dirt and corrosion and disturbs the flow. They are now limited to services where that does not matter. Consumable inserts, rings of filler shaped to fit the gap, melted by a tungsten arc to form the root, give excellent roots when fit-up is accurate, and are used with mechanised heads.
The keyhole
A welder making an open root watches the keyhole: a small hole at the front of the pool, where the arc has melted through both root faces and the metal is about to close behind it. A steady keyhole of the right size shows that the root is penetrating; if it grows, the root is about to burn through or sag; if it disappears, penetration is being lost. The welder controls it with travel speed, electrode angle and, on gas tungsten arc roots, the timing of filler additions. The keyhole changes as the welder works round a fixed pipe, because gravity pulls the pool into the pipe at the top and out of it underneath, and that is why the overhead part of the root, at the bottom of a fixed pipe, is where root defects cluster.
Root defects
Most root defects have simple causes, and most of the causes are in the fit-up or the technique rather than the welder’s luck.
| What it looks like | Usual causes | |
|---|---|---|
| Lack of penetration | root faces not melted through; a gap or notch inside | tight gap, thick root face, low current, fast travel |
| Lack of root fusion | one root face not fused | high-low misalignment, arc directed to one side |
| Concave root or suck-back | inside bead below the pipe surface | overhead position; too much heat; too wide a gap |
| Burn-through | hole or heavy drop-through inside | wide gap, thin root face, too much current |
| Excessive penetration or icicles | lumps of weld metal hanging inside | too slow, too hot, too much filler |
| Oxidised root | dark, crusty, granular inside bead on stainless | no purge, poor purge or purge removed too soon |
Lack of penetration and lack of root fusion are the most serious, because they leave a sharp, crack-like notch on the inside surface that fatigue and corrosion can grow from. Concavity in the overhead region is common and often acceptable within limits. Burn-through and icicles disturb the flow and, on hygienic lines, trap product. The oxidised root on stainless, sometimes called sugaring because of its granular look, is a defect of a different kind: the inside of the root has burned in air, and it has lost the corrosion resistance that stainless was chosen for.
The hot pass
On many procedures, especially pipelines with fast downhill roots, a second pass, the hot pass, is laid quickly over the root while it is still hot. It burns out slag trapped at the edges of the root, fuses any small lack of side fusion, and adds enough metal that the root will not crack when the pipe is moved or the line-up clamp released. The root is often ground lightly first to remove the high spots and expose any slag lines, so that the hot pass meets clean metal.
Why stainless needs a purge
When the root of a stainless pipe weld is made, the inside of the root is as hot as the outside, and if the inside is open to air, its chromium oxidises heavily. The result is a dark, rough, scaled inside bead, and beneath it a layer of metal depleted of chromium, which corrodes readily. On a product line, the oxidised root also traps product and bacteria and cannot be cleaned properly. The cure is to fill the inside of the pipe with an inert gas, usually argon, before and during the root and first passes, so that the hot metal inside meets no oxygen. That is purging.
The quality of a purge is judged by the colour of the inside root after welding and, before welding, by the oxygen content of the gas inside the pipe, measured with an oxygen meter at the vent. The procedure sets the oxygen level to be reached before welding starts. Some industries, notably food, dairy and pharmaceutical, use reference charts of acceptable weld colours to judge the inside of welds against.
How much gas a purge needs
Purging works by flushing the air out of the space to be purged with a steady flow of argon, entering at one end and venting at the other. The air is not pushed out like a piston; it mixes with the argon and is diluted, so the purge needs several times the space’s volume of gas before the oxygen falls low enough. The number of volume changes depends on the arrangement and the oxygen level wanted; the oxygen meter, not the calculation, says when it is done.
- Whole 12 m length
- internal volume about 28.3 L; flow 15 L/min
- Dams 300 mm apart either side of the joint
- internal volume about 0.71 L; flow 10 L/min
- whole length: 28.3 × 5 ÷ 15 ≈ 9.4 minutes and about 140 L of argon
- between dams: 0.71 × 5 ÷ 10 ≈ 0.35 minutes, about 21 seconds, and about 3.5 L of argon
- the dams cut the purge gas by about 97% and the waiting time by nine minutes per joint
Purging a whole length or spool is simple but slow and wasteful; purging only the region around the joint, between two dams, takes seconds. Purge dams are made of soluble paper or film that dissolves when the line is flushed, of inflatable bladders pulled out afterwards, or of reusable plugs on accessible ends. The saving is large on a job with hundreds of stainless joints, though the dams cost money and time to fit, and soluble dams must be fully flushed out before the line enters service.
During welding the purge flow is usually turned down, to avoid pressurising the inside of the pipe, which would push the root out or blow the pool, but kept on until the root and often the second pass are complete, because the inside surface stays hot enough to oxidise until then. Taking the purge away too soon gives a root that looks clean at the start and oxidised at the end.
Purge practice
Argon is heavier than air, so in a vertical pipe or a low point it fills from the bottom, and the vent should be at the top. In horizontal pipe it settles along the bottom first, so the vent and inlet are placed to flush the whole section. The gap of the joint itself is taped over until it is about to be welded, so that air is not drawn in and argon is not wasted, and the tape is peeled back ahead of the root as the welder progresses. Leaks at dams, unsealed branches and open valves let air in; the oxygen meter shows them as a level that will not fall.
Nitrogen and nitrogen-hydrogen mixtures are used as purge gases for some austenitic stainless steels where the procedure allows them, and are cheaper than argon; they are not suitable for every grade, particularly duplex and ferritic stainless and titanium, where the procedure must say what is permitted.
The purge gas hazard
Argon and nitrogen are not toxic, but they displace oxygen, and a person who enters a space full of them can lose consciousness in a few breaths without warning. Purging large pipes, vessels and tanks fills them with gas that will kill anyone who climbs in. Purge gas also escapes from vents and open ends into pits, trenches and low-lying areas. The contractor treats any vessel or large pipe that has been purged as a confined space, ventilates and tests it before entry, and vents purge gas where it cannot collect around people. Safe Work Australia’s model codes for welding processes and confined spaces cover these risks.
Carbon steel and purging
Carbon steel roots do not need purging, because carbon steel’s oxide does not matter to its corrosion resistance in the same way, and an open root on carbon steel with a covered electrode or wire forms its own protective slag or is shielded from outside. Some alloy steels and nickel alloys, and every titanium weld, need purging like stainless. The procedure states it, and where a job mixes carbon steel and stainless lines, the welding supervisor marks on each drawing which joints need a purge, so that no stainless joint is made without one by a welder working from habit on carbon steel.
Root inspection
The inside of a root can be seen only from inside the pipe, which on small pipe means after the fact, with a borescope or a camera on a cable, or by inspecting the ends of spools before they are joined. On hygienic lines, borescope inspection of a sample of welds, or of all of them, is normal, and the recordings go into the weld record. On other lines, radiography or ultrasonic testing (Chapter 9) examines the root through the wall. The contractor’s welders inspect their own roots wherever they can see them, by looking into the open end of a spool or with a mirror and light, because finding a bad root before the fill is laid saves cutting the whole weld out later.
The running case: the stainless roots
This is an illustration. The dairy’s stainless product and cleaning lines had about 300 butt joints in hygienic tube and DN 50 pipe. The contractor’s procedure called for a gas tungsten arc root and second pass under an argon purge, with soluble dams either side of each joint, an oxygen meter reading below the procedure’s limit before welding, the purge maintained through both passes, and borescope inspection of every product-contact weld against the client’s colour reference. On the first week’s welds, the borescope showed several roots with light straw-to-blue colour beyond the reference, traced to welders peeling the gap tape back too far ahead of the arc. With that habit corrected, the colour stayed within the reference for the rest of the job.
A failure: the unpurged branch
This is an illustration. On an earlier project, a welder added a short stainless branch to a cleaning line late in the job and, because the main line had already been flushed and the purge equipment packed up, made the root without a purge. The weld looked good from outside. Within months, the cleaning chemicals had attacked the oxidised inside root, and the branch was weeping at the weld.
The oxidised root had lost its chromium-rich surface and corroded in a service the pipe itself resisted easily. The branch was cut out and rewelded under a purge. The contractor now keeps a purge kit with each stainless welding set for the whole job, so that late changes are welded the same way as the first joints.
Exercise: purging a product line
A stainless product line is 38.1 mm tube with a 1.6 mm wall. A welder will purge between soluble dams set 250 mm apart, at 8 L/min, assuming five volume changes. Estimate the internal volume, the purge time and the argon used before welding, and compare them with purging a 6 m length at 12 L/min.
The bore is 38.1 − 2 × 1.6 = 34.9 mm. Between the dams, the volume is π ÷ 4 × 0.0349² × 0.25 ≈ 0.00024 m³, about 0.24 L. Five volume changes take 0.24 × 5 ÷ 8 ≈ 0.15 minutes, about 9 seconds, and about 1.2 L of argon.
The whole 6 m length holds about π ÷ 4 × 0.0349² × 6 ≈ 5.7 L, so five changes take 5.7 × 5 ÷ 12 ≈ 2.4 minutes and about 29 L of argon. On small tube the whole-length purge is not slow, but the dams still cut the argon by about 96%, and over hundreds of joints that is cylinders of gas. In either case the oxygen meter, not the arithmetic, decides when the root can begin.
