You pick a bevel geometry once and weld it for the life of the joint. Pick V where you should pick J and you will burn weld wire forever; pick J where V is fine and you spend tooling money you did not have to.
The right answer is rarely the one most spec drawings carry over from the previous project. This guide is two questions in one: which geometry to specify, and whether the J actually pays back once you run the numbers. Here is how I decide.
V bevel — the workhorse (walls 5 – 22 mm)
Symmetric single-side bevel, 30° + 30° = 60° included. Both pipe ends ground to 30° off the axis, fitted up with a small root gap, welded from one side.
| Strengths | Simplest geometry; every welder reads it instantly; accepted by API 1104, ASME B16.25, EN ISO 9692-1; cheapest tooling head |
| Weaknesses | Weld metal volume grows with the square of wall thickness — above ~22 mm the fill time becomes the bottleneck on the whole job |
| Use it for | Structural welding, low- and medium-pressure piping, ship hull plate up to ~20 mm, transmission line pipe up to typical Sch 80 |
If your wall is below 22 mm and the WPS does not call for something else, use a V. Plenty of crews try to be clever and lose money.
J bevel — half the fill on thick walls (walls 25 – 80 mm)
Curved bevel face with a near-vertical upper section opening into a radius near the root. The cut wraps inward so the open volume above the root is much smaller than a V.
| Strengths | Weld volume 30 – 50% less than an equivalent V; proportionally less fill time, wire and shielding gas; less shrinkage, less distortion |
| Weaknesses | Tooling more complex (J-prep head); needs a ground-flat fit-up; not every welder has welded one before |
| Use it for | Heavy-wall pipeline (offshore risers, refinery transfer lines), boiler headers, pressure-vessel longitudinal seams above ~25 mm wall |
The J is where most plasma-or-grinder shops stop and buy a real machine, because you cannot freehand a J inside tolerance. See how to machine RTJ grooves for the single-point tooling principle the J shares.
Compound (double-side V or J)
Same bevel applied to both faces of the joint. Only usable when access permits welding from both sides — vessel shells, plate-to-plate butts, hot-tap saddles, large-diameter manifold tees.
| Strengths | Total weld volume halves versus a single-side bevel of the same wall; symmetric shrinkage cancels distortion |
| Weaknesses | Both sides must be reachable; root preparation more critical (back-gouge or counter-V); two-side welder rotation needed |
| Use it for | Vessel longitudinal seams, large plate butts, anything where both sides of the wall are open |
If you have access from both sides, do not single-side it. Compound is one of the easiest savings on a fab job and I still see shops miss it routinely. See also double-sided plate beveling for the plate-side execution.
Picking by wall thickness (rule of thumb)
| Wall thickness | Geometry I would specify | Why |
|---|---|---|
| ≤ 5 mm | Square (no bevel) | Wall too thin to need it |
| 5 – 10 mm | 30° V | Fill volume trivial; simplest wins |
| 10 – 22 mm | 30° V (60° included) | Workhorse choice, code-friendly |
| 22 – 30 mm | 30° V or 35° J | J starts paying back the tooling cost |
| 30 – 60 mm | J (or compound if both sides open) | Fill time dominates — minimise volume |
| > 60 mm | Compound J both sides | Distortion and fill time both critical |
Use this as a starting point, not a substitute for the WPS. If the WPS quotes something different, the WPS wins.
What you are actually saving
The pitch for J-prep is straightforward: less open volume above the root, therefore less weld metal, therefore less arc time, less wire, less gas, less shrinkage. All of that is true on paper — and the saving grows with wall, because a V opens proportionally as the wall gets deeper while a J keeps the upper opening narrow.
Rough cross-section volume per metre of joint (carbon steel, 30° V versus 35° J, 3 mm land, 3 mm root gap):
| Wall thickness | V volume (cm³/m) | J volume (cm³/m) | Reduction |
|---|---|---|---|
| 10 mm | ~62 | ~58 | ~6% |
| 15 mm | ~140 | ~108 | ~23% |
| 20 mm | ~250 | ~165 | ~34% |
| 30 mm | ~564 | ~286 | ~49% |
| 50 mm | ~1,565 | ~620 | ~60% |
Read this table the right way. Below 15 mm the J saves almost nothing — do not bother. Above 25 mm the J saves a third to half the deposit, and from there the only question is whether the consumable and arc-time savings outpace the tooling and prep cost.
The cost stack you compare
For each joint you are weighing savings against incremental cost.
Savings (J vs V) per joint:
- Filler wire / electrode mass × USD per kg
- Shielding gas litres × USD per litre
- Arc time × loaded welder rate (USD per hour)
- Distortion-related rework probability × rework cost
Incremental cost (J vs V) per joint:
- One-time: J tooling head (USD 1,500 – 3,500 typical add-on)
- Per joint: longer cycle time on the beveler (J takes ~30% longer to cut)
- Per joint: tighter fit-up requirement → small extra fitter time
- Risk: WPS requalification if the J geometry is new to the shop
The break-even is (J head cost) ÷ (savings per joint). On a long production run the head cost vanishes after a few dozen joints. On a one-off project of ten joints it might never pay back at all.
Where the V-to-J break-even really sits
The wall-thickness table is the easy part. The economic break-even depends on the welding process and the consumable cost, and it shifts the line earlier than most fabricators expect:
| Material + process | J head pays back above… | Why |
|---|---|---|
| Carbon steel + SMAW | ~30 mm wall, or 100+ joints | Cheap wire, low arc cost per kg deposited |
| Carbon steel + GMAW / FCAW | ~25 mm wall | Higher deposition rate makes arc-time savings worth more |
| Stainless 304 / 316 + GTAW | ~20 mm wall | Wire $$, shielding gas $$, slow process |
| Duplex + GTAW / GMAW | ~18 mm wall | Wire and inspection both expensive |
| Inconel 625 / 825 + GMAW-P | ~12 – 15 mm wall | Wire USD 300+ per kg; one Inconel joint can pay back a head |
| P91 / P92 + GTAW | ~20 mm wall | Wire $$ and PWHT cost scales with deposit mass |
On a recent offshore project the client was welding Inconel 625 transition spools at 16 mm wall. The J head cost less than two and a half kilograms of saved Inconel wire — payback at the first joint. They had welded forty before they came to us, all V, and could not explain why the consumable budget was overrunning. The answer was on the drawing. Most clients do not re-run the economics when they switch material; they should.
Where the J fails to pay back
Three failure modes I keep seeing:
1. Hand-ground “J” tooling
A J ground freehand is not a J. The radius is wrong, the land wanders, and the welder fills the geometry he was given — which is wider than the geometry the drawing specified. The math assumes the cut equals the drawing. If the cut does not, the saving evaporates. J-prep only pays back if it is machined on a tool that holds the form. That means an ID-mount or split-frame beveler with a proper J head, not a grinder.
2. Loose fit-up
A V is forgiving on root gap. A J is not — the narrow upper opening gives the welder less room to manage misalignment. If the fitting crew is used to V tolerances and you switch them to J without a tool change, the first ten joints come back for rework. The wire savings on the remaining ninety joints are gone before lunch on day one.
3. Mixed-wall projects with one tooling pack
Real projects have a wall mix — 16 mm here, 22 mm there, 35 mm at the manifold. A J head sized for the heavy wall does not cut the thinner wall efficiently, and crews end up swapping heads constantly. On mixed projects the answer is often V everywhere except the heavy section, not J everywhere.
What this means for tooling
A V head ships standard with every pipe beveling machine we make. J, Y and compound heads are tooling options on the same machine body — typically a USD 1,200 to 3,500 add-on per geometry depending on OD range.
The decision is rarely “buy a different machine.” It is buy the right tooling pack for the joint mix you actually weld. Most clients buy too few heads at first and end up regretting it on the second project. For plate work, the same selection logic applies on a plate beveling machine. Common machines for heavy-wall and large-OD work:
- DMM hydraulic heavy edge miller for plate over 25 mm
- DMM-YG large pipe edge miller for large-OD heavy-wall pipe
- Standard split-frame and ID-mount bevelers for the rest — see the pipe beveling machine range
A simple decision tree
- Wall < 15 mm? Stay on V. The J does not save enough to matter.
- Wall 15 – 22 mm? Run the consumable math. On stainless / duplex / nickel alloy the J usually wins; on cheap carbon steel usually not.
- Wall > 22 mm? J or compound, depending on access. The V is the wrong answer here regardless of material.
- Both sides of the joint open? Skip J for compound — see double-sided plate beveling.
- Confirm the tooling can hold the geometry. If you are grinding the J, you do not have a J.
The wrong question is “V or J for the project.” The right question is “V or J for this joint.” On most real projects, the answer is different in different places. If you are sizing a beveler for a project with mixed walls or expensive alloy, send the WPS and joint mix to our quote form and we will price the machine plus the right head pack in one quote.
About the author — Dongmei Guo is founder of Kedes Machine. On the shop floor since 2010, she builds portable on-site machining and cold cutting equipment in Shenzhen and works directly with field crews in oil & gas, power, shipbuilding and high-purity tubing. She writes here when a question comes up on the workshop floor that is worth answering in full.