Ganesh Multitech EngineeringGANESH MULTITECHENGINEERING PVT. LTD.
Bending, welding and finished assemblies

Hemming and seaming: designing a folded joint that holds

Hemming and seaming join two panels by folding one edge over the other. Hems that read through paint, open in service and rust inside, and the numbers that prevent each.

GANESH MULTITECH · ENGINEERING DRAWING SETGM-BEN-952FLAT PATTERN · BA COMPENSATEDR = 2t90°BEND SECTIONNEUTRAL AXIS · K-FACTORK = 0.3–0.5 BY R/tFLAT LENGTHFLANGEBEND LINEK-FACTORSPRINGBACKBRAKE200 T / 4,450 mm5-AXISDelem controlTHICKNESSto 10 mmPROCESS ROUTECUTDEBURRBENDCHECKCOATPART / PROCESSBENDING WELDING ASSEMBLIESMATERIAL / FINISHCRCA · STAINLESS · ALUMINIUMSCALE 1:1 · REV A · ISO 9001 / 14001GANESHMULTITECH.COM/BLOGDRAWN IN-HOUSE · NOT TO SCALE
The pre-hem sets the radius; the closing tool only flattens the fold around the inner panel edge. Ganesh Multitech.

Key takeaways

  • A closed hem stack is 2t outer plus t inner, so a 0.7 mm skin over a 1.2 mm panel closes into 2.6 mm and the closing tool must be cut for that number.
  • Keep the inner panel edge 1.5 mm back from the hem shoulder; at 0.5 mm the hem line prints through 70 µm of powder as a visible ridge.
  • A 600 mm hem on 0.7 mm CRCA needs about 70 kN to air-bend and 100–110 kN to close flat, under 6% of a 200 T brake, so bed deflection governs the hem, not tonnage.
  • A lock seam holds by interlocking 5 mm hooks and stays airtight only with mastic in the fold; a bare seam leaks under a 2 kPa air decay test.
  • Above 20% outer-panel elongation a flat hem closes at 0.5t radius; below it, use an open hem or a rivet-bonded joint instead of forcing the fold.

A 0.7 mm outer skin folded over a 1.2 mm inner panel, closed flat and powder coated, is the joint on elevator car doors, genset canopy skins and machine guards. It is also the joint that comes back six months after launch with a line across the painted face where the fold sits, or with rust bleeding out of the hem, because the closing die was cut for the skin thickness and nobody calculated the stack that passes through it.

Hemming and seaming both join two panels by folding one edge over the other. A hem wraps: pre-hem to about 100°, close to flat, trap the inner panel edge in the fold. A seam interlocks: two 5 mm hooks rolled into each other on a lock seam, or two upstanding edges closed into a standing seam. The distinction matters. A hem is a stiffened edge that happens to hold two panels. A seam is a joint chosen because it can carry a seal.

Three numbers decide whether either closes. The stack through the fold is 2t of outer panel plus t of inner. The flange that feeds the fold is at least 4t of outer panel. The stretch the outer surface can take is about 1.6 times the neutral-axis path of a 180° fold, which is why elongation on the mill certificate settles the argument before the tool drawing does.

Where hemming and seaming earn their place

Wrap the hem when the assembly is seen from one side, the inner panel edge has to be hidden, and the edge needs to stop oil-canning. Interlock a seam when both panels are the same gauge, the joint runs the full length of the part, and the requirement is a water path rather than a structural one. Weld when the joint carries load in peel. Bolt when it has to come apart in service.

A hem is the cheaper joint on a door skin because it does three jobs in one stroke: it buries the cut edge, doubles the material at the boundary, and locates the inner panel without a fixture. At 0.7 mm, doubling the edge takes local stiffness up by roughly eight times over the same width, less the fold radius. That is what stops the face from drumming when the door is handled.

A seam is the cheaper joint on a duct or a long housing because it forms continuously, without a die matched to the part length. It is not a weld. A lock seam on 0.6 mm galvanised sheet holds two panels mechanically, leaks unless mastic sits in the fold, and offers little resistance to a load that tries to peel the hooks apart.

The limits that decide a hem or a seam

The values below are for CRCA and galvanised sheet with a matched two-stage tool. Aluminium moves the radius up, stainless moves it up further, and pre-coated stock fails at the fold whatever the base metal does.

ParameterTypical valueGanesh limit or note
Outer panel, closed hem0.5–1.2 mm0.8–1.5 mm on the 200 T brake; above 1.5 mm a flat hem needs a forming press
Inner panel thickness0.8–2.0 mmKeep inner at or below 1.7× outer; 1.2 mm inside 0.7 mm is the common door stack
Stack through the hem2t outer + t inner2.6 mm on 0.7 + 1.2 mm; 3.2 mm is the flat-hem ceiling
Flange before pre-hemat least 4t outer2.8 mm at 0.7 mm; below 2.5 mm the edge curls instead of closing
Pre-hem angle100–110°Radius nose sets the fold; a single-hit close splits the crown
Hem gap, closed0.05–0.15 mm0.15 mm on laser-cut edges; above 0.25 mm a shadow line shows
Hem width after closingflange minus 2.95t8 mm flange on 0.7 mm gives 5.9 mm
Inner edge to hem shoulder1.5 mm minimumBelow 1 mm the edge prints through 70 µm of powder
Sealer bead, applied2–3 mm diameter0.2–0.4 mm cured film; squeeze-out must stay inside the part
Lock seam hook width5–6 mm0.6–1.0 mm galvanised; mastic in the fold for an airtight joint
Standing seam height12–25 mmRoof and wall panels only; never a substitute for an enclosure weld
Closing force, 600 mm hem70 kN air bend, 100–110 kN flat178 kN per metre closing; tonnage is 5% of a 200 T brake, so marking and deflection govern

The stack number is the one that catches designers. A hem is drawn as a single thickness of the outer skin, but the tool closes around three layers. A 0.7 mm skin folded over a 1.2 mm inner panel needs a die gap of 2.6 mm plus clearance, and a tool cut at 2.0 mm will coin the fold, thin the crown and leave the hem edge 0.3 mm short of the inner panel.

Why the hem line reads through the paint

The hem line appears because the fold is a step, and a 70 µm powder film copies a step of 0.15 mm faithfully. Three mechanisms produce it, and they are fixed in different places.

The first is gap variation. A hem closed to 0.05 mm at one end and 0.3 mm at the other leaves a wedge that shows as a soft line under a gloss finish. The cause is usually a flange that was sheared rather than laser cut, so the blank edge wanders by 0.2 mm along 600 mm. Cutting the flange on the 12 kW fibre laser with the 40 ft × 16 ft bed holds that edge to ±0.1 mm and removes the variable before the tool sees it.

The second is the inner panel setback. If the inner edge sits 0.5 mm from the hem shoulder, the closing tool has to bend the outer skin around a hard line rather than a radius, and the mark transfers straight through. At 1.5 mm setback the fold has somewhere to flow, and the line softens.

The third is springback. A 180° fold in 0.7 mm CRCA springs back 2–4° when the tool releases, which opens the free edge by 0.1–0.3 mm. Coining the hem in a second station, or closing with a grooved tool that over-bends by 3°, holds the edge shut. Without it, the hem looks closed on the bench and opens after the coating oven because the steel and the powder expand at different rates.

Working out the stack and the closing force

Take a 600 mm long hem on 0.7 mm CRCA over a 1.2 mm inner panel. The stack is 2 × 0.7 + 1.2 = 2.6 mm. The finished hem width from an 8 mm flange is 8 − 2.95 × 0.7 = 5.9 mm, so the visible band is under 6 mm and the inner panel edge must sit at least 1.5 mm inside it.

Closing force follows the air-bend relationship, with the die opening set near 4t for a tight fold:

F = 1.42 × σb × L × t² / (1000 × V)

With σb at 450 MPa, L at 600 mm, t at 0.7 mm and V at 2.8 mm, that gives 67 kN to bend the flange to 90°. A 180° closing stroke crushes the crown rather than bending it, so multiply by about 1.6: roughly 107 kN, or 178 kN per metre of hem.

The useful part of that number is what it is not. On a 200 T brake, 107 kN is under 6% of rated capacity. Tonnage never decides a light-gauge hem. What decides it is bed deflection over the hem length, tool marking on the visible face, and whether the part can be handled at both ends. A 3,000 mm hem needs 534 kN, still a quarter of the press, but the bed has to be shimmed or the hem closes tight at the ends and open in the middle by 0.2 mm.

Seaming: the lock seam and the standing seam

A lock seam is formed in two stages on a roll former. The edges are hooked to about 5 mm, engaged, then locked and flattened so the finished seam is roughly three material thicknesses wide. On 0.6 mm galvanised sheet the seam develops 60–70% of the parent metal's tensile strength along the joint, which is enough for a duct or a housing panel and not enough for a bracket.

It is not airtight by construction. The fold is a capillary on both faces, and a bare lock seam pulls air through under a 2 kPa decay test within a minute. Where the housing has to keep dust or water out, mastic goes into the hook before locking, and the seam becomes a sealed mechanical joint rather than a weld. Standing seams are taller, 12–25 mm, and belong to roofing and cladding. They shed water and they are not an enclosure joint.

What Ganesh runs, and what we do not

Hems run on the 200 T, 5-axis CNC press brake with the 4,450 mm bed, in two stages: a radius nose at 100–110°, then a flat closing tool cut to the stack. Blanks come off the 12 kW fibre laser with the 40 ft × 16 ft bed, so the flange that feeds the fold is a laser edge at ±0.1 mm, and the first article is checked for hem width, hem gap, straightness and inner-edge setback before the run continues.

We do not run roller hemming, a dedicated hemming press, or a lock-seam roll former. A drawing that calls for a rolled hem on a painted skin, or a lock seam on a housing, comes back as a different joint: a folded return with a spot row at the manual and SPM stations, or a continuous seam on one of the three CLOOS robotic MIG cells. Both are joints we can inspect. A hem closed in a brake can be checked with a wedge; a rolled hem on a finished skin cannot be opened without destroying it.

Where the hem has to sit under a coating, the sequence is hem, then seven-tank pretreatment, then powder in the 18 ft × 14 ft booth. Coating before hemming is not an option on a closed fold, because the film cracks at the crown and the crack is the start of a rust line. Adhesion is checked by cross-hatch, and film thickness by DFT gauge, on the same first article as the hem dimensions.

The bending capability page sets out the brake and tool envelope, and the welded assemblies route covers what happens when the hem is replaced by a weld. Door and panel work of this shape sits in the elevator parts range.

What to put on the drawing

  1. Outer and inner thickness with grade, because the stack and the elongation limit both come from those two numbers.
  2. Hem type and finished width, not the flange length alone. A hem drawn at 6 mm and a flange drawn at 6 mm produce different parts.
  3. Inside radius at the fold, and which face the fold closes toward, so the cosmetic side is not marked by the closing tool.
  4. Sealer: type, applied bead diameter and cured film thickness, plus a note that squeeze-out must not pass the hem edge.
  5. Inner panel setback from the hem shoulder, 1.5 mm nominal, with the tolerance that keeps it off the fold radius.
  6. Hem gap and straightness, typically 0.15 mm and 0.5 mm per metre, with the measurement method.
  7. Stake or coin if the hem must stay closed after coating, and the test: a wedge check at first-off and a peel check on a coupon from each shift.

The hem folding guide owns the single-panel fold rules, including the 4t flange and the 2.95t allowance, and this page stops where that one starts. If the assembly is held by spots rather than by the fold, the spot welding guide covers nugget size and edge distance, and the enclosure welding guide covers what changes when the joint has to be airtight.

When a hem or a seam is the wrong joint

A closed hem is wrong on pre-coated stock. The fold cracks the film at the crown on any radius, and the crack is on the outside where it collects water. It is wrong when the outer panel's elongation is under 20%, which covers 5052-H32 aluminium, most high-strength low-alloy grades above 550 MPa yield, and any sheet where the certificate is borderline. Those parts need an open hem at 1.5–2t, a rivet-bonded joint, or a folded return with a weld.

It is wrong above 3.2 mm of stack, where the force and springback turn the fold into a press-forming job. It is wrong when the hemmed edge is a datum, because a hem carries ±0.5 mm on width and a rounded crown; locate from a hole instead.

A seam is wrong wherever the joint has to take peel, internal pressure, or repeated opening. A lock seam is a folded interlock with no fusion, so it cannot be certified as a structural joint and it cannot be repaired without cutting the hooks out. If the enclosure has to pass an ingress test, the seam becomes a weld plus a gasket, and the RFQ route should carry the test method on the drawing rather than in a covering email.

Send the two panel thicknesses, the grade, the visible face and the seal requirement. The bending and welded assemblies hub carries the rest of the joint rules, including the tolerance stack that decides whether a hemmed door still fits its frame after coating.

Frequently asked questions

What is the difference between hemming and seaming?

Hemming folds one panel edge back over itself or over an inner panel and closes it flat. Seaming interlocks two folded hooks, as in a lock seam, or closes two upstanding edges into a standing seam. A hem stiffens an edge; a seam is a joint that can also carry a seal.

How thick can a closed hem be?

The stack is 2t outer plus t inner. On a 0.7 mm skin over a 1.2 mm panel that is 2.6 mm, and 3.2 mm is the practical ceiling for a flat hem on a press brake. Above that the fold needs a teardrop form and a forming tool.

Why does the hem line show through powder coating?

Because the fold is not a straight edge. A hem gap that varies by more than 0.2 mm, an inner edge closer than 1.5 mm to the shoulder, or a hem that springs open after the tool releases all leave a step that a 70 µm film copies and gloss makes visible.

Can a hem be watertight?

Only with sealer applied in the fold before closing. Specify a cured bead 0.2–0.4 mm thick, keep the squeeze-out line inside the part, and leak test the assembly rather than assume the fold seals.

Do you need a dedicated press for hemming?

Not up to about 2.5 mm of stack. A two-stage tool on a conventional brake works: a 100° pre-hem with a radius nose, then a flat closing tool. Roller hemming is for skins that cannot take a die mark or panels longer than the bed.

Sources and further reading

Ganesh Multitech Engineering Team

Engineering and manufacturing team

The engineers, toolmakers, welding supervisors and quality inspectors who run the Ganesh Multitech plants at Mumbai (Vasai) and Nagpur. Every article on this site is written or reviewed inside this team before it is published.

Reviewed by Ganesh Multitech Engineering Team, engineering and manufacturing team. Last technical review 3 Nov 2026.

Have a hemmed or seamed assembly on the drawing?

Send the outer and inner panel thickness, the grade, the hem width and the sealer requirement. We will say whether the fold closes as drawn before the tool is cut.

Request a quote

The Ganesh engineering note

One useful email a month

Drawing checks, tolerance tables, coating failures and sourcing costs, written by the team that runs the plants. No daily drip.