Deep Drawing Irregular Parts: A Stage Planning Guide
An irregular deep drawn part is any component whose wall, footprint, or depth changes around its axis — square shells, rectangular cans, L-shaped brackets, asymmetric covers. A round cup draws the same way at every angle. These do not. No two sections see the same material flow, so one corner can wrinkle while the opposite wall tears in the same stroke.
We plan them in stages. Simulate the flow before cutting steel, optimize the blank shape, set the draw-bead and blank-holder restraint section by section, then sequence the draws and anneal only where the metal demands it. On our 5–500 ton presses a part that looks impossible on paper usually comes off in three or four hits, holding wall thickness from 0.15 to 3.0 mm and ID tolerance at ±0.05 mm (±0.02 mm with ironing).
Why irregular parts break the round-cup rules
A round cup is symmetric, so one limiting draw ratio describes it. An irregular part carries a different effective draw ratio at every section. The tight corner wants more material than the long flat wall can spare, and the long wall wants to buckle while the corner wants to split.
Two failure modes fight each other. Too much draw-in and the flange folds into a wrinkle. Too little and the sheet tears at the punch radius. The window between them is narrow, and it moves as the part draws deeper. We hold wall thinning under about 10% of the starting blank thickness; past that the corner goes thin and the next hit opens it.
The full design picture — corner radii, wall transitions, relief features — lives in our deep drawing design guide. The process mechanics sit on the deep drawing process page.
Plan the flow before you cut steel
Older shops settled irregular parts by carving a wooden model of the finished shape, wrapping it in wax-soaked gauze, and unfolding the hardened gauze to guess the blank. The instinct was sound: you need to see the metal move before you build the die. The execution is now digital.
We run a DFM / CAE simulation (AutoForm, Dynaform, or ANSYS) ahead of any tooling. It predicts where cracks and wrinkles will appear and lets us iterate the geometry and die for almost nothing. Published work shows the payoff — one multistage plan dropped from 10 draws to 6 after a simulation-driven redesign, with lower tearing potential.
We still validate the first physical blank on the press, but the expensive guessing is gone.
Blank shape is the first lever
People assume an irregular part needs an irregular blank. Often the opposite is true. For a square cup, a circular blank is the optimum — it minimizes drawability variance and keeps the flange profile uniform.
For L-shaped or clover cups there is no tidy rule, so we solve the blank by iterative FEA. The software compares the target final shape to the simulated result, measures the geometric shape error at the boundary, nudges the blank edge, and re-runs. It usually converges in a handful of iterations. NURBS-based blank optimization has cut thickness variation after forming by 81.2% in reported work — the same part, far more even walls.
Material choice sets how much slack you have. Our limiting draw ratio reference runs DC04 at 2.1–2.2, 304 stainless steel at 1.85–2.05, and 5052 aluminium at 1.80–1.95. ASTM E517 defines the plastic strain ratio r used to rank sheet drawability. When the tight corner of an irregular part pushes its effective ratio past those numbers, we split the draw or reshape the blank rather than fight the metal. The material selection guide covers the grade trade-offs.
Draw beads set the local restraint
A draw bead is a groove cut into the binder that drags the sheet as it feeds, adding restraining force so material does not flood the cavity where it is not wanted. On a round part one bead does the whole rim. On an irregular part you vary the bead by section — height, profile, and spacing all change around the perimeter.
Where the wall tends to tear, add restraint to hold material back. Where the flange wrinkles, ease the bead so it can feed. Restraining force itself is modeled: an equivalent drawbead model is accurate enough for shallow or symmetric parts, but for deep non-symmetric parts a discretized bead model predicts the deformation and the press force better.
Surface finish of the bead matters more than people expect. In tests on SPCD steel, a rough bead (Ra 6.1) produced more tearing and wrinkling than a smooth one (Ra 0.15). We polish bead surfaces and hold sealing faces below Ra 0.4 µm. Die-surface and forming-limit guidance is compiled in ASM Handbook, Vol. 14B.
Blank-holder force is not a single number
The blank holder keeps the flange flat. Too little and it wrinkles; too much and it tears. On an irregular part the “right” force varies through the stroke, because different sections reach their critical moment at different punch depths.
Modern presses run an adaptive blank-holder force profile that changes as the slide moves. Production bands sit around 150–500 kN. An optimized profile suppresses wrinkling on the long wall without starving the corner, and closed-loop servo presses hold ID tolerance at ±0.02 mm. Our deep drawing calculators turn a part’s geometry into a starting force and bead map.
What the simulation has to show before we build
Before we release a die we want four things visible in the results, not just a green checkmark:
- Thinning stays under 10% of blank thickness everywhere, with the worst spot named.
- No element in the flange enters the wrinkle regime at any stroke position.
- The punch-corner strain stays below the material’s forming limit curve.
- Springback is predicted and the die faces are over-corrected to absorb it.
If any of those fail, we change the blank or the bead map and re-run. That loop is cheap on a screen and expensive on a mill, so we spend the iterations up front.
Sequence the draws, anneal only when needed
Complex geometry usually needs more than one draw. Between stages you may anneal to recover ductility — but every extra stage and every heat treatment adds cost, so the goal is the minimum that still forms cleanly.
Forcing too few stages just moves the failure to the punch corner on the last hit. We watch the limiting draw ratio per section and add a redraw the moment the effective ratio would exceed it. Our window is depth to 150 mm and wall 0.15–3.0 mm on presses from 5 to 500 tons; outside that we redesign the part or the process, not the force.
Four ways to form the part
The shape decides the route.
Hydraulic or mechanical press with steel or cast-iron dies carries production volume. For large irregular covers we build the cavity from chromium-iron with epoxy-and-iron-powder fill in the non-feature areas, which cuts machining dramatically.
Rubber (elastomer pad) forming uses one rigid punch and lets the rubber take the female shape. It is gentle on the sheet and good for low-volume, highly irregular blanks.
Sheet hydroforming pushes fluid pressure against a single die. It suits complex curvature and thin walls and keeps tool cost down.
Incremental and servo-formed routes use segmented or adaptive tooling for asymmetric profiles, giving localized control of draw-in. This has replaced the old drop-hammer copper-zinc die, which we no longer run for production.
The stamping vs deep drawing vs die casting comparison lays out when drawing wins over the alternatives.
What we hand back
You get a validated drawing plan: the blank shape, the bead map, the stage sequence, and the tolerances, with tools made in-house. Minimum order is 1,000 pieces and we quote within 24 hours. If your part is asymmetric and nobody else will touch it, send the drawing.
Frequently asked questions
How many drawing stages does an irregular part need?
Most irregular parts take two to four draws. The count follows the limiting draw ratio at the tightest corner: each stage can advance the wall only so far before the corner tears, and we add a redraw the moment the effective ratio would exceed the material’s limit — about 2.1 for DC04, 1.85–2.05 for 304 stainless steel, 1.80–1.95 for 5052.
Why does one corner wrinkle while another tears?
Because each section has its own draw-in rate. The flat wall feeds too much material and buckles; the tight corner feeds too little and splits. Draw beads and an adaptive blank-holder force rebalance the flow section by section until both stay inside the forming window.
What blank shape works for a square deep drawn cup?
A circular blank. It minimizes drawability variance and keeps the flange profile uniform, even though the final part is square. L-shaped and clover cups need an iteratively simulated blank instead, tuned by FEA until the wall thickness is even.
Can deep drawing replace welding for asymmetric shells?
Often yes. A drawn shell is one piece, so there is no weld seam to leak or corrode and no secondary joining step. For thin, sealed, or pressure-bearing asymmetric enclosures, drawing usually beats fabricated welding on both cost and reliability.
What tolerance can irregular deep drawn parts hold?
On our presses, ID tolerance runs ±0.05 mm and ±0.02 mm with ironing, wall thickness 0.15–3.0 mm, depth to 150 mm, and concentricity within 0.05 mm TIR on guided tooling. Sealing faces are finished below Ra 0.4 µm. These bands are defined once on the capacity page, which is the canonical tolerance reference.
Do you need heat treatment between draws?
Only when the metal has work-hardened to the point that the next draw would crack it. We anneal at the stage where ductility runs out, not on a fixed schedule, to keep the stage count and cost down.
Key point: Non-round shells wrinkle or tear at corners because material flow differs at every angle around the perimeter.
What we do not claim: We do not publish leak-rate figures, autoclave cycle counts or springback angles that we have not measured. The ±0.02 mm figure applies to ironing passes only, not to every geometry. 6xxx aluminium is used for shallow forming, not deep drawing. Prices, minimum order quantities and lead times are confirmed per drawing rather than published as fixed values.
Last reviewed: 2026-11-05 · Reviewed by Sam Zhang, Process Engineer
Where to go next for a non-round shell
| If you need to decide | Start here |
|---|---|
| How many reductions the shell needs | Deep drawing process types |
| Diagnosing a wrinkle or a tear | Defect troubleshooting |
| The numeric tolerance bands | Deep drawing capacity |
| Tool and die layout | Tool and die design |
| Whether the grade is drawable | Material selection for deep drawing |
