Deep Drawn Stamping: Shells, Cans and Housings

Order a deep drawn shell and you are betting the whole part on a single press stroke. Get the draw wrong and the wall wrinkles near the flange, the base splits under the punch load, or the sidewall comes out thinner on one side than the other. Wrinkling, cracking and uneven wall thickness are the three failures that kill a deep drawn stamping program before it ever reaches volume. They are not random bad luck - they are the direct, predictable result of how much metal you are asking the blank to flow, and whether the tool, the lube and the material can deliver that flow without protest.

This guide is written for the buyer or engineer who has to spec a drawn shell and wants to know what the drawing must say, what the tool has to do, and which supplier questions separate a draw house from a press shop that will take the job anyway. The process fundamentals below are the ones that decide first-article success: draw ratio, thinning, annealing, tool geometry, and the tolerances a drawn part can honestly hold. We have run drawn shells on the same floor as our progressive and heavy stamping work since 2014 - 21 presses in a 10,000 m² facility in Chang'an, Dongguan, with in-house wire EDM, CNC, and grinder capacity in the tool room - so the numbers here are the ones we quote against, not textbook averages.

What "deep drawn" actually means

Drawing is the operation where a flat sheet blank is pulled into a die cavity by a punch, turning a disc of metal into a cup, can or box. "Deep" drawing is the regime where the draw depth is a meaningful fraction of the part diameter - generally when the draw ratio pushes past the point where a single operation can form the part cleanly. A shallow drawn lid might be 2 mm deep in a 50 mm shell. A battery can or a shield can is often 2 to 4 times deeper than its diameter.

The number that decides everything is the limiting draw ratio (LDR). It is the largest ratio of blank diameter to punch diameter you can pull in one hit before the cup wall tears. For ordinary low-carbon steel the textbook LDR sits around 2.0 to 2.2 in a well-lubricated, well-tooled press. Brass runs similarly in the soft tempers. Stainless steel 304, with its higher work-hardening and springback, lands closer to 2.0 and is far less forgiving. That band - LDR 2.0 to 2.2 - is the wall you are pressing against. Exceed it and you need a redraw station, an intermediate anneal, or both.

How the metal moves

Watch the flange of a blank during a draw and you see material flowing inward. The annulus of metal outside the punch footprint is squeezed radially inward while being stretched tangentially. The material that ends up in the wall started life in the flat flange. The material that stays in the base is the central disc that never moved. This split matters because each zone is loaded completely differently.

The flange is under compression - that is what makes it want to wrinkle. The wall is under tension, pulling the metal down into the cup - that is what makes it want to tear. The punch radius is the worst spot: the metal there is bent, stretched and sliding all at once, and it is where most splits open. Understand those three zones and you can predict nine out of ten rejects on a new tool.

Wall thinning: the number buyers never ask about

Every deep drawn part leaves the die with thinner walls than the starting sheet. That is physics, not a defect. As the flange metal is drawn up over the punch radius and stretched into the wall, it thins. A reasonable single-draw thinning at the punch-radius wall runs 8 to 15 percent of the starting gauge. Push a 0.5 mm blank through a 2.0 draw and you should expect the critical wall section - just above the base, at the punch radius - to measure in the 0.42 to 0.46 mm range.

The danger is uneven thinning. If the blank is off-center, the binder pressure is uneven, or the punch walks in the guide, one side of the wall thins more than the other. A part that averages 0.44 mm but varies from 0.40 to 0.48 mm around the circumference is a reliability problem in a battery can or a pressure housing, even if the average looks fine. We spec a wall-thickness variation of ±5 percent of nominal as a working target; tighter than that needs tighter tool control and verified stock.

Lubrication and Blank-Holder Pressure: The Two Knobs

The flange wants to wrinkle, the wall wants to tear, and the two controls that balance those competing failures are lubrication and blank-holder pressure. They are the knobs the setup engineer turns between the first draw and the thousandth, and they are the reason a drawn shell that ran fine at the tool builder wrinkles at the production shop.

Lubrication. The lubricant has one job: keep the flange metal sliding under the blank holder and over the die radius without galling, while the punch stays dry enough to grip the base. The wrong lube either starves the flange - tearing the wall - or floods the punch, which then slips and thins the base. Practical practice by material: chlorinated extreme-pressure drawing oils for stainless 304 and 316, where the work-hardening demands the heaviest film; soap-based or emulsion compounds for brass, which draws cleanly with moderate lubrication; light mineral oil or a dry-film compound for copper, which galls if run dry. The lube window is part of the process spec, not a shop-floor free choice, and it shows up on the first-article report.

Blank-holder pressure. The blank holder presses the flange against the die face with a controlled force: too little pressure and the flange buckles into wrinkles, too much and the wall tears at the punch radius. On deep draws, the pressure is tuned against the material temper and the current draw stage, and draw beads are added where the flange needs extra restraint - long rectangular shells wrinkle at their straight sides no matter what the corners do. A production draw die includes the pressure adjustment in its design, not as an operator improvisation. The defect-by-defect breakdown of what goes wrong when these knobs are mis-set is catalogued in the metal stamping defects guide.

When one draw is not enough

You want a can that is 3x deeper than it is wide. The LDR tells you a single operation will not get there - 2.0 to 2.2 is your ceiling, and a 3x depth means a draw ratio well past 2.5. The answer is progressive redrawing. The first draw makes a shallow cup. A second draw narrows and deepens it. A third tightens again. Each step keeps its own local draw ratio under the limit.

The catch is work-hardening. Every draw cold-works the metal, raising its yield strength and lowering its ductility. By the second or third redraw, annealed copper or brass has stiffened enough that forcing another draw without relief will crack it. That is where intermediate annealing earns its place. A bright-anneal between redraws softens the part back toward its original temper, restores ductility and lets the sequence continue. For stainless 304 in multi-stage deep draws, an anneal after every one or two draws is normal, not exceptional. The anneal atmosphere matters for the surface: bright annealing under controlled atmosphere keeps copper and brass clean for plating or soldering, while air annealing grows an oxide that has to be pickled off before the next draw.

Tool steel, punch radius and die maintenance

The draw tool carries the part's entire fate, and its material and maintenance schedule are where drawn-shell programs live or die. Punch and die for deep drawing need high compressive strength and wear resistance at the radius, because the radius is where the metal slides hardest and the tool wears first. Short-to-medium runs use oil-hardening or air-hardening tool steels such as O1, A2, or D2; high-volume battery-can and shield-can programs move to powder-metallurgy tool steels or carbide inserts on the radius, where a wear set is measured in millions of parts instead of hundreds of thousands.

Two tool details deserve buyer attention. First, the punch radius: a radius under 2x the sheet thickness invites splits at the radius, and the drawing should carry a radius as large as the part allows. Second, the die cavity surface: the draw die face and radius are polished and kept polished, because every scratch in the die is a galling line in the wall. The maintenance schedule sharpens or re-polishes the radius at set intervals, and the in-house tool room does the work with wire EDM, CNC, and grinders rather than sending the die out for weeks. A supplier that cannot show a die-maintenance record is a supplier that will ship your rejects in batches.

Material choice: brass, stainless, copper

The part's job picks the metal, and the metal picks the process window.

Material	LDR in soft temper	Work-hardening rate	Finish off the tool	Best for	Watch out for
Brass C260 / C268	Up to 2.2	Low	Excellent	Sensor housings, connector bodies	Corrosion in harsh environments; cost
Stainless 304 / 316	About 2.0	High	Good, needs care	Battery cans, medical, marine	Springback, punch loads, anneal stages
Copper C110 / C102	Up to 2.2	Low	Excellent	Shield cans, RF housings	Softness in handling; oxidation when annealed in air
Low-carbon steel	2.0 to 2.2	Moderate	Good	Industrial cans, housings	Needs plating for corrosion

Brass (C260, C268)

Brass is the easy winner for deep drawing. Its high ductility and low work-hardening rate mean LDR values at the top of the 2.0-2.2 band are routine, and redraws are gentle. Half-hard or soft temper draws cleanly. The payoff is a shell with excellent surface finish straight off the tool - ideal for sensor housings and connector bodies where appearance and formability both matter. The limitation is cost and corrosion resistance; for harsh environments you pay up for nickel plating or move to stainless. The alloy family and its stamping behavior are detailed on the brass materials page.

Stainless steel (304, 316)

Stainless buys you corrosion resistance and strength but fights you on the press. Lower LDR, aggressive springback, high punch loads. A 304 battery can or shield can demands careful blank-holder pressure, good lube and often a draw-bead or improved binder geometry to control the flange. We budget more redraw stages and at least one anneal for any depth past moderate. The reward is a part that survives salt spray, solvents and temperature swing without a coating. The dedicated treatment of the stainless family is in the stainless steel stamping guide.

Copper (C110, C102)

Copper's drawability is excellent - it rivals brass for ductility and takes a beautiful finish. Its killer feature is conductivity, which is why it shows up in shield cans and RF housings where the shell itself is part of the circuit. The headaches are softness (dents in handling) and oxidation if annealed in air. Bright annealing under controlled atmosphere keeps the surface clean for plating or soldering.

Tolerances you can actually hold

Deep drawn parts do not hold the same tight tolerances as machined parts, and pretending otherwise wrecks quotes. A realistic tolerance table for drawn shells looks like this:

Feature	Realistic tolerance	What drives it
Outer diameter, small precision shells (under 30 mm)	±0.05 to ±0.10 mm with trimmed flange	Die bore and punch fit; springback
Outer diameter, larger cans	±0.15 mm comfortable	Tool size and material spring
Wall thickness	±5 to ±10 percent of nominal	Draw thinning, not tool size
Depth	±0.10 to ±0.25 mm	Trimmed top versus drawn flange
Bottom thickness	10 to 20 percent thicker than the wall	The base does not draw down
Concentricity	Under 0.08 mm wall-to-wall variation	Punch guidance and centered feed

Any tighter than that and you are paying for secondary reaming, turning or grinding - which is fine, but call it out in the print so the cost lands in the right column. The broader tolerance system for stamped parts, including how to state it on the drawing, is in the precision stamping tolerances guide.

Four parts that live or die by the draw

Battery cans

The cylindrical cell can is the textbook deep draw: a long, thin-walled cup with a tight base, drawn 3 to 4x its diameter. LDR discipline, wall-thinning control and a clean punch radius decide whether you get 200 ppm rejects or 2 ppm. Stainless and nickel-plated steel dominate; the base must stay thick enough to take the crimp without splitting. The contact and terminal side of the cell package is covered in the battery contact stamping guide.

Sensor shells

Pressure and temperature sensor housings are small, often squat shells with a sealed base and a thin wall carrying the electronics. Brass and stainless both appear. The draw here is shallow-to-moderate, so a single operation often suffices, but concentricity and a burr-free trimmed edge matter because the shell is later welded or brazed. The sensor part family gets its own treatment in the sensor stamping parts guide.

Shield cans

EMI shield cans are thin, boxy or round drawn shells that clip onto a PCB. Copper and nickel-silver are common for conductivity; the walls are thin (0.15 to 0.3 mm) so wrinkling at the flange is the daily enemy. Tight blank-holder control and the right lube window keep the walls flat and the corners crisp. The RF-side requirements for the same parts, including the seam and ground rules, are in the metal enclosure and shield can guide.

Connector housings

Connector bodies are drawn shells with tight ID control so the insulator and contacts seat press-fit. Brass is the default for formability and plating acceptance. The challenge is holding the ID after springback - stainless connectors need a slightly undersized punch to spring back into tolerance. The housing family is worked through in the connector housing stamping guide.

Cost economics: drawing vs. turning vs. casting

A shell can be drawn, machined from bar, or cast, and the route is a volume and geometry decision. Drawing needs tooling and then prints parts fast; turning has no tooling but pays for chip time per part; casting carries a heavy die and a porosity risk that machining cannot fix.

Route	Tooling cost	Per-piece cost at volume	Wall thickness	When it wins
Deep drawing	Moderate to high	Lowest	Thin, 0.05-3.0 mm stock	From the low thousands of parts upward; thin-wall shells at volume
CNC turning	Low	High	Any, but chip-waste heavy	Prototypes, short runs, stepped or threaded interiors
Casting	Highest (die casting tool)	Low at very high volume	Thick, porosity limits thin walls	Very large volumes with complex external shapes and no thin-wall need

Deep drawing wins the thin-wall argument outright. A drawn 0.3 mm stainless can is stronger than a 2 mm cast wall of the same material for a pressure application, because the drawn wall is fully dense wrought metal while the cast wall carries porosity. Drawing also holds better surface finish on the ID, which matters for cans that must seal against a gasket or carry a pressed-in component. The crossover numbers follow the same logic as any stamping buy: below a few thousand pieces, turning or fabrication; above that, the draw tool amortizes and the piece price collapses. The full quoting logic, including where the tooling lands in the price, is in the how to quote a stamped part guide.

Secondary operations that complete the shell

Almost no drawn shell ships raw from the die. The sequence that turns a cup into a finished component is where the real program scope shows up, and it is also where a single-source supplier earns its keep.

• Trimming. The drawn flange edge is trimmed to length and squareness; this sets the depth tolerance the drawing quotes.
• Piercing. Side holes and base holes are pierced after drawing, when the wall is at its final thickness; piercing before drawing distorts the hole.
• Threading and knurling. A threaded neck or knurled OD converts the shell into a fitting; done in a secondary operation because the draw cannot hold thread geometry.
• Plating and finishing. Selective reel-to-reel plating places gold, silver, tin, or nickel at 2-8 µm on the zones that need it, with zinc at 5-12 µm for steel hardware and salt spray per ASTM B117 verifying corrosion claims.
• Assembly and molding. Insert molding and overmolding seal electronics into the shell, and press-fit subassembly joins the drawn part to its mating components. The assembly services page covers the subassembly side of a drawn-shell program.

Prototyping a deep drawn part

Deep drawing is the stamping process where prototyping deviates most from production, so the prototype plan matters. A prototype can be machined, spun, or drawn on soft tooling, and each route predicts production behavior differently. Machined prototypes tell you nothing about thinning, wrinkling, or springback; drawn prototypes on soft tooling use the same blank size, lube, and draw ratio as production, so the first article genuinely predicts the production part. Our rapid prototyping runs follow the production draw sequence at low speed, and the first articles are sectioned and measured for wall thickness before any volume commitment. Buyers who skip this step discover the wall-thinning surprise on the production first-article report, when the fix costs tooling rework instead of a drawing note.

QA that catches the real failures

Wrinkling you see by eye. Cracking you see by eye or by a dye-penetrant on the base and radius. Uneven wall thickness you do not see at all unless you section the part. Our first-article routine cuts a cross-section, mounts it and measures wall thickness at the flange, the radius and the straight wall on both the major and minor axis. That single sectioned part tells you more about tool health than a hundred caliper readings of the OD. In production we track wall thinning and concentricity by sampling, because those are the numbers that drift as the punch radius wears.

The quality lab backs the routine with CMM and optical measurement for the dimensional features, and the process control system behind it is IATF 16949:2016, with ISO 14001:2015 covering the environmental side of the plating and annealing lines. First-article inspection, material certificates with heat and temper, and lot traceability from coil to packed box are the default deliverables, not the premium package. The quality system itself is described on the quality page.

Design rules that save the program

• Keep the corner radius at the punch and die generous - a radius under 2x the sheet thickness invites splits at the radius.
• Limit any single draw to an LDR at or below 2.0 for stainless, up to 2.2 for brass and copper in soft temper.
• Plan an anneal into the sequence the moment cumulative reduction passes roughly 60 to 70 percent of the original blank area.
• Control the flange: blank-holder pressure, draw beads and lube together decide whether you wrinkle or tear.
• Specify wall tolerance as a percentage, not an absolute, and verify it with a cut-and-measure on the first article, not a caliper on the outside.
• Call out the secondary operations - trim, pierce, plate, assemble - on the print so the cost lands in the right column.

Bottom line for sourcing

Deep drawn stamping is cheap at volume and unforgiving in setup. The part that looks simple on a screen - a cup, a can, a shell - is a stack of material-flow decisions: how much you draw, how often you anneal, how tightly you hold the flange, how much wall you are willing to lose. Get those right and you ship a shell that is lighter, stronger and cheaper than anything turned or cast. Get them wrong and the rejects arrive in boxes.

If you are scoping a battery can, sensor shell, shield can or connector housing and want a draw sequence, material recommendation and a realistic tolerance and cost window worked out before you commit tooling, send us the print or the sketch. We will tell you straight whether it draws in one hit or needs redraws and annealing - and what that does to your piece price. Request a quote with your drawing, and our tool and die design team will return a draw-sequence and DFM review before you spend a cent on tooling.