Designing for thermoforming: draft angles, draw ratio and wall thickness

What does designing for thermoforming actually mean?

It means designing a shape a heated sheet can be persuaded to become, without stretching so far that it goes thin, and without wrapping around anything it cannot then be pulled off. Everything below follows from those two constraints.

The sheet contacts the tool progressively. Wherever it touches first it chills and stops stretching; everything not yet touched keeps stretching to make up the difference. Material never flows to where you want it. It ends up wherever the geometry sends it, and the last place reached is always the thinnest.

Our envelope at Noida is 600 × 600 mm forming area, 75 mm maximum draw, on 200 to 2000 micron sheet in A-PET, rPET, PVC and HIPS. Those numbers bound most of what follows.

How much draft angle does a thermoformed part need?

On a female cavity, one to two degrees per side is the practical minimum and three degrees is comfortable. On a male tool, start at three degrees and prefer five to seven, because the part shrinks onto the plug as it cools and grips it hard.

Draft is the taper that lets the part release. Zero draft in a female cavity drags, scuffs and sometimes tears on ejection; on a male tool it locks on so tightly the part distorts coming off. In a female cavity the plastic shrinks away from the walls as it cools, so it is already trying to release. On a male tool it shrinks onto the plug and clamps. Hence the asymmetry.

Draft is one of the few design changes that costs nothing and helps everything: release, wall distribution, cycle time and the number of proving iterations. If you are unsure, give it more.

What is draw ratio, and why does depth cost you wall thickness?

Draw ratio measures how far the sheet has to stretch. Because thermoforming adds no material, every millimetre of extra depth is paid for out of wall thickness somewhere else — usually the bottom corners, which are the last place the sheet reaches.

Two ratios are worth knowing. Linear draw ratio (LDR) is the draw depth divided by the smallest opening dimension of the feature — a pocket 40 mm deep and 40 mm wide has an LDR of 1:1. It is a quick sanity check, and it tells you where the sheet has to constrict at the rim.

Areal draw ratio (ADR) is the formed surface area divided by the flat sheet area that feeds it. This one predicts thickness directly:

• Average finished thickness ≈ starting gauge ÷ areal draw ratio
• Thinnest point (usually the bottom corner) ≈ half the average, as a working assumption

So a 500 micron sheet drawn at an areal ratio of 2 gives roughly 250 micron average and might be near 125 micron in the worst corner. If your pack needs 200 micron in that corner to survive a drop test, you need heavier sheet, a shallower pocket, or a bigger radius. There is no fourth option.

Sheets tolerate stretch differently. PVC is forgiving and holds fine detail at depth. A-PET and rPET are stiffer, less tolerant of aggressive draws, and show thinning more visibly because they are clear. HIPS draws well and hides thinning because it is opaque, which is why it is the workhorse for deep dunnage trays. Full comparison in A-PET vs rPET vs PVC vs HIPS.

What radii should you use?

Never specify an internal radius smaller than the sheet gauge. Aim for at least three times the gauge at the junction between the base and the wall, and be more generous still at the bottom of a deep pocket.

Sharp internal corners are the enemy for three reasons. The sheet stretches most tightly there, so it thins most. The corner traps air, so it vents worst. And a small radius is slower to machine into aluminium, so it costs more on the tool too.

A square pocket with sharp vertical corners is the classic mistake. It looks tidy on a drawing and it webs in production, because the sheet has nowhere to go as it constricts. Round the plan view and the same pocket forms cleanly.

Can a thermoformed part have undercuts?

Shallow ones, yes — a formed sheet flexes enough to be stripped off a small return lip or retention bead. Deep undercuts need a split tool or a moving insert, which adds tooling cost and slows the cycle.

Whether an undercut strips depends on its depth relative to the wall, the material's flexibility at ejection temperature, and where it sits. A modest lip on flexible PVC strips. A pronounced re-entrant feature on stiff A-PET does not. Before committing to one, work through this order:

1. Can a retention rib do the job? A small formed bump inside the cavity often grips a product perfectly well and is not an undercut at all.
2. Can a stepped wall do it? A step with draft holds a product against vertical movement without any re-entrant geometry.
3. Can the pocket simply be deeper? Sometimes gravity and a snug fit are all the retention you needed.
4. Can the feature move to the flange or the lid? A tab on the flange is easier to form than an undercut in the wall.
5. Only then, is it worth a split tool? Sometimes it is — a clamshell latch genuinely needs one.

Most undercuts we are asked for turn out to be step one or two in disguise. They exist because the designer was thinking in injection-moulding terms.

How do ribs and stacking features work in a formed part?

A formed rib is a shallow channel in the surface, not a solid injection-moulding rib. It stiffens the part by changing its section shape, and it costs nothing except a little wall thickness where it forms.

A flat panel in thin sheet is floppy. Two or three shallow embossed channels across it make it noticeably rigid, because bending now means deforming a corrugated section rather than a flat one.

• Wide and shallow beats narrow and deep. A rib three to four times as wide as it is deep forms cleanly and vents properly.
• Draft the rib sides at 3° or more, and radius both the top and bottom edges.
• Run ribs across the direction you want stiffness. A tray that sags along its length wants ribs across the width.
• Do not cross ribs at a sharp junction — blend the intersection or the sheet thins badly there.
• Use the perimeter. A returned lip or downturned rim around the whole tray adds more stiffness than any number of internal ribs.

Stacking features work the same way. A shallow step in the base that locates into the rim of the tray below turns a loose pile into a stable column — useful in dunnage and component trays that travel on a pallet.

How do you make trays denest reliably?

With denesting lugs: small formed buttons or steps, usually three or four around the rim, holding each tray a fixed distance off the one below so a machine or operator lifts exactly one.

Nesting saves freight and storage, but a tray that nests too well is a production problem: two stick together, the line loads a double, and someone downstream finds out later. What makes denesting work:

• Three or four lugs, not two. Two let the tray rock and re-nest.
• Place them asymmetrically on a symmetrical tray, so trays cannot rotate into a locked position.
• Adequate draft on the walls. Under-drafted walls create a vacuum grip no lug will overcome.
• A vent path. A tray that seals against the one below sticks by suction; a small notch in the rim breaks it.
• State the nest pitch you need. It sets your carton count and freight cost, and we size the lug height to hit it.

Typical achievable nest pitch as a proportion of tray height for our tooling: CONFIRM — a ratio range drawn from recent dunnage and insert-tray jobs.

What tolerances can a thermoformed part hold?

Dimensions formed against the tool are the accurate ones. Dimensions on the free face, or across an open span, carry the accumulated variation of sheet gauge, heating and shrinkage. As a default, general engineering tolerance class DIN ISO 2768-c is the sensible starting point.

Thermoforming is not a precision process in the injection-moulding sense, and drawing it as if it were leads to arguments at inspection. The hierarchy:

The most useful thing a designer can do is mark the critical dimensions and leave the rest alone. Usually only two or three matter: the ones that mate with a sealing head, a carton or an automated feeder. Tightening every dimension does not make the part better. It makes it more expensive and harder to accept.

Male or female tooling — which face gets the detail?

The face that touches the tool. A female cavity mould puts accuracy and finish on the outside of the part; a male plug mould puts them on the inside. You cannot have both, so decide which face has to be right. It is the most consequential decision in thermoform tool design, and it is usually made by default rather than deliberately.

A worked example. A medical instrument tray must hold the instrument to a defined internal position; the outside only has to fit a carton with clearance. That is a male-tool part. A face blister seals to a printed card and hangs on a hook, so the outside is what everyone sees. That is a female-tool part. Getting this backwards is recoverable only by cutting another tool.

How should the flange be designed for sealing?

Flat, wide enough for the sealing head plus tolerance, and consistent all the way around. If the pack seals to a card or a lidding film, the flange is the most important feature on the drawing.

We trace more line rejections to flange geometry than to any other single cause, and the failures are usually intermittent, which makes them expensive to diagnose. Rules that prevent it:

• Give the flange a flat land, not a curve. The whole sealing width must be in one plane. A flange that dishes will seal in some places and not others.
• Make it wider than the sealing tool, with tolerance for placement variation. Get the head dimensions from your sealing machine supplier and send them to us before tooling.
• Keep the width uniform. A flange 5 mm on one side and 3 mm on another seals unevenly however good the machine is.
• Radius the transition into the pocket, and do not let the pocket wall break into the flange land.
• Consider a moat or return lip outboard of the seal area. It stiffens the flange, keeps it flat and gives the operator something to grip.
• Decide the trim line early. The flange is formed by the mould but defined by the punch, so both tools must agree.

Sealing flange land widths we work to for face-seal and lidded formats: CONFIRM — the standard width bands used on our blister and tray tooling.

If the format itself is not settled, the trade-offs are covered in types of blister packaging and clamshell vs blister.

A design checklist before the tool is cut

Run these ten checks on any thermoforming drawing and you will avoid most of what shows up at sampling. None cost anything now. All cost money after the aluminium is cut.

1. Draft on every vertical wall: 2° minimum female, 3° minimum male, more if textured.
2. Linear draw ratio under 1:1 on every pocket, ideally under 0.5:1.
3. Areal draw ratio calculated and the minimum wall checked against what the pack must survive.
4. Every internal radius at least one sheet gauge; base-to-wall radius at least 3× gauge.
5. No square plan-view corner on a pocket deeper than about 15 mm.
6. Undercuts eliminated, or justified as worth a split tool.
7. Male or female tooling chosen deliberately, critical face identified.
8. Flange land flat, uniform, dimensioned against the actual sealing head.
9. Denesting and stacking specified, including target nest pitch.
10. Two or three critical dimensions marked; the rest at general tolerance.

Frequently asked questions