Vacuum Forming Mold Design: Materials, Draft, Vents and Tool Life

What a vacuum forming mold actually has to do

A vacuum forming mold does four jobs, and most of the problems in production trace back to one of them being under-designed.

• It defines the shape. That is the obvious one, and it is the only one most buyers think about.
• It lets the air out. The sheet has to be pulled onto the surface. Air trapped between the sheet and the mold has to escape through holes, or the sheet will not touch down.
• It removes heat. In thin-gauge forming the mold is a heat exchanger. A sheet at 160 °C has to be brought to a temperature where it holds shape, and the mold is the only thing available to do it.
• It survives. Not for one part — for the production run, at temperature, with a sheet being dragged across it thousands of times.

Those four requirements pull against each other. A material that conducts heat well (aluminium) is expensive and heavy. A material that is cheap and easy to shape (wood, plaster) insulates, so it runs hot and cycles slowly. A mold with generous draft and large radii forms reliably but may not give you the part geometry you drew.

The five mold materials and when each is correct

There is no single correct mold material. There is a correct material for a quantity, a geometry and a tolerance. The table below is the one we use when quoting.

Cost is relative to a machined aluminium mold of the same size. Lead times assume a mold under 1 m² with no complex cooling.

Wood

Wood is the cheapest way to find out whether a shape forms at all. MDF is dimensionally stable enough for a simple male mold, machines easily, and takes a sanded and sealed finish that will release from a sheet. The problems are that it insulates (so cycle time is long), it moves with humidity, it wears at the edges where the sheet drags, and the surface grain will telegraph through a thin sheet.

Wood is correct for a packaging prototype, a trade-show sample, or a first article to check fit. It is wrong for anything you intend to run for a month.

Plaster and gypsum

Used almost exclusively for very large, low-quantity, visually-driven parts — point-of-sale displays, theatrical props, architectural features. Plaster can be cast over a pattern and hand-finished to a shape that would be expensive to machine, and it holds fine detail. It is heavy, brittle, and it dusts. It also cannot take vacuum holes reliably, so plaster molds usually rely on a sintered or drilled insert at the deepest point.

Epoxy and glass-filled resin

The workhorse for medium runs and for textured surfaces. A pattern is machined or printed, a silicone or fibreglass jacket is taken, and the production mold is cast in a filled epoxy. The result is a mold that is light, dimensionally stable, takes a moulded-in texture from the pattern, and can be repaired by casting a patch.

Epoxy molds are also the standard route to a multi-cavity tool at low cost, because the second and third cavities are cast from the same pattern rather than machined separately.

The limitation is thermal. Filled epoxy conducts heat at roughly a fifth of the rate of aluminium, so the mold surface climbs toward sheet temperature over a run. A mold that cycles in 40 s when cold may cycle in 55 s after two hours. That is why epoxy tools for thick-gauge work are often fitted with internal cooling channels cast in place, or with an aluminium face plate.

Moules imprimés en 3D

For iteration speed this is now the default in our shop. A filled polymer print — glass-filled or carbon-filled — gives a mold in a day, with vacuum holes printed in the correct locations rather than drilled afterwards, and with conformal cooling channels if the printer allows it.

The limits are temperature and wear. Most filled filaments soften well below the sheet temperature, so a printed mold works for short runs and for small parts where the sheet does not have much thermal mass. A printed mold also has a layered surface that will telegraph into a glossy sheet unless it is sanded and sealed.

Aluminium

The production answer. Cast aluminium for large, gently-curved molds; machined 6061 or 7075 for smaller, tighter, textured molds. Aluminium conducts heat roughly 200 times better than wood, holds tolerance, takes a bead-blast or a chemical texture, and can be repaired by welding.

The cost is in the block and the machining time. A machined aluminium mold for a part the size of a car door card can cost more than ten epoxy molds, and it only makes sense when the run justifies it.

Draft angles: the numbers that work

Draft is the taper on every vertical wall that lets the formed part release. Without it the part shrinks onto the mold as it cools and locks. With too little of it, the part releases sometimes — which is worse than never, because it passes inspection and then fails in the field.

The amount of draft you need depends on the surface, the depth and the material.

Draft on a female cavity can be lower than on a male because the sheet shrinks away from the wall rather than onto it. The exception is a deep, narrow female cavity, where the sheet has to be pushed in and the walls must guide it.

Two things about draft that are widely misunderstood:

• Draft is measured from the direction of draw, not from vertical. On a mold with a curved parting line the draft angle on each wall is measured relative to the mold's opening direction. A wall that looks like it has 3° in the CAD model can have 0.5° relative to draw if the parting line is tilted.
• Draft costs height, not width. Adding 2° of draft to a 200 mm deep wall moves the base by 7 mm. That is the number to check against your assembly — not the angle.

Radii, fillets and the corners that tear

In vacuum forming, a sharp internal corner is a stress concentrator et a thinning point. The sheet has to stretch around the corner, and it stretches from the material immediately adjacent — so the corner ends up thinner than the wall, and the sharper the corner the worse it gets.

Corner thickness measured on a 2 mm HIPS sheet at 120 mm draw depth. The relationship is not linear — most of the loss happens in the first 6 %.

Practical rules:

• Minimum inside radius = sheet thickness. This is the floor. Below one thickness the corner will thin and, in a pressure-formed or plug-assisted process, will often pinhole.
• Preferred inside radius = 1.5 × sheet thickness, or 6 % of draw depth, whichever is larger.
• Outside corners should be at least 2 × sheet thickness. A sharp outside corner on a female mold produces a knife edge where the sheet wraps, and that edge is the first thing to crack in service.
• Fillets on ribs and bosses should be at least 0.5 × the rib width. A rib 8 mm wide needs a 4 mm root fillet, otherwise the sheet bridges the root instead of filling it.

Where a design genuinely needs a sharp corner — a machined-looking edge on a display bezel, for example — the usual answer is to form with a generous radius and then machine or trim to the sharp edge. That is a secondary operation with a real cost, and it should be in the quote from the start rather than discovered at first article.

Vacuum holes and vent slots

Vacuum holes are where most first-article failures originate. They are also the cheapest thing on the mold to get right, and the most expensive to get wrong, because a badly placed hole leaves a visible dimple in the part forever.

Size

The hole must pass air fast enough to pull the sheet down before the sheet cools, and be small enough that the sheet does not dimple into it. Both constraints depend on sheet thickness and on how soft the sheet is at the moment it touches down.

Rule of thumb: hole diameter should not exceed about one third of the sheet thickness at the point where the sheet lands. Above that, the sheet pulls into the hole and leaves a dimple that is visible on the outside of the part.

Position and spacing

The instinct is to distribute holes evenly. That is wrong. Holes should be concentrated where the sheet lands last — the deepest point of a female cavity, the top of a male mold, the inside of a tight corner, the far end of a long draw. Those are the places where air is trapped longest, and where the sheet has the least time left to conform.

• Deepest points first. Every local minimum in a female mold and every local maximum on a male mold needs a hole at the centre.
• Corners second. Each inside corner of a female cavity gets a hole within 10 mm of the corner, or a slot along the corner.
• Then a grid. Elsewhere, a grid at 25 – 50 mm pitch for thin gauge, 50 – 100 mm for heavy gauge, is enough to evacuate the volume.
• Never at a visible cosmetic surface. If the outside of the part is the show face and it sits against the mold, a dimple will show. Move the hole to a non-cosmetic area or use a slot in a trim zone.

Manifold and evacuation rate

Holes are only half the system. Behind them there has to be a path to the vacuum source with enough cross-section that the pressure at the farthest hole is close to the pressure at the pump. A common failure is a beautifully drilled mold with a 6 mm internal channel feeding 200 holes — the near holes see full vacuum, the far holes see almost none, and the part forms on one side only.

Design the internal manifold at 1.5–2× the total hole area, and give each large mold section its own connection to the vacuum tank rather than daisy-chaining them.

Thermal design: hot molds, cold molds and heat soak

Thin-gauge and thick-gauge forming have opposite thermal requirements, and a mold designed for one will perform badly on the other.

Thin gauge: the mold is a heat exchanger

In thin-gauge (roll-fed, in-line) forming, cycle time is typically 3 – 10 s and the sheet is 0.3 – 1.5 mm. The mold must remove heat fast enough that the part is rigid at ejection. Aluminium with drilled or cast cooling channels is the norm, and mold temperature is actively controlled — usually 20 – 40 °C for PS, PVC and PET, higher for PP and HDPE.

Cooling channel design matters more than mold material here. Channels should follow the part surface at a consistent depth (typically 25 – 40 mm below the surface) rather than being drilled in straight lines that wander close and far. A channel that runs 15 mm below the surface in one area and 60 mm in another produces a part that is rigid in one place and warped in another.

Thick gauge: the mold is a heat store

In heavy-gauge forming, the sheet is 2 – 12 mm and the cycle is measured in minutes. The mold is usually run hot — 60 – 90 °C, sometimes higher — because a cold mold chills the sheet surface instantly and produces chill marks, internal stress and a part that warps as it relaxes.

Hot molds on heavy gauge are usually achieved by circulating water or oil through the mold, or by building the mold from a material that holds heat. The consequence is that the mold is a heat store: it takes 30 – 90 minutes to reach equilibrium, and every part formed during that warm-up is different from the part formed after it. First-article approval has to be done on a warm mold, not a cold one.

The two processes share a name and almost nothing else. A mold built for one will not run the other.

Shrinkage compensation

A formed part is smaller than the mold it came off, because the sheet shrinks as it cools. The mold has to be made larger by the shrinkage allowance, and the allowance depends on the material, the draw ratio and the amount of orientation left in the sheet.

Values are for free shrinkage after forming. Constrained features — a part held flat by a frame or a jig while it cools — shrink less in the constrained direction.

Two practical points:

• Shrinkage is not uniform. A drawn wall shrinks differently in the draw direction than across it, because the sheet was stretched in the draw direction and that stretch is partly recovered on cooling. On a deep part the length can shrink more than the width. Compensating with a single percentage is an approximation, and it is usually good enough — but on a deep PP part it is not, and the mold should be cut with the draw direction allowance applied separately.
• Form the first article, measure it, then cut the mold. For anything where the tolerance matters, the sequence is: build the mold slightly oversize, form 10 parts, let them condition for 24 hours, measure, and then machine the mold to final. It costs one extra setup and saves the argument at first article.

Texture, engraving and surface finish

Texture on a vacuum formed part comes from the mold, but it does not reproduce faithfully. The sheet stretches as it forms, so any texture on the mold is stretched with it — more in the draw direction, less elsewhere. A texture that is uniform on the mold becomes a gradient on the part.

Heavy textures are usually specified on the non-cosmetic face or on parts where the draw is shallow, precisely because deep draw flattens the grain.

Where a heavy texture is essential on a deep part, the usual solution is to texture only the shallow areas and leave the deep draws smooth, or to accept a "texture fade" and specify it in the drawing so that the buyer and the molder agree on what the part will look like.

Engraving — logos, part numbers, date stamps — is cut into the mold as a raised feature, so it appears as a recess in the part. Minimum legible feature on a formed part is roughly 0.3 mm depth and 0.4 mm line width, and the engraving must be on a surface that does not stretch much. A logo placed on a deep draw will be distorted and possibly illegible.

Undercuts and the parts you cannot form

Vacuum forming has one draw direction and no side action. Anything that would prevent the part coming off the mold in that direction is an undercut, and an undercut needs either a design change or a secondary operation.

The routes available:

• Redesign to eliminate it. The cheapest answer and the one to try first. A snap-fit lip that would be an undercut can often become a separate clipped-on part.
• Flexible mold inserts. A silicone or elastomer insert that deforms to release the undercut. Works for small undercuts on short runs; the insert wears and has to be replaced.
• Collapsible or split molds. Two or more mold sections that separate before the part is stripped. Doubles or triples the mold cost and adds a seam line on the part.
• Form it oversize, then cut it. Form a plain wall and machine the undercut afterwards. Accurate, but it is a machining operation and it costs what machining costs.
• Form in two halves and join them. Common on large enclosures and on twin-sheet parts.

Mold life and the cost per part of a mold

Mold life is not a fixed number — it is a function of material, sheet, temperature and care. The table below is what we plan to, and it is deliberately conservative.

Heavy gauge life is lower because the sheet is thicker, hotter and stiffer, so it drags harder across the mold and carries more heat.

Converting that into a cost per part is the calculation that decides the mold material:

A mold design review checklist

The list we run before a mold is released to machining. Most first-article problems are on it.

Items 1 – 6 account for the majority of first-article rejections. Items 11 – 14 account for most of the production complaints that are not about the part itself.

Frequently asked questions