In short: three things decide it – the dimensions, the wall thickness and the complexity of the surfaces. If your part fits as a single component into 1.5 × 4 m, the wall thickness is between 0.3 and 25 mm and the geometry works without integrally moulded ribs, most of the question is answered.
largest single component – the limit comes from the autoclave
range of wall thicknesses produced, almost independent of material
from here on, instead of prepreg: infusion, wet lay-up or RTM
typically achievable tolerance depending on size and requirements
This page is for design and development engineers who want to estimate whether a part makes sense as a fibre composite – before they send an enquiry. It names the limits where it fails and the cases in which we advise against it.
Feasibility involves more than manufacturing. A part must also be measurable and testable. What we can do ourselves and what we cannot is explained further down.
“We manufacture everything from very thin prepreg laminates up to, if need be, 25 millimetres of wall thickness – and almost independently of the material.”
Before we get to laminate lay-up and verification, we clarify three things. For an assessment in five minutes we need a 3D model – on a drawing it is not possible to see in that time whether the surfaces can be demoulded.
Aviation and space, industry and automation, motorsport, marine, medical technology. Not mass-produced consumer goods.
Too large won’t fit into the autoclave, too small isn’t worth the effort. The values are in the next section.
The most difficult of the three. Without particular requirements for weight, stiffness or temperature, metal is almost always cheaper.
If the requirements don’t fit, we won’t agree on price. We would rather say so early than late.
A single component can be up to 1.5 × 4 m. This limit does not come from the design but from the autoclave – that is where the laminate cures under pressure and temperature, and the chamber is not any larger.
Assemblies can be larger. Several components are joined, and the practical limit is then no longer the autoclave but the way out: the finished part has to fit through the hall door and around the corners.
The figures are guide values from completed projects. For your part, the values we verify in the design apply.
The autoclave sets the limit: 1.5 × 4 m per single component. · Source: Connova AG
Not sure whether your dimensions still fit? Send us the model – one look and you will know.
The range extends from 0.3 mm to 25 mm and is almost independent of the material. 0.3 mm corresponds to a sandwich face sheet made of four plies. 25 mm was achieved on a structural part for the rail industry that was produced in two cures – a one-off, not a series figure.
The normal range for monolithic CFRP in prepreg is 8 to 10 mm. From around 15 mm you leave the prepreg process: pressure is then needed to get the resin through the thickness, and with it vacuum infusion, wet lay-up or RTM.
From about 10 to 12 mm, a second effect comes into play that many underestimate: the exotherm. During curing, the resin releases chemical energy. The thicker the laminate, the more heat builds up inside. This can only be controlled with a slow heating ramp and dwell points – not with the standard cycle.
The figures are guide values from completed projects. For your part, the values we verify in the design apply.
Complex free-form surfaces are not an obstacle. Neither are undercuts – they are solved with split moulds. A complete racing car front with recesses, raised areas and light and cooling ducts leads to a mould split into several parts, but it is feasible.
Conversely, an aircraft nose is comparatively simple despite its strong curvature: aerodynamically smooth, without undercuts, easy to demould. In manufacturing terms, complexity does not mean “strongly curved” but “difficult to demould”.
At least not in the prepreg autoclave process. This is the most common misunderstanding in initial enquiries. Anyone who has a milled metal structure in mind expects to be able to reproduce the same geometry in CFRP. In CFRP, ribs are made as separate parts and joined – differential construction.
Differential construction: the CFRP structure carries the load, metal fittings handle the load introduction. · Source: Connova AG
Depending on part size, complexity and requirements, the achievable tolerances are typically between 0.1 and 0.5 mm.
The way there leads through the mould. A mould made to within 0.1 to 0.2 mm delivers a part that is about 0.2 mm above that. Metal moulds are usually more accurate because they are milled directly. Composite moulds are themselves made from a mould and lose some accuracy in the process.
Then there is heat. In the autoclave, the mould and the part expand differently. This has to be anticipated in the mould design – that is where the real work lies, not in the laminating.
For surfaces, a simple rule applies: the better the mould surface, the better the part surface. That is why it pays to invest the effort in the mould as soon as it is used more than once. We do not measure roughness on the CFRP surface itself – realistically it is more like Ra 3.2, because porosity and pinholes play a role. This is only relevant where the CFRP remains visible or is painted anyway: on about a third of parts.
A general tolerance statement remains a simplification: dimensions, wall thickness, dimensional stability, position tolerance, flatness and perpendicularity are different things and are achieved in different ways. The figures are guide values from completed projects.
Gripper beam for packaging technology. Made in a closed mould; the accuracy is set via the gap. Around 0.1 mm is achieved.
Telescope structure with 4.8 mm wall thickness. Laminated in the press process, wall thickness tolerance approx. ± 0.1 mm. The contour is above that – the wall thickness is the value that is set here.
Wound products follow different rules. On the winding mandrel, i.e. on the inside, accuracy is in the hundredths range. On the outside, the raw tube is considerably less accurate. Only external cylindrical grinding brings the outer diameter into the hundredths range too – an additional process step that is reflected in the price.
Which tolerance your part really needs becomes clear at the interface. Send us the mating dimensions as well.
The cause is usually the same: parts are designed out of habit for isotropic materials. Metal behaves the same in all directions, a laminate does not.
Composite specialists only get involved when the design is already finished. By then the expensive decisions have been made.
Too localised. Every hole cuts through fibres and weakens the part locally. Where a hole is needed, edge distance, diameter, placement and locally increased wall thickness matter.
A fibre cannot be forced sharply around an edge; there it no longer transmits any force. In practice, minimum radii are the number one topic.
A complete data package: 3D data, drawings, requirements and – where available – limit samples and test specifications.
The interfaces are decisive. How is the part supported and connected? Where are loads introduced and where do they flow out? A structural part must transmit forces without being damaged – this can only be assessed once it is clear how it sits in the higher-level system.
If something is missing, we ask. Not every question can be answered straight away, because some things have not yet been decided. Then we say so: as long as information is missing, there is no binding feasibility statement. The answer is reached iteratively.
Four cases are technically feasible and still not a good idea.
Milling CFRP from solid. More than 90 % of the fibres are cut in the process. What remains holds like a pure plastic part. Possible as a design object, not as a structural part.
Very small parts such as eyelets, hooks or small inserts. They have to transmit a lot of force over a small area and offer hardly any lightweight potential. Aluminium or titanium is the better choice there. There is a good reason why screws and threaded inserts are not made from CFRP.
Small fittings in CFRP. The combination is more common and more sensible: a large CFRP structure with metal end fittings and brackets made of aluminium, titanium or stainless steel.
Extreme material accumulations. In steel, local thicknesses of 30 to 50 mm are not unusual. In CFRP this is atypical. Around 90 % of composite structures are thin-walled relative to their overall dimensions – that is precisely where their advantage lies.
Then there is the quantity. Connova AG is set up for demanding parts in manageable quantities, not for highly automated mass production. This does not apply to wound products: Connova Deutschland produces rings and sleeves in quantities of up to around 50,000 per year.
And the most common case of all: the limitation is often not technical feasibility but cost-effectiveness. What it costs and when it pays off is covered on the pages about costs and about where CFRP reaches its limits.
Initial assessment of dimensions, wall thicknesses and demouldability – before you invest effort.
Including compensating for thermal distortion in the mould – the point at which the final tolerance is decided.
Parts and moulds are measured three-dimensionally in-house, even with complex geometry.
We do not carry out ultrasonic, CT or X-ray testing ourselves. We organise them through vetted partners.
You don’t have to decide anything to get an assessment. The free consultation costs you one conversation and gives you a statement you can work with internally – even if it says that CFRP is not the right choice for your part.
3D model, industry, dimensions and what the part has to do. Drawings and requirements, where available.
Dimensions, wall thicknesses, demouldability and load introduction. Anything we are missing, we ask about in the conversation.
An assessment of feasibility with the open points named – and if it doesn’t fit, a justification instead of a quotation.
or directly to Sebastian Preiss, Head of Engineering
A single component up to 1.5 × 4 m – this limit comes from the autoclave chamber. Assemblies can be larger; the limit is then transport out of the hall.
Not in the prepreg autoclave process. Ribs are made as separate parts and then joined – differential construction.
Typically 0.1 to 0.5 mm, depending on size, complexity and mould. With CNC rework, a few hundredths are possible.
Not in-house. We work with a partner network for this.
Technically yes, sensibly no. More than 90 % of the fibres are cut, and the part loses its reinforcement.
Industry, dimensions, requirements – and a 3D model.
No. Different accuracies and different quantities apply there.
Prepreg — Fibre fabric pre-impregnated with resin that cures under pressure and temperature.
Autoclave — Pressure vessel in which the laminate cures under overpressure and temperature.
Differential construction — Ribs and stiffeners are made as separate parts and joined.
Exotherm — Heat that the resin itself releases during curing.
Undercut — Geometry that prevents demoulding in one direction.
Anisotropic — Direction-dependent. A laminate behaves differently along the fibre than across it.
Ra — Arithmetic mean roughness, a measure of surface roughness.
Non-destructive testing — Test methods without damage, such as ultrasound, X-ray or CT.
Industry, dimensions, requirements. You will receive an initial feasibility assessment – without it having to turn into an enquiry.
Heads engineering and assesses whether and how a part can be made from composites.
What does a CFRP part cost?
How long until the first part?
From what quantity is series production worthwhile?
