The basic construction method is established engineering. We are not claiming to have invented titanium-lugged carbon frames or adhesive-bonded tube construction. What we have concentrated on is how we execute it.

The production 00Kev uses double-lap bonded joints throughout the principal carbon-to-titanium connections. Instead of terminating a carbon tube against a single bonded overlap, a double-lap joint secures it across two bond surfaces. For a similar overlap length this gives roughly twice the adhesive contact area and produces a more balanced load path through the joint. That second point is important.

Structural adhesives generally work very well when the joint is arranged so that the load is carried predominantly in shear. Performance is diminished when the joint geometry introduces large peel loads at its edge. A double-lap arrangement essentially removes peel stress, transferring all the load to more shear forces more uniformly distributed through the bond area.

The technique has a long history outside bicycles. Adhesively bonded double-lap joints have been studied and used in aircraft structures for decades, in part because they allow a strong joint to be produced without paying the weight penalty associated with a much heavier mechanical connection. 

There is also good precedent within the bicycle industry. A number of manufacturers use bonded lug-and-tube construction, but Atherton Bikes is particularly relevant to us because it has pushed 3D-printed titanium lugs, carbon tubes and double-lap joints throughout frames intended for very serious mountain-bike use. Its enduro and downhill bikes operate in a load environment well beyond normal gravel riding; World Cup downhill is often close to controlled tumbling down a rocky mountainside.

We looked closely at that precedent. Our frame design, lug geometry, tubes, manufacturing processes and test programme are our own, but there is value in recognising where an engineering method already has a substantial body of work behind it.

While bikes do not face the same loads or operating temperature ranges as aircraft, care must be taken in choosing adhesives. We can’t have bikes debonding in the back of a car on a hot Australian summer day. Our preferred adhesive for 00Kev, while tricky to work with, does not materially reduce in strength across a wide temperature range.

Double-lap joints make the frame easier to trust structurally and considerably more difficult to manufacture. A single socket is relatively straightforward to design for additive manufacturing. A double-lap connection requires narrow internal geometry so that the carbon tube can sit between two titanium bond surfaces. On the 00Kev several lugs have these sockets pointing in different directions.

A metal 3D printer does not operate outside of the laws of gravity - yet. And so gravity must be taken into account when designing printable parts. People often have a view that anything can be 3D printed, but that is not true. When 3D-printing a structure each layer needs sufficient material beneath it to support the next layer, or additional support structures have to be printed with the part. Those supports are removed after printing.

That becomes difficult when the feature requiring support is inside a narrow double-lap socket. There is little value in printing a geometrically perfect internal surface if the support structure needed to create it cannot subsequently be removed without destroying the part.

Build orientation therefore becomes part of engineering each lug. We have to consider the direction of each socket, allowable overhangs, support access, surface finish, post-processing and the critical load directions at the same time.

The down tube to head tube connection is a good example. That junction is already one of the most highly loaded areas of a conventional bicycle frame. On the 00Kev it is also one of the most complicated printed parts because the lug has to provide the required head tube and down tube geometry while maintaining the double-lap bond architecture and remaining manufacturable. There are orientations that make one surface easier to print while making another effectively impossible. And changing the lug to solve a printing problem can alter its stiffness, weight or appearance. 

This is why we don't regard additive manufacturing as a process where we finish a CAD file and send it to a printer. Printability is part of the component design from quite early in the process.

The titanium parts are 3D-printed in Grade 5 titanium. Bearing seats, threads and other critical interfaces are then post-processed as required rather than relying on the raw printed surface for every dimension.

The external shape is only a portion of the CAD work. Wall thicknesses change through the lugs according to the job the section is doing. Internal sockets have to accommodate the carbon tube, adhesive and assembly tolerances. Bearing and bottom bracket interfaces need appropriate material for final machining.

Some of the features are only fractions of a millimetre in size. We have designed features down to around 0.3 mm, including small bond-separation nodules used to help control the bonded interface.

We also print identification and revision information into parts where appropriate. Version control becomes increasingly important once several generations of a lug exist. A component can look essentially identical externally while having a different internal wall, bond feature or geometry.

Another unseen feature is the steering limiter that we’ve built into the toptube-headtube lug. This limiter is there to stop the top tube getting destroyed from an impact with the handlebar in the case of a crash. The limiter allows the normal steering while riding but it stops the rotation of the handlebar before hitting the top tube - something that’s only important in a crash scenario. It was easy to incorporate this feature once designed but it’s a nice detail that will extend a 00Kev frame’s life.