LACUNA: Split Seat Tube Road Bike Design and Manufacturing

01

Design Language Research

Due to this being a personal project, I wanted to design a frame that was unique and reflected my own desired design language. I drew inspiration from a variety of sources and builds. I began this project by researching:

Throughout my research, I found that the most striking and visually grabbing bike had a multitude of small design elements working in harmony. Horizontal top tubes makes the bike seem to point forward and give a taut, disciplined posture. Slender steel tubes keep the frame looking light and nimble without collections of heavy volumes. Deep carbon rims on steel frame add visual weight and make it seem more planted and aggressive. They also add a temporal contrast of two eras of cycling meshing together, drawing on both modern performance trends and a sense of nostalgia for a previous era.

Immersing myself in these bespoke examples allowed me to build a mental library of proportions, details, materials, and visual relationships. This growing vocabulary did not prescribe a single form to copy; instead, it helped reveal the character of the bicycle I wanted to create. Like carving a form from stone, each reference provided another clue about what should be emphasized, refined, or removed, allowing the final design to emerge gradually from a broad field of possibilities.

02

Handling Characteristics & Vehicle Dynamics

Intially, I group variables into subsystems that are most relevant to each other. In reality, the geometry is a single system, and changing one variable can affect many others. However, to start to make progress and knock down some rough numbers for variables, the grouping allowed to dervie some expected rough values for certain bike parameters.

Trail is the lever arm through which lateral tire forces act to generate a self-aligning torque about the steering axis. It is the single number that most defines how a bike feels to steer: too little and the bike wanders and never settles, too much and it resists being turned at all. Industry benchmarks and rider handling studies put the useful band for a paved road bike at roughly 55–65mm (Bicycling Science, David Gordon Wilson and Theodor Schmidt, with contributions by Jim Papadopoulos), so I targeted 60mm — the middle of that band.

Because my fork offset was fixed at 50mm by what was available, trail could only be reached through the head tube angle. That inverted the usual design order: instead of picking an angle and accepting whatever trail resulted, I set the trail I wanted and solved the trail equation backwards for the angle that produces it.

Wheelbase sets longitudinal load distribution and how sensitive the bike is to load transfer under braking and acceleration, while front- and rear-center decide how the rider's mass actually divides between the two contact patches for a given fit. Mine was not a free variable: the rear wheel and split seat tube packaging require a 380mm chainstay, which is short for a road bike and pushes weight rearward. I raised the seat tube angle to 74° — the top of the typical 72–74.5° range — to move the rider forward and compensate.

Front-center matters for a less theoretical reason: toe overlap. It is the horizontal distance from the bottom bracket to the front axle — 617mm on this frame — and it has to be long enough that the toe does not strike the front tire while turning and pedaling at the same time. With 170mm cranks chosen from a leg kinematic fit calculator, the current geometry leaves ~150mm between the center of the pedal and the tire.

Static geometry only tells you what the bike is; it does not tell you how it behaves at speed. To check that, I ran the geometry through the Patterson Control Model, looking at four things: the speed at which the control spring changes sign, how firmly the bars push back against a steering displacement, how directly a steering input translates into path curvature, and how much hand force the fork flop demands as the bike rolls.

Negative steering-spring and grip-force values indicate restoring feedback under the PCM sign convention. The spring crossing is not a complete weave/capsize stability prediction.

Read together, the four curves describe the same behavior from different directions. Yaw and roll authority both rise linearly with speed, and roll tracks a fixed fraction of yaw — the coupling that makes countersteering necessary. The control spring starts positive and crosses zero somewhere around walking pace, and past that point the model has the bars resisting a steering displacement rather than falling into it. The restoring torque that produces is small at the hand: quoted per radian it reads as a large number, but a 1° input at speed works out to a few newtons, well under a pound of force. Control sensitivity climbs with speed toward the band usually associated with direct road-bike steering rather than relaxed touring geometry. What these curves are good for is the shape and the sign of each trend — whether the geometry behaves the way a road bike should, and in roughly the right speed range. The magnitudes are a different matter, and the note below covers why.

03

Frame Design

Due to the nature of the design, with custom tube profiles and a split seat tube, the frame had to be modeled for all the nuanced detail in soldidworks. I initially began by bringing over the BikeCAD geometry as a wireframe outline of the frame. From there I began modeling the tubes in their actual diameter and wall thickness. To achieve the tight fitment required to manufacture accuractely, I traced the streamlined metal tubing on a piece of paper then scanned the traced outline on a printer. After this, I imported the scanned profile into solidworks and scaled the scan based on caliper measurments. I then had the streamline tubing profile to model in solidworks.

The seat tube yoke (the piece connecting the two streamline tubes from the bottom bracket to the full size seat tube) requires adapting 2 streamline tubes to a full size seat tube. This is not easily achievable with basic solid modeling. Therefore this part requires surface modeling to attain the proper geometry. Due to the nature of the complex curves and organic geometry, I realistically had 2 manufacturing options: Metal Casting or Metal 3D printing. Metal 3D printing provided the cheapest and fastest possiblity to create this part. However, because the yoke needed to be welded to 4130 steel, I needed a similar weldable steel to properly join these materials. I went with 316l stainless steel for its ability to properly mate to the 4130 and its ease of manufacturability. A few images of that process are below.

The chainstay yoke is the most loaded and most geometrically constrained part of the frame, it has to clear the tire and the drivetrain while tying both chainstays back into the bottom bracket shell. Due to the same lofted geometry as the seat tube yoke, the chainstay yoke was also surface modeled. Notable features on the designed yoke include lines pointing to the center of the yoke and room/draft angle around the bottom bracket joint. These features greatly aid in the manufacturability of the yoke by helping with locating in the fixture and allowing the cup on the tig welding torch to easily pass around the joint.

04

Analysis & FEA

05

Manufacturing

06

Results