Custom aero carbon DIY

Who are you (no off-topic talk please)

Moderator: Moderator Team

Post Reply
Posts: 270
Joined: Sun Dec 19, 2010 4:24 am
Location: Ann Arbor, Michigan

by Stefano

Hi WW,

I've been lurking here for forever, decided it was finally time to share my project with you all! I'm currently a Mechanical Engineering student at the University of Michigan, and my friend (Aerodynamicist/graduate student) and I decided that it would be cool to try to design a cutting-edge aerodynamic bike. Since we're around the same height, we designed a geometry to split the difference and hit it with all the aero tech we could think of. I designed the layup schedule and the manufacturing plan, and after about 8 months of redesigns and optimizations, we are about three weeks away from riding completed bikes!

According to our CFD, we are within 2 watts of a Cervelo S5. Once the frame is done, we will be going to the wind tunnel to validate those claims.

One of the more interesting aspects of our design is it's flexibility-- we can produce a monocoque frame with almost any stack/reach measurement for a total material cost of ~USD$850, plus labor and machining.

Here are some pics, enjoy! Please ask questions, I love to talk about the bike :)

initial CAD model in NX- under BB brakes, designed for 3T funda fork integration at headtube
Cutting the NDS half of the clamshell
Seatstay/chainstay moulds
Finished clamshells. Just gotta gel-coat and wax
Incomplete frame jig (gotta get parts made first!!) :lol:
Seat layup-- not pretty but works like a charm :)
Straight out of the mould
Little bit of trimming-- more to come

Posts: 313
Joined: Sun Jan 09, 2011 2:35 am

by FilmAt11

Great Stefano! Can you describe the flexibility in the design to enable different sizes?
Last edited by FilmAt11 on Wed Feb 18, 2015 7:52 pm, edited 1 time in total.

by Weenie

Posts: 74
Joined: Sun Feb 22, 2009 12:31 pm
Location: Surrey, UK

by TomHelly

Keen to see how this turns out!

I'm guessing you didn't worry about meeting UCI 3:1 etc. with this design?

Posts: 2293
Joined: Tue Apr 12, 2011 8:08 am
Location: Pedal Square

by HillRPete

Stefano wrote:According to our CFD, we are within 2 watts of a Cervelo S5.

That's cool.
Can you share more details maybe? Did you do CFD with a rider (assuming complete bike, not just frame at the very least)? What's the yaw distribution for the simulation?

Posts: 3654
Joined: Wed Jan 22, 2014 11:48 am
Location: Vienna Austria

by Marin


More/better pics would be great too.

User avatar
Posts: 112
Joined: Wed Dec 24, 2014 2:38 pm

by MartyWild

Great project, keep it coming!

Posts: 315
Joined: Tue Jan 28, 2014 12:13 am

by tmr5555

Hi, seems like a great project, very intrigued,
could you post larger photos please?

Posts: 270
Joined: Sun Dec 19, 2010 4:24 am
Location: Ann Arbor, Michigan

by Stefano

Glad to hear everybody is interested! I'll try to answer all your questions as directly as possible, let me know if you don't agree with/understand something!

@FilmAt11: In order to make different sizes, we will machine completely new clamshells for the main frame. The way we constrained the airfoil and tube shapes in CAD allows us to simply enter different stack and reach parameters, convert into a tool-path, and cut out of a proprietary (not metal or tooling board) material combination and gelcoat system. The materials we are using are about 10x cheaper than high density tooling board, and in our testing were able to withstand similar usage cycles as long as the finish was polished to class A. The chainstays and seatstays remain the same shape, and if we ever did go to full production, we would machine those moulds out of aluminum.

@TomHelly: We actually designed this frame SPECIFICALLY for UCI tube shape restrictions. This is compliant with all of the UCI's guidelines on frame shape.

@HillRPete: I can't share too many details! :lol: But here are a few. We ran from 0-15 degrees of yaw, optimized around ~7 degrees (pretty standard in the industry). Our simulations didn't have a rider per-se, we had the boundary condition equivalent of legs, butt, and arms (along with wheels and the rest of the components). We also included two size large bottles in the simulation.

As for more/better pics, I don't have a very good camera but I will see what I can do for you guys :) This weekend we will do a lot of work, I will post more updates/pics then!!

Keep the questions coming, thanks for all of your kind words :)

User avatar
Posts: 1992
Joined: Tue Sep 11, 2012 6:07 pm
Location: The Lone Star State

by FIJIGabe

I love the ingenuity. I can't wait to see the progress of the build. Now, can you provide some more technical details? What BB standard did you decide to use? What wheels did you design the frame around? Is the frame designed for both mechanical and electronic shifting?

Posts: 3920
Joined: Thu Nov 10, 2005 5:31 pm

by BmanX

I can't wait to see this build.

Cervelo SLC 5960g/13.13 lbs

Posts: 141
Joined: Wed Apr 22, 2009 4:46 pm

by MattBeOne

Hi stefan,

Super cool project. I'm a Msc student Aerospace materials and structures myself at the TU Delft. Would love to start a project like this one day as well!
Keep up the good work at keep us posted! Really interested in what techniques and software you used for the design!

User avatar
Posts: 572
Joined: Tue Oct 04, 2011 5:04 pm

by Tomstr

Awesome work!

I can't wait to see the build photo's. How are you managing the tolerances of things like bottom bracket and headset?
Ride it like you stole it

Posts: 270
Joined: Sun Dec 19, 2010 4:24 am
Location: Ann Arbor, Michigan

by Stefano

@FIJIGabe: We are using a 68mm english threaded BB for this build, although we could do BB30 too pretty easily. Threaded has wider tolerances though so it's easier for a first attempt! The wheels we used in the simulation were just boundary conditions for what lower-level CFD predicted the airflow would look like off of 50mm v-shape. If we ran spinning wheels with frame and rider, we would still be on our first simulation :lol: Frame is designed primarily for electric, but I'm putting mechanical on mine. Should be easy, turn out a few cable stops on the mill and boom :) Routing is behind the headtube.

@BmanX: Thanks!! That means a lot-- your and professor chaos' cervelo builds are an inspiration and beautiful :) it's from there that I got the idea to combine the seat with the seatpost all in one.

@MattBeOne: Cool! Glad you like it. If you ever do decide to do a project like this, PM me and I can give more advice/ pitfalls that I struggled with. Software- NX, Catia, SolidWorks, Rhino for CAD, plus all the CFD software we used. Initial airfoil shapes were narrowed down/drawn in xfoil, but I'm not supposed to talk about the rest of the design process.

@Tomstr: They're coming! I'm headed to the shop in 30 mins to do some work, should have some meaningful progress by this weekend. The way we are managing BB tolerances is a machined seat in the mold that cups the BB. It has a hole in both sides that allows a little angular play, which we will take up with a bolt-through centering mechanism of my own design. For the headtube, I am grinding down a thin aluminum tube which will set the bearing seats exactly in the right place relative to each other, again using the bolt mechanism. Eventually this could be made of CF, or omitted entirely.

Thanks again for all your questions, keep them coming! I will do my best to get better pics and will report back soon :)

Posts: 1548
Joined: Fri Oct 28, 2011 3:25 am
Location: United Kingdom of Great Britain and Northern Ireland

by Valbrona

So what's the seat angle and what's the head angle? And presumably these are fixed and do not vary if you make frames with differing stack/reach.

by Weenie

Posts: 99
Joined: Mon Oct 10, 2011 8:34 pm

by 7ducati

Awesome project. Go Blue! Class of (shudder) 2001
2017 Cervélo S3
2016 Santa Cruz Stigmata

Post Reply