In this thread, I am going to document how one can build a bicycle using open source software, carbon fiber, and string.
In this thread, I am going to document how one can build a bicycle using open source software, carbon fiber, and string.
Before I get into the details on this project, I want to give some background and outline my goals for it.
I've been riding for nearly thirty five years. I started riding 'bents about seven years ago. There was a noticeable reduction in speed, which I assumed was the price to pay for comfort.
A couple of years ago I did a very long ride, and found that my average speed was about 8mph. This was completely unacceptable, so I started to learn how to improve my speed.
I ended up buying half a dozen bikes, and learned a lot about going faster. After a couple years, my speeds had nearly doubled. I average about 15mph on my two fastest bikes, but my diamond frame is STILL faster.
So, the goal of this project is to finally get that one last 'push', and craft a bike that's faster than my diamond frame.
I know that 'bent riders and DF riders tend to run in seperate circles, but I honestly have no allegiance to either camp. That's the reason I'm not posting in a recumbent specific forum. I believe the methodology and technology to build a bike is applicable to both, and cross-pollination of ideas pushes the state of the art ahead.
Based on my admittedly limited knowledge, the fast speed of my diamond frame is due to decades of refinement. Basically the design isn't the most aero device in the world, but that lack of aerodynamics is offset by the fact that the bike is ridiculously light and strong, and light and strong improves the efficiency of the drivetrain.
IMHO, this is why aero improvements on DF bikes make such a noticeable difference; the design is so mechanically efficient that big gains are found when you can improve the aerodynamics.
After studying the problem for a few years, I think recumbents have nearly the opposite problem. The reclined position affords a level of aerodynamics that's quite good, but the mechanical efficiency isn't as good as a DF. I hope that isn't an inflammatory statement; I love my 'bents, but the triangulation of the diamond frame bikes affords a level of mechanical efficiency that's VERY tough to beat.
This is noticeable in two aspects in particular. The first is frame flex. On a 'bent, there's a noticeable power loss as the power applied to the crank twists the frame. I believe this is one of the reasons Bacchetta uses stick frames; a straight pipe is stiffer than one with a bend in it.
The second think that sucks efficiency in a 'bent is that looooooong chain line. The chain weight isn't insignificant; in fact 3-4% of the entire bicycle weight is bike chain! Could you imagine if the driveshaft in your car weight 450lbs? That would hardly be efficient.
Honestly, I think this efficiency is really difficult to appreciate until you sit down with a CAD program and try to design something which is superior. It's very very tough to do. In a bicycle, it's amazing how all the variables are connected. For instance, one easy way to improve the aerodynamics is to simply lower the seat closer to the ground, but as soon as you do that you have to increase the wheelbase. And increasing the wheelbase increases the weight, which slows you down.
It's a heck of a juggling act - and it makes me give mad respect to all the bike designers at places like Rans, Cervelo, Bacchetta, Velokraft, etc.
That's the reason I am starting this thread. I believe that the only way to make a measurable leap past the existing designs is to break out the CAD program, tweak a few variables, and see if I can shave a few seconds off my laps times.
Have you looked into the Cruzbike http://www.cruzbike.com/ ?
They have complete bicycles or a kit that converts a Y-Frame moutainbike into a recumbent. One of the problems I had with both my recumbents ( Vision and Rans Statos) was transporting them and their cost. The Cruzbike option has recumbent advantages with DF dimensions. Would like to hear about your progress.
Last edited by wrbabb; 04-19-12 at 12:14 PM. Reason: misstated Frame Type in reply
I look forward to your thread. I hope it doesn't get too diluted with the same tired recumbent vs df debate, but I can't help myself from suggesting that the real advantage for the DF bike is that the position you're in allows you to climb better, and for average speed on a hilly course, faster climbing is the most important thing.
Most 'bents put you in a position that is similar to an upright position on a DF. An upright position on a DF will slow down a rider that has a lot of experience on a bike. This is probably part of the slowdown people experience when changing between df and 'bent.
At the moment I have 6-8 bikes in the stable and the only two bikes that give my DF bike a 'run for it's money' is a RANS V-Rex and my Cruzbike conversion.
If you do everything you can do to optimize the RANS I think it would end up looking something like a Bachetta Carbon Aero. That's a compelling package, but I think the Cruzbike has the potential to be faster for a couple reasons.
The first reason is that my Cruzbike is about as fast as the RANS, despite having very pedestrian components. Basically there's a lot of things I can improve on my Cruzbike conversion, but the RANS is already quite refined. The RANS is approaching "the point of diminishing returns."
The second reason that the Cruzbike is a good platform for this project is something I discovered quite accidentally. Basically, on a Cruzbike you can vary the seat angle by simply leaning forward in the seat. And when you lean forward in the seat on a Cruzbike, you can just EAT hills. It's really quite extraordinary, it almost feels like you're rowing a boat, but it's a VERY efficient way to motor up the awful hills we have around here. (Portland)
On a RWD recumbent you can lean forward, but the 'bump' in efficiency isn't noticeable at all. I think this is because on a Cruzbike the chainline is ultra short and the front end is triangulated, so when you 'row' up a hill on a Cruzbike it feels like nearly 100% of the effort is getting to the crank. Whereas on a RWD recumbent it feels more 'natural' to push up against the seat instead.
Must have something to do with the angle of the legs versus the pedals, plus the chainline and the cassette.
Also, if anyone's curious, the pic above is the current bike, not the new project. The new project will be new from the ground up.
Last edited by Patrick Bateman; 04-19-12 at 02:48 PM.
I thought they banned recumbent bikes from being eligible for the hour record because they were more aero.
Some random no-name cyclist set a record for the hour by switching to a recumbent, which incited the ban.
the UCI has their own rules for bikes, which extend far beyond just banning recumbents. They just reiterated their ban on saddles that aren't level, for example. If someone wants to compete with the UCI for holding records for recumbents, they can feel free. I'm sure Mexico City will rent them track time. Recumbent records hold their own problems though, the timed mile record is over 80mph. They'd probably have to hold the record attempts on a car track somewhere.
They do hold the faired recumbent record attempts at car testing facilities. I'm intrigued by the guy who did 90 km while facing backwards and navigating by mirror.
Imagine reversing a car at 90km an hour for an hour while pedaling.
'82 Nishiski commuter/utility
'83 Torpado Super Strada ... cafe commuter
'89 Miyata 1400
Soma rush Fixie
06 Haro x3 (son's bmx)
Electra cruiser (wife's bike)
Speaking of large frame and seat, here's some pics of my bikes progress.
Here's a schematic of the stringbike. The picture in the background is the bike that I used as my 'template.' It's called 'the magic bike' and was built by homebuilder Larry Lem from recumbent.com
Here's a detail of the front end. It's basically a carbon copy. I've found that MBB FWD bikes tend to be very critical about the angles on the front end, so I don't want to risk getting this part wrong.
If you look at the first pic, you'll notice I made a few changes to Lem's bike:
#1 - I changed the seat angle to 30 degrees
#2 - I shortened the wheelbase by about 10cm
#3 - My bike will be fiberglass, not steel. This gives me the luxury of integrating the seat into the bike, and fairing the tail.
Here's Lem's bike
Here's the Vendetta. My front end is basically a clone.
I was just going to ask why they don't just run the chain directly to the front wheel like the last pic shows.. Any reason for all the methods of running chains to every point of the bicycle before finally linking the pedals to the wheel?
The other designs have a fixed boom - so you have to add an idler near the headtube. At that pivot point, the chain twists as you turn.
The fastest bike in the world* is the latter design, but it doesn't have to turn very much
Besides using an unconventional drivetrain (front wheel drive), I am also going to design the bike in an unconventional way. Instead of using pen, paper, and a frame jig, I am going to design the bike in 3D and then print it on my printer.
This has a couple of advantages. First, it allows anyone in the world to build the bike. (I'll post the bike plans.) Second, I'm hoping it will minimize geometry errors. For instance, when you build using a jig, it's easy to get an angle off by a degree or two, and I've found that's a great way to make the bike unridable.
If you'd like to try this yourself, the 3D program I am using is open source, named Blender. My design begins in a 2D program called Xara. I can't say enough good things about Xara; I've been using it since the 90s and I'm always surprised by how few people use it. The Linux version is open source, and the Windows version is $99 IIRC. My operating system is open source also (Ubuntu.) To turn the 3D design on the computer into something that we can ride in the real world, I am using a program called Pepakura to unfold the 3D geometry.
Here's a screenshot of the bike in Blender. The basic process of getting the bike into Blender goes something like this:
- The first step is finding a bike with good geometry. At first I was going to use Larry Lem's Magic Bike for my geometry, but decided to switch to the Lightning P38 for my geometry instead. The reason that I switched to the P38 is that it has a shallower seat angle, which is generally considered better for climbing, and there's lots of hills around here. So I took a picture off the internet of the P38.
- The second step is to take the picture and make it into a 2D model. To do this, I loaded the picture into Xara, then traced the bike. Be sure to set your scale right; if you do not, the bike will be too small or too big! To set my scale, I measured the wheel that will go on the bike, and then I scaled everything else accordingly.
- The third step is to take the 2D design from Xara and load it into Blender. There are a few ways to do this. The most obvious way to do it is to export the 2D model and import it into Blender, in a format like Adobe Illustrator. I did *not* do this, and instead opted to import the 2D bike from Xara as a jpg into Blender. So basically you trace the photo you get off the internet in 2D Xara, then trace the jpg from Xara in Blender. The picture above shows an example of this.
To make the 3D model into something we can ride in the real world, I am using Pepakura to flatten the 3D design.
Kudos to Taras Lesko - his Gundam model really blew my mind. Check out his website, his videos are a real inspiration.
Please describe the course on which you are making the comparison (in your second post, 4/19) of recumbent bike speed vs. upright bike speed. I've found that the course terrain matters a lot. If the course does not include significant hillclimbs, I've found that I'm faster on the recumbent bikes. If the course is 20 miles long and include a 2 mile-long climb at 6% grade, I will definitely be faster on an upright bike since so much time will be spent climbing. If the course is flat but includes lots of stop signs, I will be faster on the upright bike since I can accelerate so much faster. Some bikes that I own are ungainly (poor turning ability) with poor visibility (looking between my knees), and these bikes, though extremely fast, are only good for certain courses (few cars, no sharp turns, no stopping...). For fastest overall speeds, I would like to have a stable of bikes and be able to switch between them as the course terrain changes.