|Click HERE to go back to the Andy Douglas Streamliner page.
Designing and building a streamliner
The Webmaster, Andy Douglas, is in the process of designing and building his first bike. This page will track the progress of the project.
I believe in using the right tool for the job, and in this case the "mission" is to build a foamshell streamliner that I can carry inside a car. Rather than try to fair an existing bike, I elected to build a bike with the specific intention of putting it inside a shell.
I've never done anything quite like this before, so I'm leaning heavily on the input of those who have traveled this road before, in particular John Tetz, whose advice and guidance is proving invaluable. I'm also ripping off... er, borrowing... a lot of his design ideas, so I figured I'd better give credit where credit is due!
One thing about the design process that I find really striking (and fascinating) is how interdependent everything is. Make one decision and it seems like sixteen other decisions are affected (or made automatically) as a result. I call it "cascading design" for lack of a better term
Begin at the beginning
The first question to answer is: What is the mission of the bike? This determines a great deal, including layout and physical size. In this case, the bike has to be a road-practical streamliner that can be transported inside a small car without a lot of disassembly... preferably zero disassembly. The bike also has to accommodate a relatively short rider with a short legs (5'7", 30" inseam). Rough roads are not a concern... this is for pavement only.
Here's where the cascade starts.
- -These requirements meant that the bike had to be low, to fit in the car, and short enough to fit in the car.
- -The need to see over the top of the shell demanded a relatively upright seat angle in order to raise my eyes up high enough.
- -The need to fit inside a shell demanded above-seat steering with as little tiller as possible, so the bars would not hit the shell.
- - That meant that the bars would have to be located up near the headset, with arms extended.
- - This bar location, combined with the low bar position needed to fit inside the shell, and the need for the legs to clear the bar, meant that the head tube angle would have to be relatively steep and the fork have minimal offset. A conventionally-raked fork would place the bars too far back, while a reversed fork would place the headset so far forward that excessive tiller would be needed to reach the bars.
So, the overall size requirements and the idea of putting the bike in a fairing drove a bunch of other design features all by themselves. As it turned out, the riding position came out almost exactly like Tetz's FWD bike... not because I was trying to intentionally copy it, but because that's the position needed to make the bike work.
Choosing a drivetrain
Originally, I wanted to do a ZOX-style FWD, 20/20, to shorten the chain and eliminate chain interference. But Tetz talked me out of it: As it turns out, putting a derailleur down between your feet makes creating a good shape for the fairing rather difficult. Plus, getting good gearing is tough with a 20" wheel. A 20/26 woudl fit inside the car, and be much easier to gear. I ruled out a high-cluster FWD like Tetz's little yellow bike, deeming it beyond my fabrication ability.
So, I settled on a 20/26 SWB, RWD, ASS bike with a fairly upright seat angle. There will be chain interference, but there are lots of bikes that seem to handle this layout pretty successfully.
After running the numbers, I decided to eliminate the front derailleur in favor of a Schlumpf Type I Mountain Drive. This is an internally geared crankset that, when the gears are engaged, provides a 2.5:1 reduction drive. When disengaged, it acts just like an ordinary crank. It provides some advantages: extremely wide gearing range (I'll be running about 16 to about 120 gear inches! This is with a 26" rear and 11-34 Megarange cassette), and the elimination of front derailleur, shifter, cable, braze-ons, derailleur post, etc. The trade-off is some loss of efficiency when engaged.
After taking some initial measurements, the layout choice was not looking too good. A problem with low SWB bikes is that the rider's legs have to be long enough to span the frame tube, plus the wheel, plus the crankset. With only 30:" of inseam to play with and a 20" wheel, one can see the problem immediately.
There are a couple of ways to address this. One is to use shorter cranks (I'm using 150 mm cranks), but this was not enough. The obvious answer is to raise both seat and crank, thereby reducing the amount of wheel that must be fit between the rider's legs. But the shell requirement starts to get in the way very quickly. So there's a tight race between keeping everything low enough to work and making everything fit.
It started to look as if I'd have to resort to using a 16" (349) front wheel. I didn't want to do this, because there are limited tire choices in that size. But it was, and still is, an option if I can't get the 20" wheel to work out.
One method that Tetz uses is to start with a human model and design the bike around it. After measuring my own body, making an articulated 2D "mini-me" model based on it and doing drawings with it, it started to look as if the 20" wheel would work after all. But just barely... with the main tube tight against my crotch and the smallest cranks I could find (150mm), there still wasn't any room for error.
By the end of February 2001, after seeking opinions online and thinking long and hard about it, I elected to go with a 349 wheel. Though there are limited tire choices, the smaller wheel gives me a lot of room to play around with. I did consider Action-Tec suspension, but I've decided to keep things simple at the outset and just use a plain fork. I picked up a used 16" fork from a P38 for this purpose.
Subsequent drawings showed that this approach will work very well, resulting in a bike that fits into the shell and has plenty of room for adjustment.
The rear end
The next challenge was to come up with a rear-end layout. Lacking metalworking skills and equipment, I felt the simplest way to go would be to use off-the-shelf rear dropouts. However, this caused its own problems.
First, I did not know if such a dropout would be strong enough to support the bike if installed into a cantilevered stay. Second, I needed to figure out a way to attach the dropout to the carbon stays in a neat, finished manner.
I went through several alternative ideas before settling on the following:
- A Henry James stainless steel MTB dropout, which comes with plugs that just happen to be exactly the right size to fit into some stock carbon tubing from Quality Composites. It's also designed to be bent a bit, allowing me to tweak the dimensions.
- A more or less conventional rear triangle (dictated by the use of a conventioal dropout), but using the seat itself as the front leg of the triangle.
- Foam core, tapered and curved chainstays. This will rotate the dropout forward a bit to keep its relationship to the cogset correct. The main tube centerline is six inches off the ground, so the stays need to come up from it another seven inches or so to meet the wheel hub. A short piece of stock tubing will be used at the chainstay end to mate with the dropout.
- Seat stays made of stock carbon tubing.
Rear brake mounting
One of the thornier problems was figuring out a way to mount V-brake studs to the chainstays. The loads on the brake studs are so high that relying on a glue bond between the studs and the carbon stays is unwise. The studs are a complicated enough shape, with enough sharp angles, that wrapping them in carbon for reinforcement would be impractical, messy, and of dubious effectiveness.
I decided to use studs from a donor MTB fork, leaving part of the fork blades in place to act as mounting plates. These will be bonded to the stays, and further wrapped in carbon.
I have a used Greenspeed boom with a chamfered BB shell (needed to mount the Schlumpf unit), but there's a problem. This tube is significantly smaller than the carbon main frame tube, so I have to figure out a way to adapt it. I also need to offset the BB to the right as much as possible to maximize chain clearance at the fork. There are a few possibilities here:
- I can create shims inside the big tube to make its inside diameter smaller. I'm thinking of longitudinal ribs made of carbon. This might cause problems with getting enough clamping pressure to hold the boom in place.
- I could use carbon and foam to create a cone-shaped adapter that will plug into the main tube. The Greenspeed boom (which I would cut to a length of 3 inches or so) would be permanently bonded to this assembly.
-Since the bike is going to be built for me only, I don't really need an adjustable boom for fit purposes. It's a lot easier to make the seat adjustable instead... so maybe I'll just do the cone-shaped adapter noted above and simply bond it in place. Much more solid that way.... The problem is that adjustability is a good thing, to leave room for design and fabrication errors.
The list of componentry to date, either in hand or ordered:
-2" ID carbon tube from Quality Composites, Sandy, Utah (VERY rare and hard to find)
- Challenge hardshell seat, bought used... heavy at 4 lbs, will probably use it as a mold to make a carbon seat.
-26" rear wheel, left over from Wishbone upgrade
- Used Lightning front fork for 16" (349) wheel
-Shimano XT Megarange rear derailleur w/11-34 9 sp. cassette
-Schlumpf Type I Mountain Drive w/150mm arms
-Used Greenspeed boom w/chamfered BB shell to accommodate the MD.