The Orion Speedtrike By Raymond Gage

Orion Speedtrike

Fairing Construction By Raymond Gage
Latest Trike Fairings Frame 2 Integration CasaGrande

Sept 29, 2004: Continued
This is the old plug from a much earlier version of this project. I won’t be using it for several reasons including that it’s 10’ 6” long (too long for the Dempsey - MacCready prize) and has warped somewhat over the years. The new fairing looks to be coming in around 9 feet. Fortunately it only took me about two weeks of after work effort to make back in 1995, so I hope that the new one will go about as fast.

I will endeavor to keep the WISIL site abreast of my work over the upcoming months as I attempt to complete this project in time for WHPSC 2005. To that end, I plan to attend the Oregon Human Powered Challenge, Memorial Day 2005 in Portland Oregon. The purpose for this date is to have a milestone to reach for besides the Battle Mountain event in Sept/Oct. That way, I still will have margin on my schedule for tweaking and training, two very necessary parts of any successful racing attempt.

Below are a couple renderings of the trike:

The next parts of this project involve several things. I am performing some design studies to decide if I wish to change the chassis configuration to lean steering with fixed front wheel. Greg Kolodziejzyk’s excellent web pages on his Trans Canada Human Powered Vehicle Record attempt have been an inspiration for these studies, especially his TCR1 lean steer trike. There are many similarities in our projects, so I can’t help but wish him well.

I am tweaking the existing chassis to improve it’s trainability and high speed capability in case I retain it for the final configuration. To that end I need to fabricate a seat, install a steering damper, improve wheel changeability, create better rear wheel axles, narrow the track, improve its visibility to other vehicles, add more midshaft gears, eliminate or reduce the various interferences with my legs, connect the front derailleur, connect the cadence sensor, fabricate and install some kind of debris shield for the front wheel, and whatever other little tweaks occur as I get more time training on the thing. 

Once I am happy with my position and pedaling action, it will be time to perform the final loft for the fairing design. Then I need to perform the fairing plug/mold/part/transparency fabrication for both the main fairing and the wheel fairings before the Portland event! Should make for a busy year!
After nearly exhaustive time spent lofting the final fairing shape for the Orion Speedtrike
in 3D CAD, I was able to come up with something that I can live with. I generated cross sections and passed the data to Alex Bethet of AVL Looms here in Chico, California. Alex did a fantastic job of converting my 'dirty' CAD file into something that AVL's CNC router would handle, and then machining them from MDF. The actual cutting of the parts went surprisingly smoothly.
Pictured are the parts as cut and then assembled into a subframe that will support the foam planks and fiberglass for the main fairing's plug. The parts went together very well with only a few minor interferences which where easily fixed. The fit is excellent. I have made a number of plugs this way before, but never with CNC, so it is nice to see how accurately everything came together.
The next step is to plank the subframe with 1 inch thick foam and then cover with fiberglass/epoxy and finish. History has shown that this step goes surprisingly quickly with an accurate subframe.

New Orion Speedtrike foam plug
Work has progressed nicely on the foam plug and is now almost complete.  But first, I recently found some old pictures from my previous speedbike projects, which show how the design and construction have evolved over the years.
Hummer Speedbike
My first speedbike was called Hummer, and was designed and built by the Chico State HPV team during 1988-1991. This vehicle was my first attempt to create a shape that had more of a 3D surface, rather than 2D, which many of the bikes of the period used. It used NACA 6 series airfoil sections stacked vertically to create an aerodynamic shape. It also featured a mold and jigless carbon fiber frame, as shown.

Inside Hummer Speedbike

Hummer Speedbike

Mold and Jigless carbon fiber frame for Hummer
Great White Speedtrike
My second design, which eventually evolved into the current Orion configuration was nicknamed Great White, and I worked on it during 1991-1994 with Rick Kirk (later of Razor Rock Racing). This design featured many design and construction concepts that have carried through to the current configuration, such as a fully 3D shaped fairing, highly laminar airfoil cross sections (still the NACA 6 series though), fully reclined rider position, three wheels, and foam plank on frame plug construction. 

Foam plank on frame construction
Great White however, was designed with only a single symmetry plane, so that the left and right sides were mirror images, but not completely symmetrical. This feature allowed for good visibility for the fully reclined rider, and possibly better airflow close to the ground, but necessitated the construction of two plugs, a left and a right. As my design experience was fairly limited at this point, the vehicle was to be 12 feet long, and with two plugs to make, they took a long time to build.

Great White plug

One trike, two plugs

Linear 95 plug and friend
A side effect of things taking too long (several years in this case) is that life situations change, and I started my professional career in the middle of this project. After some time of neglect, I returned to retrieve my plugs from storage to find that they had been lost, probably destroyed. This discovery made me a little upset, and I decide that rather than redo all the Great White plug work, I would make a newer updated version.

Linear 95 Speedtrike:
My third design was called Linear 95, as the design and construction work was done in 1995, and it's goal was was to add a semi-linear drive to the Great White configuration to make it smaller. The Linear 95 configuration also introduced the two symmetry plane feature that has carried through to Orion, allowing only half of the vehicle shape to be produced in mold and plug form, thus halving the time and cost of the mold making process. I still used the foam plank on frame plug making process, but refined its details to make it more precise.
It also used more aggressive airfoils (NACA 66 and 67 series, rather than 65 and 66) to further reduced the size.

While the cross sections of Great White were composed of several different arc segments, the cross sections of Linear 95 were true ellipses, allowing for a smoother surface. All of these changes allowed me to shorten the vehicle to 10 feet 6 inches. The Linear 95 plug is shown in grey with me in the picture above.

Orion Speedtrike:
My current design is the Orion Speedtrike, and was begun in 2001. With it, I have backed away from semilinear drive to traditional bicycle cranks, but with shorter arms to improve clearance in the nose of the vehicle. I have adapted the front wheel drive configuration that has become the de facto standard for many speedbikes, and for the first time, use my own custom airfoil sections which should allow for large amounts of natural laminar flow and drag reduction. With the custom sections, I have been able to reduce the fairing length further to just over 9 feet. Considering, that I am over 6 feet tall, and I ride fully recumbent, I consider the fit to be pretty good.

Again, I only need to make half of the plug, due to two symmetry planes, but use cubic splines for the cross sections this time. The cubic sections allow better for all of the local bulges required for a tight fairing around feet, knees, hips, elbows, shoulders, and head, while keeping the outside profile very smooth. I'm also using the foam plank on frame construction with further refinements, especially the use of CNC cut frame sections, as shown in the last update. Another important refinement with the Orion plug is a move to a reference plate table for precision alignment, instead of the wooden strong-backs used previously. Wood frames tend to warp, and I find them very difficult to align. The warped Linear 95 plug helped me decide to start from scratch again for Orion.

July 12, 2005:

New Sponsor!

Work progresses, and I am happy to announce a new sponsor, Powercranks. Powercranks are a special independent crankset, and are an important training and racing aid to improve human power, and aid in injury recovery and prevention. As I have had recurring problems with my hamstrings, probably due to a quadriceps/hamstring strength imbalance, I am very optimistic that they will improve my form, and help prevent future injuries.

Powercranks. Each crankarm spins independently.
Plug Progress: As promised in my last update, I will share more details of my plug construction. 
 Note templates to check final surface, at bottom right. The method I have now used for all of my own plugs is based upon ancient ship building techniques. I call it “foam plank on frame”, as I use foam planks instead of the more traditional wooden ones. I use wooden cross sections to help define the shape, and this time, as mentioned before, I had these CNC cut. In the past I have used the hand technique of plotting the cross sections on paper, pasting the plots onto the wood, and then cutting and sanding to final shape. While you can get very accurate cross sections this way, the CNC process is faster and more accurate.

The plug planking sequence.

Extruded aluminum table from granite surface plate reference.
In the past I have used a wood “strongback”, essentially a box, to serve as both reference and stiffener for the plug. On the Linear95 plug, this approach proved to have some problems, not the least of which was warping. For Orion, I chose to use metal. The aluminum extrusions will not deform over the temperature and humidity ranges seen in a shop environment, and are also more accurate than wood. I also chose to make a table instead of a strongback so that the same accurate and stable surface could be used as a reference for the entire project, including other large composite parts, and build-up of the final vehicle. To sweep in the surface of the table, I used a precision granite surface plate, accurate to .0002 inches.
Of course the table itself is only accurate to the total profile tolerance of the extrusions, about .030 inches.

Planking techniques usually use some kind of blank. In my case I bevel them, and then also taper them as needed to fit the desired surface curve. This time I used polyurethane adhesive to attach the planks to the cross sections. In the past I had used screws, but this approach tends to leave depressions and bumps on the surface of the plug that are difficult to fill and sand out.

Daughter Sara helps with the plank ripping on the bandsaw.

Table with blanks  and plenty of clamps, ready for planking.
Bonding meant that I needed to clamp the foam in place while the adhesive dried, and I did not have anywhere near enough clamps to do the job properly. This was where popsicle sticks and “super glue” came in really handy. The so called “superglues” aren’t really, and it is this fact that makes them great for fixtureing. As shown in the photos, I would glue the popsicle sticks where needed with the superglue, to hold the foam until the urethane adhesive dried. Then the sticks were easily popped off, leaving room for the next run of sticks.
After planking, the final step before applying the fiberglass is to sand and fill the foam to get a near perfect, smooth surface. Making a couple of curve templates helps greatly with this process, and also reveals that the method is not perfect. I like to use light weight spackle on the foam. After covering with fiberglass, I will have a smooth hard surface that only needs a few final layers of epoxy based micro to achieve final tolerance.
Left: Very tight fits due to CNC. Center: Popsicle sticks and superglue hold planks. Right: Using toothpicks to close things up.

My House is full of molds!  The molds for the Orion Speedtrike are coming in very quickly now.  I’ve spent the better part of the last month sanding and filling on the plug for the main body, and it looks pretty good now.  While that represents a lot of work, it makes for boring reading, so that is why I haven’t updated more recently.  Just a little wet sanding, and the mold will be made from it very soon.

I’m really happy with how the seat mold turned out, and the whole process went very quickly:

Finished plug in gloss black

Transferring my backbone contour from the plaster cast

Computer generated template attached and ready for cutting
I had originally tried to make a plaster cast of my backside to generate a really nicely contoured seat, but it turned out badly. So I transferred just my spine shape, with a correction for my glutes, to a board, and then to the computer. After tweaking the shape a little in the computer and cutting out two forms to the negative shape, I screwed a thin sheet of PETG plastic to the forms. The resulting shape turned out much better than I thought it would and will be the mold for the seat. Since it is in the computer, it will be easy to make trim templates for seat supports and fitting to the fairing.
Finished seat mold with primer

Wing mold as the computer sees it.
The wing molds are also turning out nicely, though they using an entirely different process than before, as they are full CNC cavity molds

Again, Alex at AVL Looms has been extremely helpful. It took a little work to get the technology do what it is supposed to (which always seems to be the case), but in the end we were able to machine these very nice wing cavity molds. They will require a little cleaning up (read: more sanding and filling), but we are quite happy with them. 

Wing molds CNC machined in MDF.

Closeup of CNC molds with finished (L) and roughed
We did the wing mold first, which covers the aluminum box tube used for the outrigger, to test out the entire process before moving on to the more difficult wheel molds. The wheel molds will be our next project, so check back soon for an update on our progress!

Update Sept 17, 2005: Moldmaking Trials and Triumphs
I‘ve had a mixed time with the mold making since my last update. As seen in the photo below, the wheel fairing plug turned out pretty well. After all the hand finishing required with the earlier wing fairing molds, I decided to do a plug this time since it is much easier to finish a mostly convex surface than a mostly concave one.

I’ve been working on the wheel fairing design for a long time, since it can make or break the aerodynamics for the whole vehicle. I also wanted to have the fairings be interchangeable left and right, so this constraint made things a little more interesting. I’ve documented most of the process below.

Here's the 700c Wheel fairing plug hot off the CNC router. 

 (photo courtesy of Alex Berthet, AVL Looms)
The CNC process starts with some kind of computer model, in this case a surface model of the wheel fairing.
A piece of software called a CAM (Computer Aided Manufacturing) package helps create the tool path for the cutter, shown in green
The tool path is downloaded to a CNC machine, in this case a large CNC router. Here the machine removes most of the material in a “roughing pass”
The CNC router then machines the “finishing pass” to bring the part to final shape.
The part after the finishing pass. Notice that the perimeter still needs to be cut to make it look like the final part at the opening of this section.
Anyone who has ever “finished” wood, knows that the surface of a wooden part is much rougher than it seems. Here, the first primer layer shows all the fuzz left from the CNC machining process
After sanding most of the first primer layer away, a surface pattern emerges that is the combination of all the errors in the process. This pattern is why I chose to make a plug this time, to give me a better surface to sand and fill on. CNC doesn’t eliminate sanding and filling, but it certainly reduces the amount of it!
Mold Trouble!

I had a horrible time with the mold for the main body. Through a combination of errors and poorly timed experimenting, I got the mold all wrong, and damaged the plug in the process. I’ve only just finished repairing the damaged plug after a week of hard work, and will try making a mold again here tonight or tomorrow. 

Way too many bubbles in the mold to be useful for the main body.
When life gives you lemons… Fortunately I have other uses for a mold with a bunch of bubbles in it. So I added support ribs and will use it as a protective cover and/or the start for an offset mold for an internal sub frame.
Good view of the repair to the plug and the finished bad mold. A problem with this plug making technique is that it makes plugs that are extremely fragile. The damage was caused mainly by the very thin sticks (tongue depressors) inserted between the mold and plug to release the part.
Obviously, with this kind of setback so late in the game, I have had to revise my plans for my upcoming race. Fortunately, I have time between applying coats of wax to the plug to write an update!
I looked at the Gorilla tape. It's the stuff made by the same people as Gorilla glue? It appears to have a fiber reinforcement like duct tape. I don't think that would be good to tape a lid on with. I've been looking into vinyl tape, essentially wide electrical tape. Stretchy, smooth, and good adhesive. Comes in colors too!
I've got two pix with the start of the new mold and the finishing of the wheel fairing plug. The new mold is going waaaaaaaaaaay better. Thanks Thom for the suggestion to divide the fiberglass into three sheets. It takes longer to go on but does so much smoother. 
Note the black wheel fairing plug in the foreground, and my "tent oven". The oven doesn't really get hot yet because the space heater has a limiter that I am considering removing. Old heaters didn't have limiters and they worked just fine. But they weren't made out of such cheap plastic either. I used to have a smaller curing oven that I could fire to 170 F, but that space heater is long gone (and not from using it to make an oven!).
I bought material today to make a new chassis for a 650c front wheel. No, no, really. Since everything is so modular on the original chassis, this is not as big a deal as it sounds. Alex will mitre the tubing on a cold saw and milling machine, so that it all aligns really well, and then I should be able to weld it in an afternoon. Of course the fit of all the parts on the different config could become a problem with only 13 days to go. I'll bring both chassis to BM. I've bought everything for the 650c part as I didn't have any before. Rim, tires, tubes, hub, fork. Going with a new chassis AND 650c actually solves a bunch of outstanding issues with the 700c chassis
Haaaaaaaaaaaaaaaaaaaaaaaaaaaa! I have a mold! Only 3 layers, so I will spend the next period reinforcing the heck out of it.

I'm so happy.

The first wing part is bagged.The wheel Fairing plug is finished and has surface coat.  I will bag several parts today. 
The new 650c chassis parts are almost all CNC'd, I keep saying I can weld it in an afternoon if the parts all fit well. Boy I hope I am right!
The newly recreated main mold is getting finished and then it's time for the mold release so I can start laying up the carbon fiber.
Here's the first wing section, fresh out of the mold.
Time to recruit more people to help! There is longer any room to move in the garage.
The first fairing half is laid up, and vacuum bagged.

Things go slow, but they go. Up til 3 last three nights. An old friend came over and we finally got a fairing shell layed up and vacuumed down last 
night . 

I just put it in the oven, so should be done in about 4 hours. Will lay up another wheel fairing while I wait. We will try the other fairing half tonight.

Finally after working on the first shell half for most of the evening, the part 'popped' out of the mold at 11:45 this evening, and looks pretty good. I just wish it didn't look so darned big. Hopefully it's just an optional delusion of being high up on the table. I'll know better in the morning when I take it outside and put it on the ground. I laid up the second wheel 
fairing today. 
Tomorrow is the final make or break. Will try to lay up second fairing half in the morning, and another wheel or two, so I can head out to BM on Wed.

Okay, I propped up the fairing to approximate the  racing height off the table and then stood on the table to get an idea about the size, and it doesn't seem very big anymore :o)

Raymond rolled into Battle Mountain with his Orion Speedbike kit. All the parts are here, all he has to do is put them together. He's planning on running a qualifier run tomorrow, and Thom Ollinger is the designated driver.

In this picture, Raymond's friend Alex works on the Orion Shell. He and many volunteers worked hard day and night (imagine the sound of electric grinders from the hotel room next door at 2:00 am) to put the bike together in time for at least one shot at a speed run.

Ultimately the amount of work required finally became apparent and Raymond spent the last day at Battle Mountain enjoying himself in a sleep deprived haze.


While Raymond didn't finish the Orion Speedtrike in time for WHPSC2005, he was very close. All the parts are done, all he has to do now is make them all work together, and tweak it a little. There is plenty of time to do this before next year's event!

More soon, after Raymond recovers from his recent marathon speedbike building adventure.


Back to the WISIL projects