by Jay Ducharme
(All text and images copyright Jay Ducharme 2000.)

When I was asked to create the 1999 parade float for my workplace, Holyoke Community College in Holyoke, Massachusetts, I told Tom Stewart, the project supervisor, that I would be delighted to build a float--if he let me construct a roller coaster on it. Without hesitation, Tom said, "You're on. We'll call the float, 'Take a Ride on Education!'" The process from inception to completion took about a year. Of that, about three months involved actual construction.

I had a small footprint with which to work. The haywagon that the college uses for parades is ten feet wide by twenty feet long. The deck of the float sits about four feet off of the ground. Because of underpasses along parade routes, area floats can have a height of no more than twelve feet. My idea was to build a roller coaster that I could actually ride. Most coaster trains are about four feet wide. Obviously, there simply wasn't enough room on the float deck to construct something that size. I recalled Kim Pedersen's legendary homemade Termite coaster, which sported a single car about two feet wide. That was about as narrow as a train could be in order to comfortably hold a rider. Comfort was important, because I would be riding this coaster continuously for over an hour during parades.

Unfortunately, even at a minimum width of two feet, there was no way to construct a coaster of that scale for this project. The ride would have room only to rise up a small lift, turn a corner, drop and then turn back toward the lift again. The float had to represent the college. Traditionally, students (sometimes as many as twenty) rode on it. Where would I fit all those people if I built a full-sized coaster? Plus, I wanted not just a coaster but other rides as well (such as a Ferris wheel and a Wave Swinger). A two-foot wide track would leave an area six-feet wide in which to fit scaled-down versions of the other rides. Those other rides would take up all of the free area, leaving no room for people. I had to decide how the float would reflect the theme.

In January, I sat down at my drafting table and began sketching out the footprint of the coaster. I wanted the ride to have an imposing height and several hills. I had to be careful that the lift wasn't so steep that the train couldn't ascend it. But by keeping the lift to less than a 45-degree angle, I couldn't get enough height. As I worked in smaller scales, I didn't have that problem; I essentially could make a taller coaster in less space.

After much pondering, I settled on a one-quarter scale working model. This gave me a train that was one-foot wide, and a total track width (from bent to bent) of one-and-a-half feet. I positioned the footings around the perimeter of the float to free up as much interior area as possible. From off of the lift, there would be a 180-degree swoop-drop down to the deck, a bunny hop and then a fan turn back to the "station." I couldn't see any way of including more hills than that without further reducing the scale. I didn't want to reduce the scale further, because then friction would become a problem. I figured those three hills would look pretty impressive.

I didn't like the slogan "Take a Ride on Education." That sounded as if we were "taking students for a ride." I wracked my brains for weeks searching for a catchier theme that could help unify my ideas. The phrase "Learning Is Life's Biggest Thrill" popped into my head, and I envisioned open textbooks for seatbacks on the coaster train. I had my theme: the thrill of learning.

I bought myself a CAD-CAM program. I decided to put the program through its paces by using it to do all of my float detail work. I began by finishing the groundplan I had started drafting. All the footings would be constructed out of pieces of 2x4 lumber cut to a length of 1 1/2 feet and spaced 1 1/2 feet apart. I was amazed at how much time I saved doing this on CAD-CAM, compared with drawing everything by hand.

When the layout and length of the track was set, I created a profile view using the same technique I used for creating my miniature "Shooting Star:" I found the highest and lowest points in scale, connected those points with a line and then filled in the hills up to that line. I placed vertical lines (the bents) at 1 1/2-foot intervals and then adjusted the height of the lines until there appeared to be smooth transitions into and out of the hills. I measured the lengths of the lines, which represented the heights of the ledgers off of the float deck. I used a scale of 1 mm = 3 in. Why I did this, I now can't recall. I think I simply wanted to cram the entire coaster onto a single sheet of paper. It was a pretty dumb scale to work with, and caused me no end of grief because it was simply too small. It was impossible to find the exact height of any bent. I had to estimate all the heights.

Once I was satisfied with the profile, I created scale views of each bent section complete with ledgers and diagonal bracing. The bents themselves and the diagonal braces were to be constructed out of stock 1x3 wood. There were 35 footings, so the coaster required 35 bent sections. The ledgers were to be made of 2x2 stock. For those of you who are unfamiliar with woodworking terminology, a length of wood referred to as 1x3 indicates that the piece is theoretically one inch thick by three inches wide. The wood is always a bit smaller than this. For instance, a length of 2x4 is usually one-and-a-half inches thick by three-and-a-half inches wide.

I was given a purchase order to buy the supplies I needed. The store I went to was out of 1x3, so I bought 1x2 stock instead. All totaled, supplies for building the coaster cost just under $1000, a lot less than building a real coaster (and less than I was expecting) but still a pretty hefty chunk of change. I then handed the plans over to my assistant, Carcus. He was in charge of cutting, assembling and painting the bent sections, which took about two weeks. He created a jig and basically made an assembly line.

During that time, I began to work on the design of the train. I originally sketched out by hand a very detailed train car in the style of a standard PTC three-seater, complete with both upstops and side-friction wheels. As I progressed into the construction stage, I realized that building a standard laminated coaster track in 1/3 scale would require an enormous amount of time. Additionally, wood grain doesn't scale down; smaller sections of wood become increasingly harder to bend around corners and instead snap like brittle twigs. So I opted to create a side-friction design, similar to Lakemont's Leap the Dips. That required only one layer of running track and one layer of side-friction track.

I calculated the size of the train based on the width of the track. I knew the footings and ledgers were cut to eighteen inches. Each of the bents, which would attach to each side of a ledger, was a 1x3. So by subtracting six inches for two bent sections I would end up with a track width of exactly one foot. I wanted a little bit of play between the side friction rails, but not too much. So I designed the train to be eleven inches wide, with the load-bearing wheels ten inches on center. If I made the track three inches wide, that would allow plenty of surface on which the train could ride. The side-friction rails would be the same width, attached to tops of the bents which would stick up six inches beyond the ledgers. I wanted to build a Prior-and-Church-style train, not just because I was inspired by the Great Coasters International recreation for the coaster Roar but also because I had a very tight footprint to work with and knew I needed articulated cars.

Perhaps the greatest concern when building a working model coaster is friction, which doesn't scale down. As the size of a model decreases, it's as if friction is increasing. I had to choose wheels that would have as little drag as possible, and I would have to be careful that the sides of the train didn't come in contact with the side friction rails.

I called my friend Rich Boehmer and ran my idea by him to see if he could think of any problems I might have. He thought my choice of rolling stock was a wise one, but cautioned me to make sure I beveled the front of each car so that the train had no problem taking corners. In two days, using standard 2" nylon casters, 3/4" plywood and 2x4 scraps, I fashioned the five-car articulated train. I used short strips of metal perf (flexible steel band with holes punched in it) and bolts for the couplings. Thanks to the beveling, the train had an insanely tight turning radius. It would have no problem negotiating the curves.

By the end of August, all the bents had been built and painted. I had Carcus number them as he constructed them. More and more, I began to focus my ideas on the theme of education as an amusement park. I didn't think the circle swing would fit in. Carcus came up with the idea of putting graduation caps on the Ferris wheel instead of seats. That sounded like a great idea. Then all I had to do was find a way to thematically tie in various curriculae at the college. I thought of building three ticket booths such as what you'd see at a carnival, each with a major on it. I chose nursing, chemistry and electronic media. Each booth would have "3 Credits" written on it, similar to a sign for "3 tickets." The booths would serve two purposes: they would give people on the float a place to sit, and they would reinforce the theme. As a bonus, they were easy to build.

The next big construction project was the Ferris wheel. I built this one myself in two days. I drew up some CAD drawings, but ended up using mostly my instincts. The wheel was made entirely of 3/4-inch-thick plywood pieces that I pre-cut. The buttresses to support the wheel were made from 2x6 stock. The assembly went smoothly. The finished structure stood an imposing six-and-a-half feet high and was extremely heavy.

When the wheel was completed, I cut all the track and side-friction rails. I took sheets of 1/4-inch-thick plywood and ripped them down to 3" wide strips. I needed 240 board feet of track and rails to complete the 60-foot circuit. I then loaded all the pieces into the college's dump truck and hauled them up to the garage, where the float was stored.

I sorted out the pieces that I needed. Then I covered the float deck with fresh half-inch thick plywood. I stapled down fake grass matting. Then I began "pouring" the footings for the coaster. Using the groundplan I drew up, I measured the distance from the end of the float deck to the first footing. I lined up the footing (a 2x4) with a carpenter's square and secured it to the deck with 3" multipurpose screws. Then I measured eighteen inches forward and secured the next footing after squaring it. I did this for both sides of the float where there were straight runs of track. For the turns, I positioned the footings at the center, perpendicular to the float deck, and secured those. Then I lined up the remaining footings, stood back, eyeballed their position and adjusted them until the curves looked correct. It's difficult for me to describe the process. I simply knew when the footings were properly aligned. Securing all of the footings took me about a day-and-a-half.

Erection of the bent sections took another day-and-a-half. The process was fairly simple: I started with bent section number one, set it against footing number one, drilled a pilot hole through the bottom of each bent and then secured each with a 2" multipurpose screw and fender washer. It was important for stability that all bent sections faced in the same direction, with the diagonal bracing pointing down toward the edge of the deck. I noticed that one of the bent sections on the lift hill suffered from severe warpage (from poor-quality wood). I was able to correct the problem by adjusting its position on the footing. Occasionally, an unseen knot would cause the screw heads to snap off. Outside of that minor problem, the coaster structure took shape quickly.

When Carcus returned, I had him work on the structure's stompers, diagonal pieces attached from each bent down to the footing in front of it. These braces "plumbed" the structure so that the bent sections rigidly stood straight up. When that was completed, we applied the ribbon bracing to the structure. The ribbon bracing in this case didn't have much structural integrity; it was made from 1-inch wide plastic moulding. But it was easy to shape and helped make the structure look like a real coaster.

The next phase of the project, tracking the coaster, posed some interesting challenges. I took two of the three-inch wide strips of track I had cut and placed them against the inside edges of the bents going up the lift hill. Then, as a test, I took the lead car of the train and placed it on the tracks. The car fell through between the tracks. It didn't fall through by much, but it would never stay on the tracks if it moved. I discovered my earlier misjudgement: the structure was designed with 1x3 stock in mind. I purchased 1x2 stock instead. The track was now effectively two inches thinner than it was designed. If I had cut the plywood track to four inches wide, everything would have been fine. By then all the plywood had been cut; there was none left. I stood staring vacantly forward for quite some time. If I were able to slide the track closer together by two inches, everything would work out. But then the side friction rails would be too far apart.

The solution I finally came up with wasn't very elegant, but it worked: I cut seventy 3/4"-thick plywood shims and placed one between each bent and the side friction rails. This allowed me to bring the track closer together by one-and-a-half inches and keep the train centered.

I began tracking the coaster backwards, so that I could test the train at each hill to make sure it would make it back to the lift. I found that each time I tried to bend the 1/4"-thick side friction rails around a turn, the wood snapped from the stress. So I went back to the shop and cut two-and-a-half-inch wide strips of scrap luan that I had. Luan is a very thin and supple plywood. I chose to go 1/2" thinner with these simply because I didn't have very much; I would have run out by making the side-friction rails 3" wide.

Generally, the track contoured well. I affixed it to the ledgers with 1" multipurpose screws. One of the college's maintenance workers offered a suggestion for bending the track around the turns: he recommended cutting entire turns out of a sheet of plywood. At first, that sounded like a good idea. As I attempted it, though, I found that it was difficult to transfer the exact curvature to the plywood. Another problem was that the edges of the plywood chipped and broke apart as I cut across the grain. I found that the best way to shape the curves was simply to cut short sections of the 3"-wide plywood strips, long enough to reach from the center of one ledger to the next, and butt the edges together to gradually form the curve. I did this on the final turn, and then sanded down the areas where the two pieces met so that there were no bumps in the track. This method was very effective.

I applied straight sections of the track and side friction rails so that one piece overlapped another. It was as if the train would be going down small steps along the entire route. The small downward steps in the track didn't appreciably increase friction. In retrospect, however, I wish I had assembled all of the track the same way I did the final turn, which was a bit more work but produced much better results and looked much more elegant.

I made an error in calculating the heights of two ledgers (which I thought was pretty good overall!). They were too low and would have produced a triangular spike in the track. So I simply used some spare footings turned on their sides to effectively raise each ledger up three-and-a-half inches. That generally fixed the problems. The ledger at the bottom of the second drop was still a little low though. The 1/4" plywood track kept snapping in half when I screwed it down. So I used luan at that point, which contoured more smoothly.

When I had completed the top of the second hill, I placed the lead car there and let it go. It slowly traveled down the hill and nosedived into the track, nearly flipping over. Even though the lead car had four wheels, it needed the other cars in the train to keep its nose pulled up. So I assembled the entire train at the base of the lift. I tried pulling it backwards. The train kept jackknifing and the wheels dropped through the track. That was not encouraging. But little by little, I worked it backwards to the second hill. I gave it a slight push forward and held my breath. With a roar, the train careened down the hill, turned gracefully up the next hill and thundered toward the lift, where it promptly ran off the unfinished track and came to a crashing halt. Seeing the train successfully complete that part of the course gave me a big rush of adrenaline. The train wasn't damaged. I needed to complete the first drop track so that I could test the it on the entire circuit. Unfortunately, time was running out. It was Friday, October eighth. The parade was October tenth.

I frantically worked into the night. The change in height from the end of the lift hill to the beginning of the first drop was so severe that I couldn't bend the 1/4" plywood track to fit it. So I once again used strips of luan sandwiched together for each track side. After a few hours, the entire course was complete. I stood back and looked at it. I couldn't believe it! I had designed and built my first working coaster. Or so I thought.

I pulled the train slowly forward to the top of the towering lift. I took a breath and released the train. In a deafening instant it had reached the bottom of the first drop, at which point it smashed through outer side friction rail and stopped dead. My heart sank. I expected it to lose its momentum to friction and fail to crest the second hill. But I hadn't expected this disastrous result. The couplings on the train were all bent, but there was no damage beyond that. I removed the train car by car and repaired the couplings. Then I repaired the damaged side friction rail. While doing so, I discovered that the reason the train crashed was that it fell through a space between the track and the outer side-friction rail. So I shimmed in the outside rails by another 3/4" and pulled the train back to the top of the lift. I took another breath and watched as the train zipped down the first drop once again. This time, it rounded the corner beautifully, started up the second hill and stalled out before cresting it. There was too much friction at the bottom of the drop. The train would scrape against the outer side friction rail and basically drag itself to a halt. By then it was late in the evening. I called it quits for the night. The rest of the float still had to be assembled.

The next morning, I arrived early at the garage with Karen. She added some decorations to the Ferris wheel as I made sure everything was properly secured. It began to rain very hard. Tom Stewart arrived. We hitched up the float to the college pick-up truck and drove off to the parade site. It continued pouring all day. I simply left the train inside a "ticket booth" to try and keep it dry. The glue with which I assembled the Ferris wheel began dissolving. Instead of riding on the float, Karen and I stayed in the truck's cab. Our son, dressed as the school mascot, and his friend stood in the pickup bed. When we drove by the reviewing stand, the announcer bellowed, "And here's the Holyoke Community College chemistry float." He proceeded to describe all the things I didn't have time to put on the float, like the graduation caps. I refused to go out like that. The next parade was a month away; I had that much time to get it right.