by Jay Ducharme
(All text and images copyright Jay Ducharme 2018.)

THE ROLLER COASTER

I assumed that this was going to be the biggest job. Although I had constructed coaster models before, both big and small, this would be the most extensive recreation I'd ever attempted. Plus, I wanted it to be accurate, matching the original ride.

I began by separating the plastic ladders from their molds. I started assembly at the lowest point on the ride, the below-grade bents (vertical members) just before the lift. I wanted to work backwards from their, installing all the smaller hills and leaving the big hills (the outbound track) for the end. I kept my iPad nearby, displaying the coaster's original blueprint. I had to keep reminding my math-impaired mind that whatever the height was on the blueprint, I had to deduct 14 feet to account for the original elevation change. So a 17-foot-high bent was actually only 3 feet off of the ground on my scale model. I also kept a ruler nearby so I could verify the height of the sections.

I wasn't sure what I was going to use to affix the plastic ladders to the Homasote. Wood glue wouldn't work. The caulk would be too flexible. I decided to use my hot glue gun, and I'm glad I did. After widening the footer holes with an awl (and repositioning them to be 7/16" apart instead of 1/2"), I'd squirt a little hot glue inside them and smoosh in the ladder. The hot glue gave me about thirty seconds to get the bent positioned before it set. The plastic ladders were then impossible to remove without tearing up the Homasote. Within about four hours, I had completed about a third of the ride.

As I approached the turnaround, I noticed a discrepancy. The blueprint had fewer footers than what I had punched into the table. In other words, the coaster was a bit shorter than the way I had rendered it. So I decided to simply stretch out the two small hills before the turn. That was ironic, because that's what happened to the ride when it was actually built: the original blueprint was reduced in length by 85 feet. My render had added even more length to the ride!

Constantly checking my measurements with the ruler was getting tedious. I finally realized that the ladders were reliably consistent: the ladder "rungs" were 1/4" apart. In scale, that equaled four feet. All I had to do was count the rungs and I'd know the bent height. That saved me a lot of time. But when I got to the turnaround, the bents didn't seem to scale properly to the blueprint. They looked absurdly tall. I noticed that the blueprint listed some of the lower bents near the turnaround as being 5 feet above grade. If I subtracted 14 feet from that, the ride would have been in a deep pit. And it wasn't at that location. So I assumed that the grade had changed and the bents really were at five feet. But then when did the blueprint shift to a different grade? That wasn't clear. So I decided to follow what was written in the blueprint. But that didn't quite work out because, as I mentioned, the structure looked absurdly tall. So rather than 59 feet, I made the tallest part of the turnaround 45 (reverting to that 14 foot deduction). It still looked a bit tall. When I compared it to the photo I took of the actual turnaround, it should have been much flatter. But then I looked at the profile in the blueprint; it looked like my model. So I guess that over the years, the turn must have been flattened to slow the train; the turn took a real beating and had to be rebuilt each season.

I worked into the evening to finish the turn and stopped for the night. The coaster was more than half done, and it was starting to look like the real thing.

The next morning I resumed work on it, completing up to the second hill before lunchtime. The heavy rain of the previous day had let up, so I took the opportunity to spray paint the new pack of ladders I received the day before, in case I needed them. One thing I noticed on the model was that the hills were coming out spiky, like triangles instead of arcs. To make the hills, I had been increasing the bent height at pretty consistent four foot increments. But that's not how Herbert Schmeck designed the ride. The bents sharply increased and decreased height going into and out of the hills, but had a very slight change of height along their tops, creating that signature parabolic arc. I wasn't happy with the way mine came out, but changing them would mean ripping out (literally) every hill. I pondered the dilemma over lunch and then got back to work.

All I had left to do were the "speed bump" at the bottom of the lift and the lift hill itself. I got into a pretty robotic rhythm: clip the ladder to the right height, file down the flashing, widen the footers with the awl, fill the holes with hot glue and then position the bent. The lift was one hill where my consistent increments worked. Ironically, that hill ended up with a slight dip in the middle. But after a few hours, I had connected the structure back to the station. The Mountain Flyer was finished -- at least, this stage of it. I still needed to attach ribbon bracing, batter bracing, handrails and track. But that would happen at Heritage Park. I just wanted to get the structure up for now.

I did further modify the turnaround, getting it to match the photo I had. I also tweaked some of the bents on the big hills. That helped a little, but it would require more serious surgery if I wanted to create true parabolas. But I still had a lot more work to do on the model before it left my basement in two weeks. For now, I figured this version would have to do.

Well, it didn't do. The next morning I looked at the profile and couldn't live with the spiky hills. So I tediously tore out the offending bents and cut new ones to maintain the arc at the top of the hills. I modified both the second and third hills, plus the tall seventh hill heading back to the station. It wasn't perfect, but it looked a heck of a lot better and I could live with it now. The way to make it more accurate would have been to continue adjusting the slopes going into and out of the hills. But I was concerned that would lead to another big problem: my model was actually longer than the blueprint. If I increased the slope on the hills, that would leave me with long stretches of ground-level track. And remember that the actual coaster had been shortened even further than the blueprint. So I was willing to accept this compromise and keep moving on to the next phase of the model.


Remaking Mountain Park