Too Complicated? I Finally Finished My Car Models!
This should be my last blog post about car engines for quite a while, because after way too much time, I have finally finished what I promised myself I would not do: connect my already complicated internal combustion engines to my existing model car chassis. The result is a new pair of complete car models, the most complex of which has 609 (!) parts and takes maybe 5-6 hours to put together. (I can do it faster, but that doesn’t count.)
When I first designed a model car body, the only engine I had to connect to it was the radial engine (one of very first models I ever designed). While there were a few experimental radial engine cars, my car body was clearly crying out for a more conventional style of engine. But it would be two years before I got to the point of making such engines: I had a lot to learn at multiple levels before I could attempt such complex designs.
Parts list for the Radial Engine (click to expand).
For every kit, I have a Mathematica notebook that lists every part, including the part number, a small image, the cost, and the required and included quantities. (Different because sometimes we include more than needed of the small parts.) I show here the parts list for the radial engine, one of my simplest kits. (I’m in there under Labor, with a time estimate, so that this is included in the calculation that ultimately determines the list price of the kit.)
From these master parts lists I distill production parts lists that show only the information needed when actually putting together kits (with the count in huge font because that’s the easiest thing to overlook). This gets put on a clipboard with a magnet that gets slid down over each line as the part is added to the batch. (You can read here how I make these kits in batches of 20 at a time.)
Getting the production parts list right is of course crucial: if it’s wrong, then every kit is wrong. For a simple kit like the radial engine, that’s not too difficult. It’s pretty easy to look at the prototype and just count how many of each part it contains. With my early, simple kits, I would generally go directly to making a batch of 20 after the design was completely finished and I’d made up the parts list and checked it a couple times against the final prototype.
But as the kits get more complicated, this does not work anymore. The radial engine has 21 different kinds of parts and a total of 58 parts. That’s pretty easy to count even in an assembled model. At the bottom of this blog post I leave for your amusement the un-redacted production parts lists for the dual overhead cam car, which has 147 different kinds of parts and a total of 609 parts. It’s just not possible, for me at least, to correctly count the hundreds of individual parts in the most complicated kits. Even taking a prototype completely apart, sorting the parts, and counting them as separated parts is not 100% reliable. There are so many, and they are so easy to lose, that I never completely trust such counts.
So I work in the other direction: I make up the best list I can, pick the parts for a kit based on that list, and then try to build it. This never works the first time. So I go back and fix the list, and then make, usually, 5 kits based on that list. When I put the first one together, it may be correct, but often not, in which case I fix the list, and very carefully retrofit the remaining four test kits. Only after I’ve built all five of the test kits, and all of them came out correct with exactly the expected number of left-over parts, do I trust that the parts list is correct.
Then the parts list becomes an object of superstition. Even if I look at it and think, there is no way that part is actually used, I won’t change it, because too many times I have been proven wrong: the part is used, but in a really obscure place that I couldn’t think of at the time. (The most recent example of this was when I tried to remove the 3/8” screw that’s used in the friction bushing of the motor mount in the car models. I had to put it back in the list, because, yes, that screw does actually exist whether I remember it or not.)
The n-squared aspect comes from the fact that, not only do I need to make more copies of the kits before I trust the parts list, but making those kits takes correspondingly longer as the number of parts increases. Making two radial engine kits takes maybe half an hour. Making half a dozen complete cars is days of work. Here are just some of the cars and engines I built in the parts-list-finalizing process for the two car models I just added.
Phew, just looking at this picture is exhausting. And then we had to record the assembly instruction videos.
