[QUOTE=Tuckerfan]
One could say the same thing about steam engines in the 1600s (Hero having invented them in ancient times). Or the computer prior to the 1940s (Babbage’s difference engine, a mechanical computer, having been designed in the 1800s), or any other of a number of inventions. Just because nobody’s doing it, doesn’t mean it’s impractical.
[/QUOTE]
I see your point, but the reason I said it that way is because unlike your examples, we’re already experts at sheet metal fabrication, and already experts at casting, and already experts at gasoline internal combustion engines, and presumably we’re not looking for a new invention or even fabrication process, but only doing what we already do, to build what we already make, in a different manner.
[QUOTE=Tuckerfan]
Good question.
Ever been to a modern car plant? Very little welding is done by humans. Robots pick up the sheetmetal, clamp it together and then another robot comes along and welds it, and the whole process takes a couple of seconds. All the robots are isolated so that nobody can wander into the area as no human could react fast enough to get out of the way of the robot’s motion.
[/QUOTE]
Sheet metal tolerances are sloppy loose on all modern car bodies, at least when you compare them to engine tolerances. I’m talking in the neighborhood of ±0.5 mm being about the best case for repeatability in most subassembly marriage operations. This is all designed into the car, so you get good leaking sheet metal alignment (except for all but the most recent Saturns – yuck!) despite these tolerances. In any case, the robot has nothing at all to do with dimensional control (unless its position is grossly, grossly off); it’s all controlled through a series of different types of locater points, which can be holes, slots, edges, and so on. Every buildup of subassemblies up to the marriage points as a body-in-white depends on locaters interfacing with the locater points, i.e., some external fixture (which could be a robot end effector on a non-welding robot). In some cases vision systems are used for roofs, since locater holes present an obvious problem there.
Engine blocks then have the added problem of weld integrity. I left the body shop and built steel fuel tanks for about four years, and with all of the best technology and large budget, there was no way to ensure that every single fuel tank that was welded would be acceptable; as a consequence, they were subject to 100% inspection and rather high scrap rates compared to most stamped assemblies. Even something simple like a resistance spot weld – when you weld 2,000,000 welds per shift – leads to rework and scrap, and spot welding is an easy, simple, proven, ancient process (in relative terms, of course)!
Could we invest the money to guarantee absolute sheet metal precision for a stamped engine block? Maybe… (I don’t think so, but am willing to accept the possibility)… but would it be worth it? The only thing that can go wrong with castings is bad dies, bad material, voids. It’s dirt cheap, proven, reliable.
In the end, yeah, you could build a sheet metal engine block, but you couldn’t currently make it work economically.
(Reminds me – the American Welding Society Semi-Annual Sheet Metal Welding Conference is next week in Plymouth; feel free to drop in! It’s full of welding people, sheet metal people, materials people, stamping people, and I’ll ask if there are any engine people.)
Cost, could be a real bugaboo, I’ll admit. After all, if (ignoring things like R&D and start up costs) you wound up will a block that cost more than your typical Mercedes car, there’s really not much point in doing it.