Ford's Law?

April 24, 2008 5:44 PM

Is there a precursor to Moore's Law for internal combustion engines? If so, can you provide details, please?
posted by Exchequer to Travel & Transportation (7 answers total)

I suspect there is.

Figure 8 in Meindl's 1984 paper on Ultra Large Scale Integration (IEEE Trans. Elect. Dev. 31, 1555, 1984) plots production of structural materials (steel, plastic, aluminum) along with single crystal electronic materials on semilog axes. All have exponential behavior, with about the same slope. I could swear I have seen a similar graph for engines (and other technologies) but I can't remember where.

Interesting you should ask this, I made just this point in my class this morning.
posted by Wet Spot at 6:08 PM on April 24, 2008

I would guess, no. One of the problems is that unlike computer chips, not every auto engine is designed to run at it's maximum performance potential. What is the metric that you want to plot? My guess would be hp/L. In that case, there were some turbo 4-cyl F1 engines from the '70s that put out over 1000 hp/L, and that's never really been exceeded since. Except maybe for top-fuel dragsters, but those engines disintegrate from the moment they're turned on and never make more than one trip down the strip.
posted by hwyengr at 6:32 PM on April 24, 2008

Response by poster: I was thinking further back, from the the origin of the IC engine to today and beyond.

And from a metric standpoint, maybe efficiency? Energy in vs. energy out?
posted by Exchequer at 6:40 PM on April 24, 2008

Questioning the metric is the key here. Integrated Circuits all share the same goal. Faster instruction cycle. Although recently "while consuming less power" has come up too.

Internal Combustion Engines have too many different goals for the whole industry to push together in the same direction and achieve the same kinda growth. More powerful, more fuel, efficient, cleaner burning, these have all been the goals at different times/companies rather then there always being one solid goal above all others.
posted by oblio_one at 7:39 PM on April 24, 2008

I suspect that the closest you'll get is the progress of the Rolls-Royce Merlin engine, during World War 2.

If you look at the graph on the Moore's Law page that was linked, you could get pretty similar lines for the Merlin, for both horse-power and power-to-weight ratio, when plotted over the lifetime of the basic design.
posted by veedubya at 7:58 PM on April 24, 2008

This is stretching my rudimentary knowledge of thermodynamics, but I'm pretty sure there's a firm ceiling in the efficiency of the process of converting liquid fuel into mechanical motion (friction losses, heat loss, etc). There have certainly been advances in reducing the internal friction and heat transfer of an IC engine's parts, but it doesn't seem likely to me to get exponential improvements in efficiency, except perhaps during the early development.
posted by hwyengr at 10:04 PM on April 24, 2008

This is stretching my rudimentary knowledge of thermodynamics, but I'm pretty sure there's a firm ceiling in the efficiency of the process of converting liquid fuel into mechanical motion

Yes, since they're all Carnot heat engines, they follow the corresponding theoretical limit.

I should mention that Moore's law has nothing to do with speed -- it's transistor density. He claimed (truly exceptionally at the time) that they would squeeze more transistors onto the same area chip. The speed thing is a nice side-effect.

The closest you can get in engines is the incremental increases in efficiency (about 5%/year). Again, we run into the metric-designation problem. This is not exponential growth on the same order as Moore's law, but over time you might see something similar (an increase factor of 1.05 in engines versus close to 1.33 in transistors).
posted by spiderskull at 2:07 AM on April 25, 2008