SandySandfort on June 25, 2011, 07:49:40 pm

dough560 on June 27, 2011, 12:45:09 am
Interesting,  The original design specs should be public domain by now.  I believe there are two motorcycle companies who have cycles with similar wheel and drive systems.  If they haven't patented the steering system or its not similar to the D-Car, I wouldn't think there would be a problem building a modern update.

Back in the 70 / 80's there was a company experimenting with a aerodynamic cab-over semi-tractor.  Allegedly it got about 15mpg (compared to 5 or 6 mpg).  I remember there were no doors in the conventional  positions.  The door was located at the back of the cab. 

quadibloc on June 27, 2011, 09:36:53 am
There was a discussion of the Dymaxion Car over on the Modern Mechanics blog. Apparently, the design is flawed, despite the prestige of its creator.

Here is one article about it from there:

http://blog.modernmechanix.com/2006/09/13/the-automobile-of-the-future/

Apparently, though, this one doesn't have the discussion I recall seeing. Perhaps it was on another site that I found when Googling. Ah, now my memory is refreshed. A prototype was involved in an accident which did not have anything to do with its design, and this led to its being abandoned.

But rear-wheel steering will be hard to get used to.

wdg3rd on June 29, 2011, 09:28:34 pm
There was a discussion of the Dymaxion Car over on the Modern Mechanics blog. Apparently, the design is flawed, despite the prestige of its creator.

Here is one article about it from there:

http://blog.modernmechanix.com/2006/09/13/the-automobile-of-the-future/

Apparently, though, this one doesn't have the discussion I recall seeing. Perhaps it was on another site that I found when Googling. Ah, now my memory is refreshed. A prototype was involved in an accident which did not have anything to do with its design, and this led to its being abandoned.

Drunk driving Chicago politician.  His crew had the connections to spin it against Bucky.

Quote

But rear-wheel steering will be hard to get used to.

Mastered driving a forklift several decades ago.
Ward Griffiths        wdg3rd@aol.com

Men will never be free until the last king is strangled with the entrails of the last priest.  --  Denis Diderot

BMeph on July 01, 2011, 03:42:54 pm
I wonder who owns the rights? Imagine a Dymaxion built with modern technology and materials.

Dammit, I want my Flying (Dymaxion) Car!  :P ;D

ZeissIkon on July 04, 2011, 04:16:15 pm
But rear-wheel steering will be hard to get used to.

Mastered driving a forklift several decades ago.


However, there's a lot of difference between driving a fork lift at 6-7 mph and driving a car at 60-70 on the interstate.  In the latter condition, divergent steering (the big flaw in the rear-steer Dymaxion) is a no-warning killer; let the nose wander by as little as a degree, and the next thing you know, you're rolling down the freeway sideways (as in wheels-over-roof) after pulling a 90 turn in a fraction of a second.  Even a fork lift can sneak up on you at 6 mph.

All the reverse-trike motorcycles, as well as the few cars (Morgan, Mitsubishi, one or two others I don't recall immediately) with that layout that have been even mildly successful, have been front-steered, effectively like a modern car but with only a single rear wheel (interestingly, the motorcycle derived models are all rear drive, the ones built from ground up as cars mostly front drive).  And there's very little earth-shaking advantage to a front-steered reverse trike except that it's a little lighter than a comparable four-wheel machine, and can be registered as a motorcycle in most states (cheaper, but requires a special endorsement on the driver's license in many states, as well as a helmet).

I'd have to characterize the Dymaxion as one of Fuller's biggest failures.

NeitherRuleNorBeRuled on July 06, 2011, 04:09:43 pm
[D]ivergent steering (the big flaw in the rear-steer Dymaxion) is a no-warning killer; let the nose wander by as little as a degree, and the next thing you know, you're rolling down the freeway sideways (as in wheels-over-roof) after pulling a 90 turn in a fraction of a second.

I'd have to characterize the Dymaxion as one of Fuller's biggest failures.

This could be resolved today with minimal problems; simply add "computer assisted steering", and a solid state gyroscope.

quadibloc on July 06, 2011, 06:15:31 pm
and a solid state gyroscope.
They have gyroscopes with no moving parts these days?

Well, I suppose it's possible. After all, electrons can precess, because they have spin...

So it isn't necessarily true that the plans for a solid-state gyroscope would show one how to distinguish between inertial reference frames (and, hence, very likely also show one how to design a faster-than-light drive)... but it would be worth a look.

NeitherRuleNorBeRuled on July 06, 2011, 07:01:42 pm
and a solid state gyroscope.
They have gyroscopes with no moving parts these days?

Well, I suppose it's possible. After all, electrons can precess, because they have spin...

So it isn't necessarily true that the plans for a solid-state gyroscope would show one how to distinguish between inertial reference frames (and, hence, very likely also show one how to design a faster-than-light drive)... but it would be worth a look.

OK, not completely solid state (except for fiber optic gyroscopes, which are rather expensive), but very close.  Look at MEMS gyroscopes, which are encased in IC packaging; they can be had in bulk for ~5USD each.  They are used in video game controllers (Wii, Sony), smart phones and tablet PCs -- and yes, modern automobiles:

http://www.eetimes.com/electronics-news/4197686/Automotive-applications-fuel-MEMS-market

They have been around nearly 10 years.


ZeissIkon on July 20, 2011, 03:38:25 pm
This could be resolved today with minimal problems; simply add "computer assisted steering", and a solid state gyroscope.

Will that work like the "computer assisted engine management" in a 2009 Toyota?

I have no desire to drive a car controlled by software; I've written some and used much more, and never seen any program of significant size that was bug free (I'm convinced "bug free software" is an oxymoron).  Given that, I'd rather not have software controlling my engine/transmission or my steering.  If I find I absolutely must have a vehicle that'll get 150 mpg on straight ethanol, I'll start with an 80 cc two-stroke engine and a bunch of bicycle parts, and get there without a single computer component in the critical systems...

dough560 on July 21, 2011, 04:47:13 am
That sound like something that's already been done.  Popular Mechanics a couple of years ago?  Mother Earth News?

J Thomas on July 21, 2011, 06:01:05 am
This could be resolved today with minimal problems; simply add "computer assisted steering", and a solid state gyroscope.

Will that work like the "computer assisted engine management" in a 2009 Toyota?

I have no desire to drive a car controlled by software; I've written some and used much more, and never seen any program of significant size that was bug free (I'm convinced "bug free software" is an oxymoron).

If you have a problem to solve that demands a program of significant size, chances are the specification of the problem has many bugs.

But if all you want is a fairly simple result that happens quicker than a human is likely to do it, or one that's counterintuitive for a human without lots of training, then that might be quite feasible. So for example, computer-controlled anti-lock brakes are probably going to be fine.

Quote
Given that, I'd rather not have software controlling my engine/transmission or my steering.  If I find I absolutely must have a vehicle that'll get 150 mpg on straight ethanol, I'll start with an 80 cc two-stroke engine and a bunch of bicycle parts, and get there without a single computer component in the critical systems...

If it's simple enough, you can design mechanical feedback systems for anything you might otherwise use a computer. That might even be cheaper if you're building hundreds of millions of them. The big advantage of software for simple control systems is that it's cheaper to tweak. Cheaper to use a microprocessor with software than build an electronic circuit because bugs are easier to find and fix. Cheaper to use a microprocessor with software than mechanical feedback systems for the same reason, except for the parts of the design that were carefully arranged to allow easy tuning.

But of course the more complicated you make it, the more room for things to go wrong. One advantage of a design that leaves out software completely is that you have to keep it simple to get it to work at all. So the temptation to make it too complicated is less....

ZeissIkon on July 28, 2011, 04:02:38 pm
One advantage of a design that leaves out software completely is that you have to keep it simple to get it to work at all. So the temptation to make it too complicated is less....

Another advantage is that interactions between subsystems also tend to be simpler (though not always -- I've seen stuff in nail guns that's pretty counter-intuitive to trace).  Still, I'm a firm believer in being able to see the parts of anything I'm going to trust my life to.  And no, I don't drive a 1960s vintage car, but the electronics that control the fuel injection in my van will pretty much just kill the engine if they fail -- yes, that could be fatal, if it happens fifty miles from the nearest town in the middle of Nevada, but it's a lot easier to manage than the car computer suddenly deciding I've commanded top speed, and then ignoring inputs to brakes, accelerator, or even ignition switch.  As for ABS, you can build a pretty good ABS without a true CPU (you'll probably use something like an ASIC to detect impending wheel lockup, but those run on simple enough programs to test rigorously) -- but I learned to drive before ABS existed even in semi rigs, and learned to modulate the brakes, even in conditions like loose gravel, deep mud, and glare ice.  I can't quite beat a modern ABS equipped vehicle of similar weight and with the same tires, but I can come very close...

J Thomas on July 29, 2011, 06:24:04 am
One advantage of a design that leaves out software completely is that you have to keep it simple to get it to work at all. So the temptation to make it too complicated is less....

Another advantage is that interactions between subsystems also tend to be simpler

They usually have to be simple. Mechanical parts are made to tolerances and then they wear. Despite anything you do they collect dust. To make them reliable you need ruthless simplicity.

But software is supposed to be cheap, and it's easier to add complications than to remove them. Usually the money isn't there to keep it simple. But it can be done, and it should be done.

Quote
And no, I don't drive a 1960s vintage car, but the electronics that control the fuel injection in my van will pretty much just kill the engine if they fail -- yes, that could be fatal, if it happens fifty miles from the nearest town in the middle of Nevada, but it's a lot easier to manage than the car computer suddenly deciding I've commanded top speed, and then ignoring inputs to brakes, accelerator, or even ignition switch.

Yes, the hope is that when it fails it will fail safe.

Quote
As for ABS, you can build a pretty good ABS without a true CPU (you'll probably use something like an ASIC to detect impending wheel lockup, but those run on simple enough programs to test rigorously) -- but I learned to drive before ABS existed even in semi rigs, and learned to modulate the brakes, even in conditions like loose gravel, deep mud, and glare ice.  I can't quite beat a modern ABS equipped vehicle of similar weight and with the same tires, but I can come very close...

Sure. So, a relatively simple and cheap electronic design can give anyone the chance to match your expertise. That's a good thing. This particular time, the goal was simple enough to easily allow that.

I say, software is potentially quicker and cheaper to tweak than other designs. So if you can maintain simple criteria and simple interfaces, it can be quicker and cheaper to get things working that way.

But there's always the temptation to have the software do more, and then get lost in so much complexity that you can't find the errors or even be sure which results are errors. And generally our software guys wholeheartedly glut themselves with that temptation. Their training is in how to handle complexity and not about how to avoid it.