Building a 100mpg consumer vehicle - Page 10 - Fuel Economy, Hypermiling, EcoModding News and Forum

botsapper · 01-04-2012, 12:32 PM

We see you prefer the tandem seating and the front track width has to be configured to 'original' width specs because of steering geometry & engine clearances.

I sometimes wonder why there are very few attempts & examples; of using a relatively intact FWD section mated to trailing rear wheel. Just offering alternative configuration that could be 'easier' and still offers the 'comfort' criteria. The Peugeot 20Cup basically keeps relatively intact the FWD engine/drivetrain of a Peugeot 207 and integrated it to a motorcycle trailing arm. The seating is tight side-by-side, but could be use an aero roof. Also see the next gen EV handling dynamics test (narrow two-wheeled rear track) - Peugeot EX1.
256 Fifth Gear - Peugeot 20 Cup - Car Videos on StreetFire

http://www.youtube.com/watch?v=FTXbRZeQ2GA

The VW GX3 used a VW Lupo for their engineering exercise.

http://www.youtube.com/watch?v=e7DXoe5mgFc

Ken Fry · 01-04-2012, 02:40 PM

Quote:

Originally Posted by drmiller100

Back seat seating. Yes, there are issues without great solutions.

Do I want the car narrow? Yes.
Do I want a 1.5 inch roll bar above the driver? Yes.
Do I want a back seat? Yes.
Do I want side impact protection? Yes.
Do I want to build a custom door? No.
Do I want the front door to open inside of a standard parking spot? Yes.

Which one of the above am I willing to sacrifice?

Narrowness: You definitely will not want to narrow it at the front, because if you do, you will have to make the suspension brutally stiff to have enough roll resistance for good handling. (And as you point out, narrowing the powertrain is effectively out of the question, anyway.) Even with its incredibly wide track and (eventually) lowered stance, the Aptera rolled too much with suspension geared for comfort. Just to get through a simple handling test that any sedan can pass easily, (in the X Prize) they had to stiffen the suspension. The T-rex has much wider track than you do, for this reason.

BTW, don't repeat all the Aptera mistakes with suspension geometry. Unless you intend to give your wife a wild ride, go for consistent understeer. That means the front wheels must tilt at the same rate as the rear wheel as the car corners. The original Beetle geometry is good in this respect. Ditto 2CV and DS series. The Civic geometry would be expected to cause dramatic oversteer, when applied to a car in which the rear wheel tilts adversely under cornering loads.

1.5 inch roll bar: This should be 1.75" .120 DOM. Even at that, it will need to be triangulated to function correctly. If you figure a 4 G load, and a loaded weight of 2000 lb (1500 empty vehicle + 500 passengers and groceries) then you need to resist 8000 lb at the center of the roll bar. You can test this with an instrumented portipower. I'd guess that your current structure will permanently deform easily at fraction of that load. The narrowness of your cabin helps, because the span from left to right is less than it otherwise would be.

I gather you plan to provide roll protection for the back seat person too. The current longitudinals are too long to have meaningful stiffness in bending, so you will put a roll bar behind her head, I assume.

Side impact protection: This is another reason to use larger tubing. Loads in side impacts go up to about 50 G (almost 100G in the worst-engineered cars). Engineered cars have very sophisticated structures that are impossible to match weight-wise with a ladder frame. (This is one reason that some SUVs and trucks have had atrocious side impact protection.) A properly designed fully-triangulated space frame can be both light and strong, but encroaches on passenger space. In most modern cars, even the windshield glass is used as a load bearing element, providing both torsional and bending stiffness. Monocoque is hard to beat.

I'm sure you do not fall into this category, but many people think that a fiberglass body hung on a frame makes a light structure. The old Corvettes show that this is not the case at all, with a mid 70's Corvette weight of about 3500lb. This is unbelievably heavy for such a small two-seater car. (The five seat Accord of 1976 was about 2000 lb, due to its safer and more sophisticated structure.)

Custom Door: I agree. Building a door is a hassle, and the doors from your hatchback are already longer, if I recall, than those for the sedan. Getting into your back seat should not be more difficult than for the Civic Hatchback it you keep the door location relative to the front seat tracks the same. The Civic already has about the minimum headroom, so you may need to raise the door sill and roof height somewhat (or lower the floor to a point below your shop floor). No need to reinvent the wheel here: I'd just duplicate the arrangements in the Civic: the hatchback version was fairly low outside, if I recall, and rear seat headroom was just barely adequate. An extra few inches height will mean almost nothing in terms of aerodynamics, but will ensure marital bliss.

The stock hatchback should have a (structural) advantage in length, because the rear wheels can be further forward relative to the back passenger backrest. This, in combination with the larger circumference of the hatch back body, gives the hatchback a huge advantage in weight-to-stiffness ratio. If you are willing to put up with some handling imprecision and creaks and stress cracks in the body, perhaps you can match the weight of the Civic. Otherwise, the expectation would be that your vehicle will be heavier, due to the inefficiency of a long, thin structure, and the lack of monocoque advantages. But even matching Civic weight will be an accomplishment to be proud of, with a home-built car.

It will be interesting to see what you have planned to avoid wakes behind the front wheels. (Perhaps Aptera style wheel fairings... you might need to park with the wheels cocked to one side to clear the door?) Going from full Civic width to the door width (within only a few inches of car length) without creating huge wakes will be a challenge. Perhaps the door hinge could be out where it was in the Civic, but the doors could taper a bit more toward the back of the car. If you can somehow handle the transition from front width to side width, you may be able to get better-than-Civic aero. It would be interesting to know if the dimpling on the back of the Corbin sparrow fenders actually had any effect in keeping the flow from being completely separated and turbulent -- I doubt it. But perhaps you can dig up CFD studies on it -- or put a Sparrow in a wind tunnel... or borrow one to drive around with tufts. I have a couple friends that have them, so let me know it you are thinking along those lines.

I'm really impressed with the progress you've made! It will take really extensive engine mods to get close to 50 mpg combined (as measured today), but you will have a very cool car, nevertheless. The current Accord V6 is variable displacement, and that can help a little -- maybe 1 or even 2 mpg, vs standard VTEC. (The variable displacement system on the Accord includes audio from the sound system to cancel the odd sounds of an engine running on few cylinders... pretty cool stuff.)

It's already starting to look like a car! Very Cool.

drmiller100 · 01-04-2012, 08:56 PM

Interesting how you equate wider track with increased roll stiffness.

Further interesting how you equate understeer stability characteristics with rear suspension design, and actually consider a lateral rear swing arm suspension a good thing. How would one go about building a lateral swing arm suspension, along the lines of old VW bugs for a mono rear tire? Most people considered the Corvair and early VW bug "unsafe", and Ralph Nader actually became famous for his opposition to it.

The traditional solution for straight line stability in rutted pavement involves attention to static camber, and keeping the front rollcenter lower then the rear.
As I am sure you are aware, it is relatively easy to get the front roll center below ground level, and the rear roll center of a trailing arm suspension, with the pivot perpendicular to vehicle travel, is by definition ground level. Therefore, it "feels" steady.

Another way to make a car feel stable is to make it narrow, and long. I think I have succeeded in this case.

Your recommendations on rollcages are just silly. Spectacularly silly. I hope you mistyped your recommendation for a 1.75 cage of .125 for a 1000 pound car.

Heck, I don't think NHRA requires more then 1.5" for a 2000 pound car into the 9's. SCCA says if I wanted to race it I'd be fine with 1.5 .095 even up to 2200 pounds.

As for weight, I certainly hope I can get the completed car lighter then 2500 pounds.

My goal is actually less than half that.

Ken Fry · 01-05-2012, 02:07 AM

Sorry Doug,

I didn't intend to confuse you so much.

I wrote:

"That means the front wheels must tilt at the same rate as the rear wheel as the car corners. The original Beetle geometry is good in this respect. "

You reacted to that with:

Quote:

"Further interesting how you equate understeer stability characteristics with rear suspension design, and actually consider a lateral rear swing arm suspension a good thing" ... and you continued ranting.

I would have thought it was stunningly obvious that we were talking about front suspension geometry here, because with a rear trailing arm, you don't have anything to work with at the back: the wheel must lean with the car's roll angle: as soon as the car goes into a turn, the rear wheel is disadvantageously positioned with the road. To avoid oversteer, the fronts need to also lean with the car, as provided by the front suspension of a VW, the front suspension of a 2CV, and the front suspension of a DS (most Citroens, in fact) ... and, of course, the Morgan three wheeler.

With the Civic front suspension, the geometry is designed to keep the outer wheel in a turn roughly perpendicular to the road, (so its top must lean into the body, called negative camber) to give it better traction than it would have if it simply leaned with the body. Likewise, the rear suspension on the Civic does the same thing. If instead, the Civic's rear suspension were straight trailing arms, then the front suspension would also need something very close to trailing arm geometry, to avoid oversteer.

Here is a pic of a 2cv cornering. You can see the wheels are far from perpendicular with the road, but are instead parallel to the cars rolled centerline plane.

From the Wikipedia article on camber:
"Camber angle alters the handling qualities of a particular suspension design; in particular, negative camber improves grip when cornering. This is because it places the tire at a better angle to the road, transmitting the forces through the vertical plane of the tire rather than through a shear force across it."

No other factor of suspension geometry has as large an effect on the cornering ability and cornering characteristic (oversteer, neutral, understeer) as camber management vs roll angle and the relation of front and rear camber. (Much of the rest has to do with feel and feedback, anti dive, etc. -- and if you screw that stuff up, at least you're not likely to swap ends while taking a bend.)

You rambled on about straight line stability, which I had not brought up. I am only concerned about safety aspects, and could care less if it wanders a little: there are plenty of factory fresh cars that wander a little, and almost everything from GM has crummy feel -- those are little things.

Roll cage issues:

You wrote:

Quote:

Your recommendations on rollcages are just silly. Spectacularly silly. I hope you mistyped your recommendation for a 1.75 cage of .125 for a 1000 pound car.

Your roll cage is not a roll cage, because the spans are far too long and you have no triangulation. To make up for the long spans you need larger diameter tubing. If your car were to actually weigh 1000 lbs this would be less of an issue, but it will not weigh 1000 lbs, even empty.

Your structure is far less efficient that the Civic's structure, but also longer, which makes it whippier yet. So to to keep it at Civic weight would be hard, without sacrificing safety and torsional and beam stiffness. The X Prize Aptera was 2300 lbs, and it was at least a partial monocoque.

These are the Auto Power Industries specs for a Civic roll cage: 1.75 x .120 DOM. Their cages are engineered and tested, however, so you would do well to overbuild rather than underbuild, because your structure is different. Here's a link that shows a Civic roll cage built of 1.75 inch tube: Roll Cages for the Honda Civic from HorsepowerFreaks

I did not recommend .125 wall. Nor was I recommending a roll cage for a 1000 lb car. I have no reason to believe that your car will be 1000 lb, even unloaded. I am also not recommending any roll cage dimension -- that requires engineering, which no one will do in casual conversation over the web. The safety cage for a passenger vehicle needs to be engineered. I am not saying that 1.75 inch tube is "enough." I am saying that anything less requires more extensive engineering and testing. At a minimum, you should do the testing for rollover strength spelled out in the Fed regs.

Your roll cage has no provision for accepting the load from the door guard beams, and even a 1.75" tube is probably inadequate for that, if it occupies the place where your existing downtube is. Read the regulations, apply the test loads.

Narrowness:
Read about rollover. This may be difficult to understand, but imagine moving the front wheels closer together, and eventually so close that they merge. You would then have no roll resistance at all. Now, as you move them apart again, the car does not go from having no roll stability at all, to having "plenty." One width is "way too narrow". Another is "pretty close", etc.

Think about Explorers flipping over, and why the online rollover threshold calculators ask for width.

I mention these things for your safety, and that of your wife. I see that all this falls on deaf ears, however. So I will trouble you no further.

drmiller100 · 01-05-2012, 12:08 PM

Ken, I encourage you to learn about roll centers, roll stiffness, and tipping. cars don't tip because they lean. they don't lean because they are narrow or wide.

they lean because of the torque created between the rollcenter and the center of gravity.

They tip because the torque created by the center of gravity about the outside tire.

Leaning and tipping are unrelated. If you make a car track wider, it will indeed be less likely to tip, but you can say nothing about how much it leans.

As for the rollbar, if I were to add 1.75 DOM I would indeed have a much heavier car.

G2TDI · 01-05-2012, 12:40 PM

I haven't read all the pages yet but I read the initial post and then some and this one came to mind see the poster's other videos too!
http://www.youtube.com/user/spiderbke/videos