Torsional Stiffness of 1988 911 coupe

[burgermeister]

around 2700 ft*lb / degree.

Edit: 5300 ft*lb / degree after further corrections (detailed later in the thread)



I decided to measure rather than hum-haw and wonder. The pictures pretty much speak for themselves - basically, jackstands under the rear jack supports (directly under the torsion bars). A beam across the front A-arm mounts, balanced on an edge. Apply load to one side of front beam and measure deflection.

As the vertical loads from the suspension are carried at these locations, they seemed like the most sensible to me.

[David]

Will all the engineers please stand up



Looks like someone has too much free time



Are you preparing to build a rollcage perhaps?

[Cooper911SC]

Different car but still relevant...

Here's the way a friend of mines Co. did a twist test of a Subaru STI. The CarLab manufactures a product for the STI called the X-brace...improves torsional stiffness by a whopping 12%!

The picture kind of tells the story here too.

But to clarify what might not be obvious.

Front and rear springs and shocks are removed from control arms and car. Solid tube the length of the shock when the car is sitting at regular ride height are bolted in.

The rear of the car is bolted to the ground using the wheel studs and concrete anchors.

The front is supported by jack stands.

A long steel 'I' beam is bolted to one of the front suspension members. The 'I' beam is marked at 'X' distance from the car and weight in 100LB plates is added to the beam. The beams sag is measured at 'X' distance and the calc. is done to figure out rigidity.

Kind of cool to figure this stuff out.

It is something manufacturers DO NOT like to divulge to the public or competitors. Chassis engineers get paid big bucks to build these products.

Thanks for posting the 911 info.

Cooper

[Cooper911SC]

Here's the STI graph
Cooper

[RoninLB]

I think this is a great topic.


I eyeball the sag on 3 floor jacks and so far so good with it being a stiff straight body. [I do my own wheel alignments.]


any educated opinions on the 3 point sag distance? In general... Even a wild guess?

another one is how much of a road slam into the tub from the wheels before it twists?

stiff suspension track tubs understably twist out of shape but if i slid off the road into dirt and start slamming around. I was told the front strut brace would help stop the front wheels from folding under.

my wild guess is that if the wheel didn't fold it can't be that bad?

jacked at the rear tub and fr tub behind wheel.

[Bullet Bob]

2700 ft*lb/deg is veeery bad for a sports car :-( The WRX values look halfway decent though.

[burgermeister]

Yes, it's quite the noodle. I have it from a reputable source that modern day full size pickup truck frames are in the 2000+ realm (just the frame by itself). SUV frames are stiffer yet. Bolt a cab & box on it, and the pickup trucks are stiffer than our 911's, and at significantly longer wheelbases to boot.

Though, it is a 1965 design, and it was probably quite stiff for its time.

Ronin, I measured linear stiffness. Permanent set caused by overload is an entirely different subject, and the two do not necessarily correlate. You could figure linear corner sag by taking the front axle weight (lb) * track (ft) / 2.0 / 2700. That gives you degrees of sag. divide by 57.3 to get radians, multiply by track width, and you have predicted sag.

125shifter, no rollcage for me - street use only. (Currently it appears to be garage use only...). I'm just a curious geek, and while I had the back on jackstands, I thought I'd spend the hour (though it took twice as long as the planned hour ... ) to measure it. Might come in handy some day.

[livi]

Just an ignorant thought:

Does torsional stiffness and overall vehicle weight relate to each other in terms of performance? That is, will generally a lighter car require less torsional stiffness than a heavier vehicle?

Just speculating, as our vintage 911 are pretty light.

[Peter Bull]

Interesting work!

There are a couple of sources of compliance in your setup which will reduce the measured torsional stiffness some. Possibly a lot.

The pad between the wooden block and car body at the rear; is it rubber?

Since you are working with relatively small deformations, the elastic compliance in the wooden blocks will affect the measurements too.

Are you sure there's contact between the right front support and the cross member when you start applying load?

There is also some compliance in the axle stands, small compared to the rubber and wood, but it will affect the result.

Measuring torsional stiffness is difficult, and your setup looks good. But it could be worthwhile to measure how much the car body moves relative to the supports, and maybe use steel rather than rubber and wood in the supports.

/Peter

[hcoles]

it brings to mind... when I see pictures of 911's with one front wheel in the air... at that point how much torque is in the chassis? 2700* .4 * 2.5 = 2700 ft-lbs = ~ 1 degree

[hcoles]

also brings to mind when I back out of my driveway at one point diagonal wheels are up compared to the other diagonal set... and the front window squeeks. Anyone else notice this?

[450knotOffice]

Yup. Even my 996 windshield squeeks in its frame going up my driveway entrance at an angle.

[911pcars]

Here's a jig on an open wheeled car:



I'll also add this link that connects this discussion to another on structural rigidity:

Structural Reinforcement

Sherwood

[Bullet Bob]

The torsional stiffness of the chassis has a great effect on the ability to change the handling balance (from oversteer to understeer) using the anti roll bars or spring changes. Even though these cars are light they do require a great deal of stiffness to perform well. The car shown inthe CAD picture might weight ~450 lbs and the chassis stiffness design target would be ~3000 ft*lb/deg. This is with a very low CG as well.

A well designed and fabricated Porsche race car would probably be in the 13k-18k ft-lbs/deg range

[Bullet Bob]

Here is one that would be really nice to know. What is the CG height of these cars?

[burgermeister]

The lighter the car, the lower the required spring rates and the lower the roll moment, so the less stiff it can be. Obviously the 911 has at least decent, and certainly fun, handling with its stiffness. I'd be curious what the stiffness is for a Miata, Boxster, Solstice, or other convertible sports car with reputedly excellent handling?

(Edited - swapped Peter Bull for Bullet Bob)
Peter Bull, yes, there are many sources of error in my setup, starting with the bathroom scale and ending with the lack of dial indicators. The pictures are there so everyone can see them. If you care to donate 3 or 4 more dial indicators and mounts to my toolbox, next time I have the car on stands I'll get a more accurate measurement . As it stands, my budget is small and my time is limited - you're always welcome to try it yourself if you'd like a more accurate answer.

There is no rubber between the wood blocks and the car. And I will not put my car on jackstands without wood blocks...

My car creaks quite a bit when doing driveway ramp events. I can feel the door move in its frame if I touch the roof and door at the same time.

According to SAE 840561, CG height in the 80's varied from 437 to 579mm. I'd guess the 911 to be in the 450 - 500 range, with stock being close to 500, and everyone's lowered-to-the-weeds "euro height" being closer to 450. I have NO basis for this statement - it's a guess.

[Jay Laifman]

I've always wondered about this. It seems every year R&T is saying how great it is that some new BMW, Porsche or other car is 30% stiffer than the last year. Really, year after year. How stiff can these cars get!? And, really, for sporting fun driving on the road, my 35 year old 911 doesn't seem to be waffling around. Go drive a Morgan! (yet they seem to do pretty good on the track). So does this stiffness really have ANY impact on a car that is not raced? (of course I realize that question sends one down a path of questioning pretty much all the features of our beloved cars).

[hcoles]

Originally Posted by Jay Laifman

I've always wondered about this. It seems every year R&T is saying how great it is that some new BMW, Porsche or other car is 30% stiffer than the last year. Really, year after year. How stiff can these cars get!? And, really, for sporting fun driving on the road, my 35 year old 911 doesn't seem to be waffling around. Go drive a Morgan! (yet they seem to do pretty good on the track). So does this stiffness really have ANY impact on a car that is not raced? (of course I realize that question sends one down a path of questioning pretty much all the features of our beloved cars).

I'm thinking the same thing. Probably after you get above 3000 ft-lbs/degree per ton or some formula like that it really isn't noticed on a street/DE car. The Corvette over the years comes to mind... I think in the last C-5 and C-6 they finally got the chassis stiffened up. My memory is that the Corvette marketing people made a pretty big deal about it at the time.

[Bullet Bob]

Burgermeister, I didn't have any issues with your setup, that was Peter Bull. The slope of the graph looks pretty linear so I would say the setup worked well.

[Bullet Bob]

The stiffness demands of the chassis are dictated by the roll stiffness of the suspension. Basically the chassis needs to be stiff enough so that it is not acting like a suspension spring itself. I don't think it really matters on a street car unless you really want to balance the handling via tuning sway bars etc...

With a loose chassis stiffening the front or rear suspension will only cause the chassis to twist more and it will not alter the load transfer characteristics during cornering.