Jump to content


LMP3 torsional stiffness

  • Please log in to reply
6 replies to this topic

#1 Monstrobolaxa

  • Member

  • 187 posts
  • Joined: April 04

Posted 03 April 2008 - 10:08

Well I'm finishing a preliminary LMP3 chassis and will FEA test it in a few days!

Now does anyone know or have information on the torsional stiffness of any LMP3/Group CN cars? And if the info is from a space frame chassis it wouls be great! I need the info so I can set a baseline value from where I can start optimizing.

Earlier spaceframe chassis in F1 where around 1500 Nm/deg...acording to what I've found!


#2 zac510

  • Member

  • 1,713 posts
  • Joined: January 04

Posted 04 April 2008 - 09:11

Hey Monstro, there is a reference here that says the Juno SS1 has a stiffness of around 22000Nm/deg.

It's a pretty competitive market full of private manufacturers so I wouldn't be surprised if they didn't talk much.

Maybe you could email Ewan and ask him what ballpark he is in? :) You'd think the other manufacturers would be around a similar figure.

#3 Lukin

  • Member

  • 1,953 posts
  • Joined: January 03

Posted 04 April 2008 - 10:06

I remember reading in a roadcar mag that one of the new Saabs at the time has a stiffness around 21-23,000 Nm/Deg. Once you weld in a roll cage you could probably double that.

Peter Wrights book about the F2001 says 7000 Nm/deg axle to axle; the tub was higher but once the gearbox/engine assembly was installed the stiffness dropped a fair bit.

Though looking at that link above, 22000 could be a little optimistic?

I think you also need to look at how the figures were gained. I've seen people measure torsional stiffness two methods:

1. Constrain the rear axle and apply a force upwards at the LF and the same force downwards at the RF. This seems like the best way for measuring pure torsional stiffness.

2. Constrain both rears and one of the front, applying a load to the other front. To me this is a combination of bending and torisonal load. Milliken says if you have sufficient torisonal stiffness then you shouldn't have a problem with a lack of bending stifffness; whether this is the case is another matter.

Both are probably a good measure of chassis construction/design though I would imagine measuring via method 2 would give a much higher value than method 1!

If your not comparing apples to apples your fairly well urinating in the wind if your using other people's figures as a guide. Deakin, Crolla, Ramirez and Hanley did a paper about finding the point of limited returns for chassis stiffness based on the overall roll stiffness required in the car.


This is another thread (maybe two) on chassis stiffness. One of them I started back when I was young and stupid (as opposed to middle aged and still stupid) where Greg has some good ideas on working through chassis stiffness optimisation.

#4 zac510

  • Member

  • 1,713 posts
  • Joined: January 04

Posted 04 April 2008 - 10:16

Lukin, the figure is probably just for the spaceframe..

#5 Greg Locock

Greg Locock
  • Member

  • 4,907 posts
  • Joined: March 03

Posted 06 April 2008 - 03:04

All numbers off the top of my head, Nm/deg or ft lb/deg

1955 Tbird 2700 - chassis only, body would add a little

1960 F1 1500 reputedly

1978 911 Targa top 3000

1988 Lotus Esprit 5000 (ish)

1990 Lotus Esprit 10000

FSAE car at least 1500

Typical 1995 large sedan 15000-25000

some 2007 BMW 40000 (claimed)

I concur with Lukin, deciding how you measure it will affect your results, to get good correlation I have been forced to model the rig as well as the chassis. Glass matters a lot for production cars.

Here's some more results for the Tbird


x axis is distance along the chassis, y axis is the vertical deflection at various points on the chassis, measured in the same place on the FEA and the real test.

#6 GeorgeTheCar

  • Member

  • 376 posts
  • Joined: September 03

Posted 06 April 2008 - 17:39

My recollections is that in 1967 a good F1 car was 6000 and a Can Am car a little better

#7 mariner

  • Member

  • 1,542 posts
  • Joined: January 07

Posted 09 April 2008 - 20:10

One point worth mentioning is "how old is the chassis you are measuring?"

On road cars thee have been some scary demonstrations of loss of stiffness due to rust and cracked welds. It was mostly aimed at crash worthiness ( hence the scary bit) but the same logically applies to torsional stiffness.

With racing cars I think the old rivetted chassis deteriorated quite quickly as the rivets fretted. It is hard to stop that happening. Carbon should be better but if you tie suspensions to gearboxes to engines to chassis all through mechanical fasteners the quality of the detail design of mating surfaces and how many dowels etc. you use could affect torsional stiffness over time.

So I would be interested to know whether all the quoted data was from brand new cars and if anybody ever runs the same test as part of planned maintainence as the chassi builds up milage.

The torsional test data I find most intersting are the graphs plotted at stations along the chassis length, I know it is much arder to do but it gives good clues as to where to improve things.