Roll Resistance

[El Guapo]

Bob,
Seems to me that with (wheel rate) = (corner weight), and also assuming the CG mass centroid is horizontal (the same height both front and rear), then the UNIT roll resistance:

(Ft-Lbs of roll resistance per degree) / (total car weight at that end)

would be 50/50 split. This is without any bars. However I think most cars have a higher CG in the rear, so that would effectively soften the rear roll resistance relative to the front. Most cars seem to run stiffer roll resistance in front, exactly why is beyond me, but I think it has something to do with more rear traction under acceleration. Right or wrong??

[El Guapo]

I don't want to shoot this thread off on a tangent away from Bob's question, but last year I tried calculating roll resistance not from the simple standpoint of outside wheel rate, but also considering the roll-assist effects of the inside wheel. I got bogged down and dropped it, it was more complex than I first thought, or I am looking at the problem wrong. Has anyone tackled that monster? How is it done?

[Mike Ahrens]

We have all read the posts regarding spring rates as various percentages of corner weights. While playing with various set-ups on paper I got to wondering about roll resistance distribution, as this directly effects where the weight that is transfered goes. If my numbers are correct, with wheel rates = to corner weights on my Z10c and my roll bars set to mid position, I get a 67%F 33%R roll stifness distribution. Does this sound in the ballpark? One big assumption I had to make was the center of gravity height being ~13.5". (I know how to measure it, I just don't have a set of scales) Any insight would be appreciated.

[Mike Ahrens]

[quote]Seems to me that with (wheel rate) = (corner weight), and also assuming the CG mass centroid is horizontal (the same height both front and rear), then the UNIT roll resistance:

(Ft-Lbs of roll resistance per degree) / (total car weight at that end)

would be 50/50 split. This is without any bars. However I think most cars have a higher CG in the rear, so that would effectively soften the rear roll resistance relative to the front. Most cars seem to run stiffer roll resistance in front, exactly why is beyond me, but I think it has something to do with more rear traction under acceleration. Right or wrong?? [/quote]Dale, I'm using the formulas from Staniforth's book [i]Competition Car Suspension[/i]. The formulas are pretty straight forward, if difficult to follow. He does not take into account any droop limiting or preloading.

As for the Cg for the entire car, it is a point so it would not have a slope. If you were looking at individual slices of the car, computed their Cg's and then connected the dots, yes you would get a choppy line that sloped from back to front. As a matter of fact Staniforth dicusses this technique and abandons it in favor of using the Cg of the entire car.

I zeroed out the anti-roll bars and calculated a roll stiffness distribution aproximately equal to the weight distribution, i.e. 41%F/59%R.

With the roll bars the roll stiffness can be varied from
59%F/41%R to 45%F/55%R
The difference between the two settings, in terms of weight transfer, is 25#'s moving from the outside front to the outside rear tire.


Is this a large amount? I don't know. Are there any chassis gurus out there?

[Jeff Smith]

bob,
You really need to attend one of ICP's seminars. Because of discussions just like post, most attendees are still waiting for their head to stop throbbing. Also Steve has developed (maybe borrowed) a program that calculates exactly what you are looking for. It even goes so far as to need the chassis stiffness for a more complete number. Sorry but I am not going to even attempt to answer your question. You also need to figure in the spring rate for the tire as well.

Richard is at a seminar in Indy this weekend, I believe, he would be considered the Guru, especially seeing as they build citations.

It must be a real cold winter....when you guys gonna start racing and quit tinkering?, Geesh our off season is approx 6 weeks and damn near gone...lol

John

[Mike Ahrens]

Hi John, As far as the seminar goes, been there done that. Although it was about two years ago so I'm sure the course has changed some.

I'm working on taking the tire rate into account, except Hoosier doesn't provide a lot of data and its too damn cold in the garage to go experimenting or work on the car hence the cerebral exercises. They did publish some info on their site for the SAE guys but they did not include the R60 compound used by CFF in CenDiv. Between that info and some other data I've found I can probly get pretty close to the tyre spring rate.

I'll get around to measuring the torsional stiffness of the chasis one of these days also. I'll have to figure out how to model that next; two spring systems connected by a torsion bar??? Maybe Richard will see this and offer an opinion.

[Jeff Smith]

Bob,
I also attended a seminar last year. Based on the difference in the material and the way it was presented, I am planning on attending every year. Cheap price to pay for the all trial and error that the seminar eliminates. Atleast I know which direction to try.

John

[scott 71]

Bob: I was waiting on a guru to respond as well. But since none did, I'd like to compare notes with you. I did some rough calculations a few years ago. First this is the basic formula I used for roll stiffness: ((wheel rate) * (track) * (track)) / 2. I believe there are a few assumptions in there and also a conversion factor is needed to convert to (ft-lbf)/deg. On a '79 VD with wheel rates numerically equal to corner weights, and no roll bar, the roll stiffness was also close to 40F/60R. What surprised me was the change your roll bars made to your roll stiffness. On mine, the stock bars can only change it by +/- 5%. I'm wondering if my calculation for the rollbar rate at the wheel is correct. I plan on trying to measure the roll bar rate on the car. The plan is to use a setup similar to that used to measure the torsional rigidity of the chassis. After going to the ICP seminar I overheard Steve Lathrop talking about using that setup to measure the rollbar rate. BTW, the roll resistance of the chassis using the ICP method was around 1500 (ft-lbf)/deg.

[Mike Ahrens]

Steve,
I'll have to go back and doublecheck my calculations just to make sure. I was also suppried by the amount of the change as the bars on Z10's are generally regarded as ineffective tuning aids.

After doing some digging on the various websites, a fellow by the name of Ted James (Seems to have "Guru" status according to homage paid to him by Leo Stohr)suggested 55% of the roll stiffness be on the front. Said this was applicable to CSR, DSR and FA. Not sure how this would apply to FF's but it may be a good starting point.

I a going to measure my chassis stiffness as soon as the garage is warm enough to work in it. I believe modern FF's are around 6000 ft-lb/degree.

[Curtis Boggs]

The amount of front roll stiffness from bars and springs will vary according to the height of the roll center - the higher it is, the greater the jacking effect and the lower the needed roll resistance for any given degree of roll desired.

The 55%-65% numbers are in the ball park, depending again on roll center height. On the Z10, the front ride height can be lowered down to around 1.25" using 1000# springs. The bar should be a minimum of 11/16ths, and it may eventually prove necessary to go bigger at that ride height - don't know yet as we have yet to get good feedback from a decent driver with that sort of spring rate.

The complaint about the Z10 bars being ineffective is typical of the engineering level of the car owners of the time - they kept softening the front bar in hopes of getting rid of the push, when the problem was elsewhere.

[Mike Ahrens]

Thanks Richard.
BTW I give up, where did the problem lie with the push in the Z10?

[Curtis Boggs]

Some cars were set up with anti-dive. If you have two mounting holes for the front lower trailing links at the inboard end ( you should), make sure that you use the lower one.

The car also had a very high front roll center originally at the 2 1/2" ride height, making for a lot of jacking effect. If you go to the stiff springs and lower the RH down to 1 1/8" - 1.25", the roll center will go to about 5/8" below ground, a very GOOD place for it. You CANNOT try to run it up around 1 1/2 - 1 3/4" as it will pass thru ground level at times, making for a real fun ride.

You also need the full ackerman kit as the original arms had zero ackerman.

[Mike Ahrens]

Thanks Richard, thats good info. Got the Ackerman and I am using zero anti-dive. I'm going to have to recheck my calcs on the roll center. I had figured it much lower at that ride height so I was inclide to run the car higher. I tried real low, 3/4" and did not like the handling. It felt like the front was jacking itself down in the corners.

[Curtis Boggs]

Which suspension version do you have, A, B, or C? I'm not sure which one Steve ran the analysis on, but I'm betting it was the C version.

If you were trying to run that low, you'd probably need a lot stiffer springs to keep it from pounding into the ground, but most likely would lose grip as the shocks can't properly handle such rates with so low a motion ratio.

[Mike Ahrens]

I have a "C". My front springs were only 500# at the time. I plan to follow your advice and go to wheel rates = corner weights