beanbag

Nuts, I was hoping those were measured numbers, but at least our results agreed.

I'll add in the %'s.

beanbag

In the context of explaining how the program works- A blurb about chassis deflection/stiffness:

Even though there are many kinds of chassis flex and deflections, in the context of corner balancing, there is only one. If you have 4 points (shock towers) of which you can only push upwards against, then there is only one relevant flex, and that is that one diagonal bends up, and the other bends down. Hence if you include flex in the corner balance eq's in my first post, one diagonal will get +D's, and the other will get -D.
Also, the equation I mentioned in the second half of the first post changes to:

(dh + ds) / mr = +/-D (depending on which diagonal) since now weight transfers are changing the corner heights. Whichever corner gets more weight gets HIGHER. Yes, higher, because this corner is now pushing the chassis upwards "harder", while the other corners are "sagging".
Going thru the equations again as usual, there aren't any additional unknowns that would make them unsolvable, since the chassis flex is proportional to the weight transfer.

There are a couple of nice benefits to looking at chassis flex in this manner (as opposed to looking at each corner individually getting bent up or down), mainly having to do with ease of math and coming up with equations. One is that E4 is still an eigen operation. If one diagonal bends up, and the other bends down, does it affect the overall chassis height? Tilt? No. This is nice because it shows that chassis bending is a function of diagonal weight transfer ONLY. Like I mentioned earlier, all this eigen stuff and having "separate" operations is a nice foundation to work off of, because it makes it easy for you to add in other effects, like I did here for the chassis flex.

One other thing to note: As Descartes mentioned earlier regarding using the fender heights for corner balancing instead of the jacking points. This becomes more relevant if there is flex between the shock tower and the jacking point. I read some posts recently regarding "fender braces" and whether there is any flex at the firewall. If this is true then the program will give an error (although likely small) where you either will not get as much weight transfer as expected, or the expected corner heights will be off. I wouldn't worry about it, though. Close enough.

A blurb about the right side of the spread sheet:

Even though I call it the "stupid" way, it's only stupid in the sense of changing one corner at a time and trying to guess what happens everywhere else. The height and weight of any corner depends on the spring rate and MR of all the other corners combined. I guarantee you the equations are a mess.

Luckily, we already have a set of relatively simple equations from the left side of the spread sheet that tell you what the perch heights and weights should be given the corner heights. IOW, h1=blah c1+blah c2 +blah c3, etc. What we need to do is flip these equations around so we now have c1= blah h1 + blah h2 + etc. I guarantee you this will not be pleasant if you try to do it by hand. However, we all learned about systems of equations and matrices back in 10th grade. (Right?) Just write the equations in matrix form, and if you invert the matrix, you get the new set of equations. This is pretty nice because all the variables are actually numbers that you've already plugged in, so all Excel has to do is numerically invert a matrix, and not actually solve any equations (analytically).

That's what is going on in that box of "Random numbers and stuff". You'll notice that the matrix has a lot of 1's in it. That's coz, conveniently, there are a lot of factors (such as the factors F and R) that cancel out. I called them the "spring rate - chassis interaction" because their equation form looks a lot like the chassis stiffness in series with the spring rate. Fancy that.

Anyway, one of the uses for the right half of the program is if the left half tells you to change one of the spring heights by 0.05", for example, and you're too lazy to do that. So plug in 0 for that corner and see if the consequences are horrible. Or you only feel like mucking with one corner since the car is already pretty close to where you want it.

beanbag

Program has been uploaded here:

http://www.filedropper.com/cb4a

MD5 hash of the file is 9a1137d9db7e21ea7924b5612a0e65a7

I guarantee that this program will get your car exactly balanced in one shot, or you get double your money back.



Note: this is an updated version

descartesfool

Good work and original thinking.

essex

Good woork beanbag!

descartesfool

Speaking of corner balancing, one of the engineers Bryan I think from JRZ suspensions said on another forum that once a car is properly dialed in, they use rake or front to rear ride height adjustments for fine tuning. The theory here is that if you raise or lower the rear end say, while corner weights don't change, what does change is the kinematic roll center height (or forced based roll center height if you prefer). If you change the roll center height, you change the moment arm length between the center of gravity and the roll center, and thus change the balance between elastic (body roll via spring compression) weight transfer and non-elastic (no body roll but forces transmitted through the links) weight transfer. Thus a rake change would change the handling balance of the car without changing springs or bars. I have never tried this, and I was wondering if anyone had.

Here is what Bryan of JRZ said about tuning an AWD car when I asked about the effect of changing just the rear spring perch height with stiff rear springs:

"My best guess is that since the rear travels much less than it did as stock, there was some roll center (another semantic boondoggle) movement split front and rear built into the car that isn't going on anymore. Since the rear RC is more under control, setting ride height has a larger effect on the kinematic mechanical balance through the range of travel/roll. 1-3 rounds (1.5-4.5 mm at the damper) is quite effective. We can use 1/6 turns for very fine adjustments. " With Beanbag's nice little spreadsheet, rake adjustments are easy to do.

Of course this is only going to work if car is set at close to optimal balance with the proper springs, bars and damper settings. Perhaps on a FWD car the place to try rake adjustments is on the front axle rather than the rear, since at least both of those tires are always on the ground. Sounds like a good question for Mike Ortiz!

beanbag

I updated the program. Thanks to DF for finding an error in one of the cells. At his insistence, I also added the ability to measure at the jacking points. Adding this just ruined the elegance of all that "eigen" stuff involved in doing it "my way". Forget about using the back of an envelope to calculate your next corner balance operation. Bring a ream of paper and a big eraser. Better yet, bring a computer. At least I was able to salvage some pride by only using a 3x3 matrix inversion instead of 4x4.

I also added tire stiffness to the equation. It adds in series (with a motion ratio) to your spring rate, and really cuts down on the spring rate if you are using stiff springs. I have read that tire stiffness values are somewhere around 1000 lbs/in and higher. Anybody have a better estimate?

The chassis and tire stiffness factors don't have a big effect, but they can work to reduce the expected weight transfer. Since nobody measures chassis stiffness around here, just treat it as a fitting parameter, ok? Shift some weight, and see how much the chassis deflected. Or set it to a high value and forget about it.

dirty19

iTrader: (1)

Now i have a headache. Time for a beer!

Unbreakable6161

Good thread, goes hand and hand with what I was taught in statics.

SleeperGSR

Ok beanbag...how good are you really? Awesome thread although i have to say some of it is over my head still. I have a 96 integra gsr and it has been corner ballanced. I want to change the ride height by raising the back and lowering the front. Both about by 1/2 inch. Is there anyway i could do this without changing the corner weight proportion and keep it at 50%?

beanbag

Use the program. It will tell you what to do. Don't worry about inputting the correct corner weights and heights initially, since all changes are relative anyway.

You will end up raising the back adjusters and lowering the front, with equal left/right adjustments.

If I were you, I would adjust the heights myself and not even take it to the corner balancer. Assuming your car was properly balanced the first time, any corner weight error you get would be due to random errors from suspension stiction and settling.

SleeperGSR

This is what i was thinking, so if i lower both front's 1/2 inch and raise the back both 1/2 inch i should have the cross weight % should still be 50% right?

beanbag

Depends on your faith in the E2 (front back tilt) operator.

rotten

God damn it, you guys have sucked all the 'fun' out of corner balancing.

descartesfool

A nice article on chassis twist:

http://www.theoryinpracticeengineeri...ing%20fsae.pdf

SleeperGSR

not exactly sure what the E2 operator is/does. So i guess i have faith in it? ha

beanbag

A valuable life lesson indeed

SleeperGSR

i guess i will keep reading

beanbag

Thanks. Good to get some actual numbers.

Mattyg2

iTrader: (1)

this thread is confusing lol, i think i understand it and how i can use it towards my car.

Ultraspeed DC2

6 year bump.

Does anyone have a copy of the file?

Has anyone tried it?

Beanbag check PM