[4banger]

Originally Posted by MGizzle

Springs hold the SPRUNG mass, hence damping is important. Swaybars are directly conected to the sprung & unsprung masses and move via a given motion ratio. They don't seperate the sprung & unsprung mass like springs do. When you look at a quarter of half model of a car you don't see swaybars in the equation because they don't support the sprung mass.

I see what you are saying. If you wanted to model the sway effect, I think you would have to use a three mass system. One mass being the sprung weight, one mass the tire and wheel, and the last mass being the rest of the unsprung weight. That is a lot more detail than the basic DHO examples that I've seen.

Originally Posted by MGizzle

A great test that we could do on our cars would be to buy some cheap string potentiometers and bolt them to our frame & the ground. Push on the roof pilars until you get the car going side to side quiet a bit and than let go. The string poteniometers will tell you how long it took for the car to settle and it will give you the damping curve.

You would be finding the damping rate rate of the chassis, right? I'm not sure I understand why you would want to move the car from side to side.

I'd also like to hear your thoughts on the use of the damped frequency. Most of what I've been able to read only talks about the natural frequency which doesn't seem to have much use considering it doesn't take in to effect the dampers, and I always drive around with my dampers on my car. When choosing spring rates, do you think it would be better to choose them based off the undamped frequency? It's awesome to finally be able to talk to someone who has practical and theoretical knowledge.

[ace996]

Originally Posted by stretch

I deliberately didn't make a conclusion to prevent certain individuals from skipping to it, reading it as a blanket statement, and then derailing this thread by arguing about it out of context.

My friend, I don't skip...I just haven't been around the past few days...last day of the season today at Limerock. I needed to video and datalogg the new 600/500 springs...I even kept my stock tires on so I could direct compare from my previous events there.
Needless to say it was a great day and there's much to discuss...but not here, I'm not looking to de-rail anything... but I am glad to see some thought going on about the sways, but am disappointed at the approach.
I believe you need to consider the sways as individual corners, not together, as that's what will upset the chassis...

Be good,
TomK

[ace996]

Originally Posted by 4banger

I don't know why I didn't think of that earlier when Ace asked about adding in the swaybar rate. I'm not going to put words in anyone's mouth, but I think the thought there was that the sway and spring would just add together, which would result in a much lower %cd. I think what really is happening is that spring rate would go way down. I think we need to nail this down first before we continue on.

4b,
Good thinking....and I agree with you. They're another form of spring, and should be considered....you'll need more rebound with bars than with none.
Be good,
TomK

[BigSky]

4Banger- I'll post up my setup and impressions from your recommendation on ~ 65% critical damped

I'm in a 04 WRX wagon ~ 3000 lbs w/o driver, SA Koni's w/ STi Tarmac springs (324/276), by my calculations my frequency is front 1.9/rear 2.1 (ignoring sway bars which are 27mm front/22mm rear)

It was suggested I soften up my damping to full soft as this would yield close to 65% damped for these spring rates. I ran two autoxs at full soft (I was running previously at just shy of full firm f/r). The car handled better overall Where I felt it was lacking was in fast transitions. I've since dialed them up 1/3 turn from full soft (my Koni's have just shy of 1.5 full turns), and like this setting better, but this is still much softer than I was previosuly running. Clearly the Koni's are adequately damped to run much firmer springs than I'm running, ideally I'd like to get my car to ~ 2.2/2.5 (450/400)- maybe next season

[MGizzle]

Alright, see attached image of the system I drew up.



I think that this way we can see exactly what is hapening. From this you can see that the two UNSPRUNG masses have a spring in between them, which is the ARB (a rotational spring). Hence, you don't need to dampen it if you are worried about the sprung mass.

However, this also shows us that if we want to move a SINGLE corner up, our overall spring rate will INCREASE (wheel rate, ride rate will go up), hence why ARB's can affect ride comfort. In this case, the damper is dampening the overall spring system, which includes the ARB. However, remeber that even in the single bump case the other side of the car will cause some resistance since the ARB wants to go in the opposite direction, hence you are leveling the sprung mass between the two unsprung masses even in a single corner bump case. Again, the ARB is thinking that your car is rolling and it is doing exactly what it has to, which is PREVENT ROLL, hence even in the single bump case it causes less roll of your sprung mass due to the increased effective spring rate.

Now, I don't have a clear mathematical explanation on why the ARB shouldn't be included in your damping calculations, probably should be, but from research everybody states that the effect of the ARB is so small in this case that it can be disregarded in the damper calculations. I am going to try and confirm that and I placed a few calls with some Tier 1 supplier friends.

So, your ARB will cause the body to "level" even in a single bump case and the main frequency is going to be created by the two corner springs, which are being damped.

As far as why would I shake my car side to side...well..that is what you are trying to prevent with dampening. If I didn't have dampers on my STi I could do this until I reach the natural frequency and theoreticaly (without friction, air resistance etc.) my car would wabble infinetely since it doesn't have any damping. However, if I introduce dampers, the car will level to equilibrium and depending on what my dampening factor is, it will determine how fast it will do this. Please keep in mind (lol) that I don't care about comfort and that I car about the "setting time" or "transient time" response. I want my car to reach steady state AS FAST as possible, because this is where I am pulling max G's, be it in lateral or longitudinal directions. Hence, if I measure how long it took my car to set with the string potentiometers, I could go through dampers settings and see which one would settle it the fastest. I could do the same in the lateral direction. Also, this way I have everything included in the system (ARB spring rate, springs, friction, damper friction, bushing friction, tire dampening, tire spring rate, etc. etc. etc.).

Think of it as a cheap seven post rig hehehehehe. Instead of placing forces on the unsprung masses I excite the sprung mass to get something similar.

Anyhow, I will see what I get back from my friends (of which one worked on the suspension for the new C6RS Z06, yeah the black Jay Leno MONSTER!!!) so I think I can get some good stuff from them.

Talk to you guys later.

[4banger]

Originally Posted by BigSky

4Banger- I'll post up my setup and impressions from your recommendation on ~ 65% critical damped

Thanks for the long term feedback. I'm glad you like the new settings.

[4banger]

Originally Posted by MGizzle

Alright, see attached image of the system I drew up.



I think that this way we can see exactly what is hapening. From this you can see that the two UNSPRUNG masses have a spring in between them, which is the ARB (a rotational spring). Hence, you don't need to dampen it if you are worried about the sprung mass.

Thanks for the feedback, but I think that you really have to include the tire if you want to include the swaybar. If in your picture, the Mu1,2 also had another spring/masss/damper attached to them for the tire spring/damping rate, the overall rate of the spring acting on Mu1,2 from the non-coilover side would be found by adding the swaybar spring rate with the tire spring rate. Since the tire spring rate is so much larger than the waybar rate, it effectively masks the swaybar rate. I have no idea if this is right - what do you think?

I see what you're saying about shaking the car side to side now. I was so fixated on the frequency of the car during bump that I forgot that there are also frequencies for the other modes of car movement. Very interesting idea.

When you talk to your friends, maybe you can ask them about thier thoughs on damped vs undamped frequency. All of the info that one can find on the internet talks about choosing springrates based on the natural (undamped frequency). It seems like using the damped frequency would be the more logical choices since it better shows what your car is really doing.

Here is an interesting observation for everyone. If you look at the equation for damped frequncy, you will notice that the answer is undefined if you are over 100% critically damped. When your car hits a bump, it doesn't oscillate any more, it just slowly moves from one steady state to the next.

[BigSky]

here's an interesting snippet

evolutionm.net - View Single Post - ultimate suspension engineering

"here is a logarithmic decaying relationship between damping ratio and natural frequency"

[stretch]

Originally Posted by 4banger

Here is an interesting observation for everyone. If you look at the equation for damped frequncy, you will notice that the answer is undefined if you are over 100% critically damped. When your car hits a bump, it doesn't oscillate any more, it just slowly moves from one steady state to the next.

That would seem to happen noticeably on the stock struts once the grease wears away. The increased friction essentially creates high damping at the lowest (and still) damper velocities.

Looking at your damping charts, I bet it would also happen noticeably on Teins due to the almost immediately high damping forces. I don't think you'd get this on Ohlins, even though they're over 100% critically damped, because those damping forces drop as piston velocities slow.

That's an interesting thought though, and I think is a good way to explain why overdamping is bad.

Originally Posted by BigSky

here's an interesting snippet

evolutionm.net - View Single Post - ultimate suspension engineering

"here is a logarithmic decaying relationship between damping ratio and natural frequency"

Interesting, good find! I'd bet he's referring more to high-speed damping than low-speed. Perhaps that's where OptimumG's recommendations come from for a damping coefficient of 0.3 to 0.5 for high-speed. 0.3 sounded very low, but maybe that's appropriate for really stiffly sprung cars to get some compliance out of them.

[4banger]

Great find, Bigsky.

Originally Posted by stretch

Looking at your damping charts, I bet it would also happen noticeably on Teins due to the almost immediately high damping forces. I don't think you'd get this on Ohlins, even though they're over 100% critically damped, because those damping forces drop as piston velocities slow.

The Whitelines, Ohlins and the stock dampers all show a slight increase in %cd in the ~5 in/second region. Maybe it is so that the initial response of a large bump doesn't upset the suspension so much because of the high damping force, but the later part of the reaction is more controled by the suspension and allows for a smoother and compliant response.

Originally Posted by stretch

Interesting, good find! I'd bet he's referring more to high-speed damping than low-speed. Perhaps that's where OptimumG's recommendations come from for a damping coefficient of 0.3 to 0.5 for high-speed. 0.3 sounded very low, but maybe that's appropriate for really stiffly sprung cars to get some compliance out of them.

I see what you are saying here, and it makes perfect sense. The higher frequency allows for a quick response by the suspension, which makes up for the fact that the suspension has to oscillate a little more to completely dampen out a bump. This way you get a compiant ride which quickly responds to disruptions in ride.

[MGizzle]

Well I just heard from a few friends in the industry and they all agree that the ARB is just an addition to the spring rate, as in the small drawing I posted. I guess depending on how much the ARB contributes to the overall spring rate will change how much damping you need.

Short answer as far as damping: you need to consider the addition of your ARB in damping if it is a significant contributer.

Now as far as the damped frequency question. Why would you care about this as long as you know you are damping your natural frequency as fast as possible?? You want your transient time to be as fast as possible until steady state is reached hence your damped frequency is directed by this.

Even after all this I am still confused as to what spring/damper set up would be great for our STi. Talk to you guys later.

[ace996]

Originally Posted by MGizzle

Well I just heard from a few friends in the industry and they all agree that the ARB is just an addition to the spring rate, as in the small drawing I posted. I guess depending on how much the ARB contributes to the overall spring rate will change how much damping you need.

Short answer as far as damping: you need to consider the addition of your ARB in damping if it is a significant contributer.

Thanks for the info. Good to hear that the ARB needs to be considered, and it makes perfect sense....
Go around a corner, car is leaning, a bump goes under the front outside tire -
You then have a compressed spring, a compressed ARB, and a requirement to damp the release of that energy on rebound.
Here is where the ARB' rate, which in the above example is well beyond 'insignificant', must be considered if one is lookng to be at a certain CD level. Run the numbers without the ARB and a set-up that comes out to 'perfect' will, in reality, be underdamped. Once you add in the ARB, the CD result will be lower....so a previously assumed 'overdamped' strut may just be 'dead on the money'.
Yeah...what I said a few pages back...

Be good,
TomK

[mofoSTI]

info overload is a good thing

[4banger]

MGizzle, thanks for checking with your friends.

I'm not going to post up new graphs quite yet because I have run into a few problems.

While doing more research, I came across an SAE (Society of Automotive Engineers) article on race car (and they specifically state that American autocross applies) design. It does not include ARB in the calculations of spring or damper rates. This article calculates ARB rate after the damping rate has been determined based on the additional stiffness needed to combat roll.

The real problem is this - the spring rate that they used is 571 lbf/in. The estimate that I was using from stretch's spreadsheet was around 5000 lbf/in. Stretch's info is based just on tire sizes and pressures and not materials or a specific tire. I did some more searching and on one engineering website I found a recommendation of 800 - 1430 lbf/in for "automotive tires." Obviously this is a huge disparity and unless someone has a scale that reads up to 1000 lbs and a vice big enough to put a tire/wheel in we're not going to be able to determine the exact number to use. The only manufacturer that posts their spring rate data is Hoosier, which is located here: Hoosier Tire. We all know that the stock tires are pretty stiff, so I am thinking that we should go with something close to the 1430 lbf/in number, but I would like to hear everyone's thoughts. EDIT: I found another estimate for a street car of 2285 lbf/in.

The second problem that I have run into is much, much less significant. The tire problem has a huge effect on the results where this one does not. I found this thread on Nasioc with calculation of unsprung weights: Anyone curious on what the usnprung weight of a WRX is? - Brakes, Steering & Suspension - NASIOC. Thanks for finding all these numbers Bigsky. The numbers in this thread are very close to what stretch and I have been using, and are also almost exactly the same as a WRX. I understand that their is some weight savings on the STi's aluminum parts, but the tires and rotors are heavier by a good bit, and I would assume that the hubs and axles are beefier as well. I also think that the entire weight of the strut/spring assembly should be used since our MR is so close to 1.

[4banger]

I also wanted to give a few reasons why we should be cautions in adding in the ARB rate:

-Not considering the ARB rate is perfectly fine for 2 wheel bump.
-When considering the ARB rate you have to descide if you want to analyze 1 wheel bump or body roll.
-If you want to analyze body roll, you now have to account for the change the mass of your given corner. You would really need to recalculate the weight for each corner of the car. Not that hard, but someting to consider. Also, for different "g" turns, the weight on each corner will change. A single graph would not tell you much.
-In the body roll case, the ARB adds to one sides spring rate and subtracts from the other sides spring rate (I think, someone please back me up on this). On the stock vehicle, the sway bar itself has a higher spring rate than the spring and would yeild a negative number in the calculation for Kwheel. I'll have to do some more thinking on what that means.

Just some things to keep in mind.