Strut tower bar questions
Strut tower bar questions
Preface:
Was discussing this with an auto-x buddy who drives a CSP miata. He was saying that few of the prepped miatas run (rear) strut bars, but that bars at both ends really improved his old STi. He also mentioned that there are comparatively few rear strut bars available for the miatas. Anyway, our talk turned to the theoretical impact of a bar.
Main discussion:
Assuming steady state cornering,
Red is friction (most of the cornering force coming from the outside wheel). This acts of the wheel which acts on the suspension arm (blue). Now, I don't think I'm working things right with the strut bar. I have it as a moment (wrt the arm) on the top of the strut, both from the momentum of the chassis and from the force being transmitted from the wheel. This is the part that is confusing me. It seems to me that the bar should be in tension and that implies that the outside wheel should be pulling the inside wheel outwards, right? Yet, I always was told the strut bar prevents the towers from squeezing in on each other, which implies the bar is in compression. I can convince myself either one is right, HELP!
Either way, the it seems that the idea is bar preventing the towers from shifting relative to each other. If this is the case, why are so many of them designed in a manner that suggests buckling (or a more obvious spring action)? My cusco bar is arced towards the front of the car, yet is more of a "-" cross section that in the end probably makes it bend upwards since the forward arc isn't severe. Why aren't more bars straight with a circular or "+" shaped x-sections if the idea is to prevent the bar from buckling at all (aka making it have an extremely high spring rate and essentially preventing motion between the ends)? Is there some sort of black art to this (obviously minor) tuning? I figure it has a alot to do with how rigidly the strut mounts are attached to the rest of the body, and probably has much less effect on a SLA suspension, since the shock is very lightly loaded due to effect of the upper arm (hence miatas not using them). Still, with a little work, could you mildly affect the roll or transfer rate by extremely preloading it (vs say a control of no bar)? I'm sure this doesn't have nearly the effect of say a sway bar, but is it really something that should just be blindly added? I'm hoping some of the track guys might have some data on this, or even better, have hooked up a strain gauge at one point. I'm probably over thinking this, but still curious.
Was discussing this with an auto-x buddy who drives a CSP miata. He was saying that few of the prepped miatas run (rear) strut bars, but that bars at both ends really improved his old STi. He also mentioned that there are comparatively few rear strut bars available for the miatas. Anyway, our talk turned to the theoretical impact of a bar.
Main discussion:
Assuming steady state cornering,
Red is friction (most of the cornering force coming from the outside wheel). This acts of the wheel which acts on the suspension arm (blue). Now, I don't think I'm working things right with the strut bar. I have it as a moment (wrt the arm) on the top of the strut, both from the momentum of the chassis and from the force being transmitted from the wheel. This is the part that is confusing me. It seems to me that the bar should be in tension and that implies that the outside wheel should be pulling the inside wheel outwards, right? Yet, I always was told the strut bar prevents the towers from squeezing in on each other, which implies the bar is in compression. I can convince myself either one is right, HELP!
Either way, the it seems that the idea is bar preventing the towers from shifting relative to each other. If this is the case, why are so many of them designed in a manner that suggests buckling (or a more obvious spring action)? My cusco bar is arced towards the front of the car, yet is more of a "-" cross section that in the end probably makes it bend upwards since the forward arc isn't severe. Why aren't more bars straight with a circular or "+" shaped x-sections if the idea is to prevent the bar from buckling at all (aka making it have an extremely high spring rate and essentially preventing motion between the ends)? Is there some sort of black art to this (obviously minor) tuning? I figure it has a alot to do with how rigidly the strut mounts are attached to the rest of the body, and probably has much less effect on a SLA suspension, since the shock is very lightly loaded due to effect of the upper arm (hence miatas not using them). Still, with a little work, could you mildly affect the roll or transfer rate by extremely preloading it (vs say a control of no bar)? I'm sure this doesn't have nearly the effect of say a sway bar, but is it really something that should just be blindly added? I'm hoping some of the track guys might have some data on this, or even better, have hooked up a strain gauge at one point. I'm probably over thinking this, but still curious.
I don't have strain gauge readings, but I can tell you from suspension programs that I have run, that there is an incredible amount of weight transfer to the outside front wheel. This would result in a force that would tend to move the strut tower inward.
In a 1700# racecar with equal corner weights, 425# per corner, the weight transfer in a 1.3 g load was nearly 800#s on the outside wheel. This is affected by roll center and center of gravity, so will vary depending on your set up. All this being said, any time you can stiffen the chassis, it will help by reducing uncontrolled movement or flexing. The suspension should be doing the work, not the chassis.
In a 1700# racecar with equal corner weights, 425# per corner, the weight transfer in a 1.3 g load was nearly 800#s on the outside wheel. This is affected by roll center and center of gravity, so will vary depending on your set up. All this being said, any time you can stiffen the chassis, it will help by reducing uncontrolled movement or flexing. The suspension should be doing the work, not the chassis.
Well since Miatas don't have struts they are just trying to put a brace across big holes inthe chassis.
As for a strut bar, its doing the same thing. The strut towers try to move in every direction depending on what the car is doing. All you need to worry about is keeping them from moving if you can. You're on the right track with your design thoughts. Usually rules or space get in the way of making a better design.
If your chassis is about 4-5 times stiffer than your suspension you're fine. That can be figured with some weights and a long 2x8 hung off of the strut towers like a lever if you really want to get into it deep.
As for a strut bar, its doing the same thing. The strut towers try to move in every direction depending on what the car is doing. All you need to worry about is keeping them from moving if you can. You're on the right track with your design thoughts. Usually rules or space get in the way of making a better design.
If your chassis is about 4-5 times stiffer than your suspension you're fine. That can be figured with some weights and a long 2x8 hung off of the strut towers like a lever if you really want to get into it deep.
77333
Joined: Aug 2006
Posts: 221
Likes: 0
From: Burien, WA
There's an E30 guy that did some similar analysis and came to the conclusion that in cornering conditions on a smooth track a shock/strut tie-bar would be in tension:
http://www.e30m3project.com/e30m3per...bar_theory.htm
I'm no engineer so I can't confirm or refute the findings, but it does seem to make sense to my vaguely mechanical mind.
http://www.e30m3project.com/e30m3per...bar_theory.htm
I'm no engineer so I can't confirm or refute the findings, but it does seem to make sense to my vaguely mechanical mind.
That is basically what I was trying to get at from a theory standpoint. I'm glad I'm not the only one with that conclusion. I need to find some way to log some strain gauges.
If this is the case, why are so many of them designed in a manner that suggests buckling (or a more obvious spring action)? My cusco bar is arced towards the front of the car, yet is more of a "-" cross section that in the end probably makes it bend upwards since the forward arc isn't severe. Why aren't more bars straight with a circular or "+" shaped x-sections if the idea is to prevent the bar from buckling at all (aka making it have an extremely high spring rate and essentially preventing motion between the ends)?
Though, some might say that something is better than nothing, and these bars do have a slight value in chassis rigidity. Their construction is more designed for easier production and lower tolerances, as well as ease of installation and packaging contraints (putting on and taking off my Mazdatrix one-piece bar is a PAIN. And it sits very close to my alternator)
Now, since these cheaper bars do absorb some of the energy like a spring, they can serve to lower the vehicles suspension frequency, which would result in lower NVH and a smoother chassis. To that effect, Subaru puts a rear strut bar on some J-spec STi editions that has a damper built in (I would have to search to find pics, but I have seen it as a result of another discussion of the same topic). Why they chose to do that instead of having a rigid bar is a bit perplexing, but in the previous discussion we wrote it off to a "rally driving" issue and left it there.
-Outside wheel sees a largish outward force at the top of the shock tower.
-Inside wheel sees a medium inward force at the top of the shock tower.
Because this means that one shock tower tries to flex more than the other, on net a strut tower bar goes into tension in a corner. This is why turnbuckle style preloading is a good idea (and why pivots at the attachement point of a bar don't matter.)
Note that this whole commentary really only applies to FCs due to the use of MacPherson struts. If you look at how forces are carried, it should be obvious why strut tower bars are of limited value on FDs.... There are no significant lateral forces at the top of the shock tower because of the use of dual A-arms. This is also why pillow ball uppers are a great idea on an FC, but just create unnecessary harshness with no upside in an FD.
