GRC (roll centre)

[JohnL]

I think you're right, new thread:

55EXX wrote:
“i see my understanding as being the rc is the point at which the car rotates around”

This is the first and most common misconception about geometric roll centres (GRC, of which there are two, one located in a vertical plane defined by the contact patches at either the front or rear axle lines).

Despite their name, GRCs are NOT the points about which sprung mass ‘rolls’, except in very specific circumstances. These circumstances are:

1) At the infinitely small moment at which weight transfer begins to occur and before the springs / dampers / ARB have begun to deflect.

2) When the GRC and the CG are located at the same height (though in this instance no roll motion will occur).

3) When the springs etc provide zero roll stiffness (e.g. some *Formula Vee ‘zero roll stiffness’ rear suspensions where the springs only have bump stiffness and contribute zero roll stiffness - see footnote).

A GRC is a theoretical point defined by suspension geometry that is involved in dictating what % of the weight that is transferring at that axle line is transferring either geometrically (through the suspension linkages) or ‘elastically’ (via the springs, dampers and ARB). Spring rate etc is also involved in dictating the ratios of geometric and elastic WT, so GRCs aren’t the sole determinant of the %s of transfer occurring through both vectors (i.e. geometric and elastic).

The relative %s of WT occurring geometrically and elastically vary constantly during any corner, and the amount transferring through either vector does matter because WT that occurs geometrically is ‘instant’ and WT that occurs elastically is ‘slow’.

Geometric weight transfer occurs at the rate of the lateral force (acceleration), i.e. if the lateral acceleration increases by say 10% in one second then the geometric WT also increases by 10% in one second. However, whatever % of WT is occurring elastically will occur more slowly, so if the lateral acceleration increases by 10% in one second then the % of WT that is occurring elastically may take say 2 seconds. This may seem trivial, but it has huge implications for the transient handling characteristics of the car, and the car’s reaction to bumps etc.

To make things more interesting, as lateral acceleration increases the % ratio of geometric vs elastic WT will change, with the geometric vector tending to become a lesser % of the total WT and the elastic vector becoming the dominant vector of WT.

I’m only scratching the surface here, this is only a hint of how complex this stuff is, it can do your head in. Even when you have an understanding of some facets of roll theory, it’s hugely difficult to describe it to other people.

“and the cg is the mass centre and determines the length of the moment arm around the rc. the further the cg is away from the rc the longer the moment arm and therefore the greater roll force produce from the same cornering g forces.”

More or less.

“so by creating a environment that the rc comes closer to the cg as beneficial for increasing roll stiffness geometrically by shortening that moment arm other than mechanically by the use of sways and springs rates etc. this would then mean with softer springs and sways a softer ride can be enjoyed and i guess the wheels would follow the undulations in the road better but what other advantage is there? like i said before increased front roll resistance can be increased by bigger sways, and which with soft springs can increase front roll resistance and follow undulations in the road.”

Suspension design is rarely that simple. Advantages of high GRCs are very few, and the downsides are substantial and more numerous. It does tend to be advantageous to bring the GRC and the CG closer by means of lowering the CG, but not by raising the GRC (at least not very often and only in specific instances for specific reasons).

The higher the GRC the greater the % of WT that will occur geometrically and thus the less responsive the car will tend to be to changes in suspension rates, so the handling becomes difficult to tune with spring / damper / ARB rates etc. The lower the GRC the more WT will occur elastically and the car will be more responsive to associated changes to springs etc. This is especially important for racing cars, but also important for ‘performance’ road cars.

Higher GRC is also strongly associated with lateral scrub of the contact patch (with independent suspensions), i.e. as the wheel rises it also moves laterally outward, so when you pass over a bump the chassis tends to get ‘pushed’ laterally (inward when cornering and the outer wheel hits a bump). Force goes both ways (ask Isaac), so this also means that the chassis mass / inertia can ‘push’ abruptly and hard on the contact patch, potentially causing it to lose grip suddenly at / near the cornering limit. Lower GRC results in less lateral scrub with vertical wheel motion.

Then there are the problems with abrupt WT caused by the GRC being high, wheel jacking effects etc…

An instance where changing the GRC height is used as a tuning aid is in V8 Supercars, where the Watt’s Linkage (lateral locating system for the rear axle) can be raised / lowered by means of a lever in the cabin, which raises / lowers the rear GRC. The drivers use this as a means to fine tune turn-in as grip conditions change (yes, rear GRC change affecting front grip at turn in…), but this is also associated with the mandatory use of locked differentials (i.e. I doubt it would be so needed or effective if a more ‘normal’ differential were used).

-------------------
*Formula Vee cars are forced to use the VW swing axle rear suspension that suffers from the GRC being way too high and thus the rear geometric roll stiffness being way too high. The last thing they want is more roll stiffness from the springs, so the spring (only one rear spring with this type of suspension) bears (via rods and cranks) against the suspension on each side but not against the chassis, i.e. the left suspension bears through the spring against the right suspension, thus giving bump rate but not roll rate.

Note; roll theory is very important for most cars, but tends to become less important the less suspension motion there is and the more the car relies on down-force for grip and handling balance etc. Modern F1 cars have really terrible suspension geometry (front suspension most particularly is truly awful) because the down-force is so great and the suspension motion very small (to keep the under-tray at a stable distance from the track), and the down-force generated grip totally overwhelm any geometric affects. The aero considerations are far more important than the suspension geometry, so the suspension members tend to get placed wherever they will interfere least with the air flow, not with all that much regard for optimum suspension geometry.

It’s interesting to know that our cars have substantially better suspension geometry than current F1 cars…

[nd55]

ughh! This is more than I read the whole of 7th grade....

Nick.

[string]

Engineering is the art of compromise. I think that's all that needs to be said in regard to the above comparisons between modern F1 and 90's Honda suspension.

[JohnL]

ughh! This is more than I read the whole of 7th grade....

Nick.

That's more or less my point. That rave just scratches the surface and doesn't really investigate the physics in any meaningful way, and it gets harder and harder to understand the deeper you go into it.

It's not so much that each aspect of weight transfer and roll motion is hard to understand, but there are always a number of simple things going on at once and it's the interactions between these things that gets complex.

At the level of the amatuer enthusiast it's probably better to try to work with understanding the affects of changes to the elastic vector of WT (ie. springs, dampers, ARBs etc), rather than the geometric vector. There's not all that much we can do with the geometric vector in any case, other than to understand that while the roll geometry created by the designer may not be perfect, it's probably a lot better than we can manage ourselves, and that when we mess with it we will more than likely make it worse...

[Sp00ny]

Nice Writeup John.

Before we know it, you'll be have a sticky up about "Flux Capacitors" haha xD

It would be interesting to hear your view on the roll centre adjusters from aftermarket manufacturers and there advantages(if any)/disadvantages.

[JohnL]

When you substantially lower the ride height, the front wishbones become angled upward from the chassis to the ball joints. This changes the static location of the 'instant centres' of the 'virtual swing arms' and thus changes the static location of the geometeric roll centre (lowering the GRC). This isn't necessarily a bad thing in and of itself, and in fact a somewhat lower static GRC location may actually be a good thing (especially at the front axle line), though it would be advisable to keep it from going below ground level.

However, when the wishbone angles are changed substantially from stock, there is also an increased tendency for the dynamic locations of the VSA ICs to change location with suspension motion (as occurs in roll) in a manner that tends to cause the dynamic location of the GRC to change more than with stock wishbone angles (maybe much more).

This can cause erratic (or at least unintended and adverse) changes in the distribution of dynamic weight transfer in transient cornering, which could make for 'difficult' handling characteristics that could be impossible to sort using spring rates , dampers etc etc.

If 'roll centre adjusters' are used then this reduces the angle of the lower wishbone (lowers the ball joint relative to the chassis mounts), so in some degree will negate some of the adverse geometry change associated with a substantial lowering of ride height.

So yes, I think 'roll centre adjusters' would tend to be a good thing, but since they only affect the lower wishbone they only partially address the problem. A better option (but more expensive) would be to use uprights / knuckles (whatever you want to call them) that have both the lower and upper ball joints located lower relative to the axle axis.

This means we can lower the ride height (i.e. lower the CG!!, which is always a good thing if it can be done without making a hash out of the roll geometry) without excessively affecting either the lower or upper wishbone angles, thus keeping the roll geometry more or less stock, which is more or less a very good thing (most probably, in most cases).

Changing the as designed roll geometry may or may not be a bad thing, but chances are that you won't accidentally improve it, and probably would make it much worse unless you really understand what you are doing...

[Sp00ny]

When you substantially lower the ride height, the front wishbones become angled upward from the chassis to the ball joints. This changes the static location of the 'instant centres' of the 'virtual swing arms' and thus changes the static location of the geometeric roll centre (lowering the GRC). This isn't necessarily a bad thing in and of itself, and in fact a somewhat lower static GRC location may actually be a good thing (especially at the front axle line), though it would be advisable to keep it from going below ground level.

However, when the wishbone angles are changed substantially from stock, there is also an increased tendency for the dynamic locations of the VSA ICs to change location with suspension motion (as occurs in roll) in a manner that tends to cause the dynamic location of the GRC to change more than with stock wishbone angles (maybe much more).

This can cause erratic (or at least unintended and adverse) changes in the distribution of dynamic weight transfer in transient cornering, which could make for 'difficult' handling characteristics that could be impossible to sort using spring rates , dampers etc etc.

If 'roll centre adjusters' are used then this reduces the angle of the lower wishbone (lowers the ball joint relative to the chassis mounts), so in some degree will negate some of the adverse geometry change associated with a substantial lowering of ride height.

So yes, I think 'roll centre adjusters' would tend to be a good thing, but since they only affect the lower wishbone they only partially address the problem. A better option (but more expensive) would be to use uprights / knuckles (whatever you want to call them) that have both the lower and upper ball joints located lower relative to the axle axis.

This means we can lower the ride height (i.e. lower the CG!!, which is always a good thing if it can be done without making a hash out of the roll geometry) without excessively affecting either the lower or upper wishbone angles, thus keeping the roll geometry more or less stock, which is more or less a very good thing (most probably, in most cases).

Changing the as designed roll geometry may or may not be a bad thing, but chances are that you won't accidentally improve it, and probably would make it much worse unless you really understand what you are doing...

Ah, you see this write up in technical terms should be basically changing the idea of "lowest handles best" impression, within reason of course.

All in all, I don't think 99% of people on this forum including myself would have enough knowledge to improve upon the stock GRC or even improve upon it once the car's ride height is dramatically changed.

Most Tyre/Alignment places wouldn't have a clue either. Expessially the K-Mart tyre and auto type places.

If we get a RC Adjuster then where would you suggest to get it "fine tuned" and adjusted at? A suspension specialist?

[nd55]

> If we get a RC Adjuster then where would you suggest to get it "fine tuned" and adjusted at? A suspension specialist?

?????

The ones I've seen aren't adjustable. You either bolt them on or you don't.


Nick.

[JohnL]

^ That's my understanding.

Lowering the CG is always a good thing (in itself) because it reduces weight transfer, so more of the car's weight is distributed over more area of rubber when cornering (i.e. it increases grip).

The problem is that substantially lowering the CG nearly always creates substantial concomitant geometry changes which can create problems that may outweigh the positives of the lower CG, which then have to be dealt with properly in order to realise the full benefit of the lower CG (or just to make the car handle as well as it did before, let alone better).

A moderate lowering isn't likely to make a pig's ear out of the roll geometry (though some suspensions are bound to be more sensitive than others), and would usually be a worthwhile modification (as is commonly found). The trouble arises when people start thinking that if if a little bit is good then a whole lot must be much better, which is only the case if you know where the pitfalls might be and how to deal with them.

Suspension isn't simple, and shouldn't be modified without due care and consideration because it is primarily important to the car's safety. It should never be ignorantly changed just for bling appeal...

[55EXX]

so in affect the lesson is if you want better handling a bit lower not dumped is the way to go. like the whiteline flat out and control springs. the flatout spring are dumped and whiteline doesnt recommend them for best handling whereas the control springs have been designed with the suspension geometry in mind to create the best compromise for better handling tho they aren't as low as the flatout ones.

i understood all that you said. one line that stands out to me.


roll theory is very important for most cars, but tends to become less important the less suspension motion there is

that would be me in the bracket of the less suspension motion the less important. rc adjusters would ony be a thing i would look into if i were to seriously dump my car. if i were a track guru and done everything and wanted to squeeze out everylast ms then i would look further into it. yeah i'm sweet to stop here with roll theory while i'm on top. i don't wanna catch the bug and want to know it all then end up more confused. quit while i'm ahead.

thanks john L wish i didn't need to spread rep points around.

[EfiOz]

Nice write up John on a complex subject. I had to write an article similar to this for a mag and it was one of the hardest subjects to explain in lay mans terms I've come across.

A couple of addons.

We don't let drivers touch the RRC in V8 Supercar. They get front and rear ARB adjusters and they stuff that up enough as it is. The rear roll centre is used to induce a jacking moment to unload the inside rear so the car will turn in. This coupled to a fairly low FRC gives the car a mild "dog-at-the-lamp-post" attitude.

We used to use spacers on the lower ball joint on Dallara F3 cars to alter the FRC. Quick and easy!

[nd55]

> We used to use spacers on the lower ball joint on Dallara F3 cars to alter the FRC.

more info, please?

What spacers, what knuckles etc
What was the criteria to raise and lower the RC?
That's a highly stressed point. DId they ever break?
What material?

Nick.

[JohnL]

Nice write up John on a complex subject.

Thanks, I tried at least...


I had to write an article similar to this for a mag and it was one of the hardest subjects to explain in lay mans terms I've come across.

In any terms it's gobsmackingly difficult to understand weight transfer and roll, and even harder to explain. I don't pretend to have a definitive understanding, but I have given the subject quite a lot of thought. IMO anybody who claims they understand weight transfer thoroughly is probably full of 'it', mistaken, or quite highly paid...

I started thinking about this when I was trying to understand weight transfer with karts. If you're masochistic you can find pages and pages of my ramblings on this on ekartingnews.com (internal search for 'John Learmonth' and 'roll', but be warned, it's not all completely baked.... ).

I got the bug to try and figure this out after I found that the commonly available information on roll centres (etc) just didn't add up. I couldn't make heads nor tails of it until I figured out that the GRC couldn't possibly be the point around which the sprung mass actually 'rolls', that is unless it was the only source of roll stiffness at that axle line (e.g. aforementioned FV rear ends, or tractor pivot beams (where the pivot is by default also the GRC), or at the instantaneous moment before any WT has actually begun...).

Once I realised this it all made a lot more sense, but became a lot more complicated (and this is just in 2 dimensions at a single axle line, when in reality we're dealing with 4 dimensions at two axle lines (4th dimension being time of course)).

To understand the nature of WT we need to understand elastic WT (through 'elastic' vectors such as springs / dampers et al), as well as geometric WT (through 'rigid' vectors of suspension members according to their geometry), but, we can't really understand either vector without understanding the other vector, i.e. to understand X we first need to understand Y and Z, but to understand Y we need to understand Z, but to understand Z we need to understand Y....

It's hard, and it doesn't take long before I've just about clogged my brain's working RAM and it gets even harder to think clearly...


A couple of addons.

We don't let drivers touch the RRC in V8 Supercar. They get front and rear ARB adjusters and they stuff that up enough as it is.

Fair enough. My source is obviously somewhat incorrect, but then I shouldn't completely believe everything I read, even on eng-tips.com...


The rear roll centre is used to induce a jacking moment to unload the inside rear so the car will turn in. This coupled to a fairly low FRC gives the car a mild "dog-at-the-lamp-post" attitude.

So, with the locked diff a V8 Supercar (I hate that name...) is in effect an oversized kart that by dint of it's longer wheelbase to track width ratio doesn't need to unload the IR quite so hard as a real kart? Do V8SCs also use SR and caster to jack the IR? I hear a lot of caster is used, is it all for steered camber compensation purposes?

IMO it's the relative GRC height difference front / rear that affects turn-in. Anecdotally it seems the typical result of raising the rear GRC / lowering the front GRC is to sharpen up turn in, which fits with my understanding of what this ought to do.

My thinking suggests this is because the lower front GRC / higher rear GRC creates an initial lateral front weight transfer that is not only less but also slower and a rear weight transfer that is not only greater but also faster, which gives the front end more grip and 'leverage' to change direction more sharply at turn-in and early entry (this has a lot to do with effective 'rubber on the road' and the manner in which initial lateral forces are distributed into / resisted by the four sidewalls of the front tyre pairs relative to the rear tyre pairs, I think...).

However, as lateral acceleration (and WT) increases the influence of the geometric roll stiffness rapidly falls away (less so the closer the GRC is to the CG, in which case GRS is always a larger % of total axle roll stiffness) and the elastic roll stiffness becomes far more influential.

I.e., in the very early stages of the cornering process before much weight transfer has occurred, at both the front and the rear, GRS is the dominant vector of WT. This is because weight transfers 'instantly' via the geometric vector (and 'slowly' by the elastic vector). As time passes (quickly), more and more weight is increasingly transferring via the elastic vector, and the amount of weight transferring geometrically becomes a lesser and lesser % of the total WT until we reach a steady state (at which point elastic weight transfer 'catches up' and both vectors become stable values).

This is why (in my understanding) it's possible to tune turn-in / early entry with RC height changes without having a substantial affect on mid corner / exit, and to tune mid corner / exit you'd be more likely to alter relative front / rear elastic RS (springs / dampers / ARBs / maybe even tyre pressures (i.e. tyre 'spring rate')).

Of course this is just theoretical speculation on my part. I've left out pretty much all of the detail that might explain why I think / suspect the things that I do (about the nature of weight transfer / roll), but once you start writing any of the detail down it becomes hard to stop, and hard to keep going, on and on and on...

With roll theory, just about anything you say begs another question, it becomes impossible to find a place to stop, you either run out of steam or run out of time, or start making less and less sense...


We used to use spacers on the lower ball joint on Dallara F3 cars to alter the FRC. Quick and easy!

The joys of not being saddled with tapered joints, the joys of having an F3 to play with ...

[JohnL]

What spacers, what knuckles etc

My understanding (fwiw) is that F3 cars (and many other types of specialised racing car) will have a lower 'ball joint' that is similar to a spherical rod end (aka 'Rose' joint), probably a 'caged spherical bush' fitted into the outer end of the wishbone. This spherical bush would then fit onto a stud protruding from the bottom of the 'upright' (aka 'knuckle').

Placing spacers between the bush and the bottom of the upright will change the angle of the lower wishbone and thus change the location of the 'virtual swing arm' instant centre, and thus change the static location of the GRC (and most probably change the manner in which the GRC moves in roll). This will also affect the camber curve (and static camber), and for the unwary, probably affect bump steer in some degree...


What was the criteria to raise and lower the RC?

My guess, to tweak turn in...?

[EfiOz]

No, it never broke. We would use a maximum of 20mm spacer.

It was mostly used as a function of turn in ability and track surface. At somewhere like PI you'd use it as the surface is kind to tyres even when throwing large instantaneous loads through them whereas Oran Park you'd go as low as you could so the tyre was loading in a more gentle fashion as it has a very abrasive surface that damages tyres that are loaded to quickly/heavily.

[55EXX]

this is slowly sinking in so this explains why by lifting the height in the rear of my eg hence its rc turn in is improved due to the geometric quicker response in WT via geometric means. so a kits like the front rc adjusters would decrease turn in due to its raising of the rc?

so also to by having a more heavily braced rear would that also increase turn in?

[EfiOz]

JohnL, you're pretty much on the money with the cornering sequence. You also have to factor in a tyre which is ludicrously fragile for what it has to do and really needs to be treated gently to get anything out of it.

The canary on the coal mine for RRC height is when you start to lose drive coming out of corners, i.e. excessive wheelspin. With a FRC that deep, it's almost impossible to make them too "pointy".

[JohnL]

this is slowly sinking in so this explains why by lifting the height in the rear of my eg hence its rc turn in is improved due to the geometric quicker response in WT via geometric means.

This is what my understanding suggests should theoretically occur (when raising rear GRC and / or lowering the front GRC), and is also what I subjectively found to actually happen when I raised the rear of my Accord, i.e. somewhat improved turn-in response, leading (either directly or indirectly) to noticeably less understeer in early corner up to apex.


so a kits like the front rc adjusters would decrease turn in due to its raising of the rc?

In theory, yes, but I don’t think this is the most important thing here. Of more importance (in all probability) than the static GRC location is the dynamic GRC migration (GRC movement as the chassis rolls), which is very likely to become greater and potentially more problematic as the chassis is substantially lowered.

If this GRC migration is large and / or abrupt and / or in an undesirable direction then it’s likely to have at least some noticeable affect (and possibly a substantial affect) on the linearity of front vs rear weight transfer during transient cornering (i.e. what % of total weight transfer is at any given moment occurring at the front relative to the rear).

If the ratio of front vs rear weight transfer changes in a non progressive manner (due to substantial GRC location changes in roll), and the change is strong enough, then the driver may find transitional handling to be ‘difficult’, with the car generating fluctuating levels of understeer, or even understeer and oversteer in more extreme cases.

This is unlikely to be an issue in steady state cornering as the car will have ‘taken a set’ and the front / rear GRCs will have stopped moving.


so also to by having a more heavily braced rear would that also increase turn in?

And handling response in general. Chassis can never be too rigid, though not all of the various aftermarket stiffening members actually do much (some are little more than bling and extra weight).

The rear of most chassis tend not to be as stiff as they ought to be. With sedans (my CB7 being a good example), there is usually an unbraced aperture behind the rear seat back to allow long cargo to be passed into the passenger compartment (an artefact of the marketing department, I expect at the protest of the engineering dept…).

Unfortunately this is a high load area of the chassis, important for structural stiffness as it braces the rest of the structure (or should do). There are sheet metal braces at each side of this aperture, but really they are pathetic. Properly bracing this aperture makes a big difference to chassis stiffness, as does fitting a robust rear tower brace. I have both on my car, i.e. rear tower brace and a home made ‘X’ style seat aperture brace, and both are very worthwhile additions (I can definitely feel the difference when I’ve have to remove them to carry long cargo).

Hatches are even worse in this respect, with not even a bonded in rear screen to improve stiffness (the rear door isn’t attached rigidly enough to effectively brace anything). Hatchback chassis resemble a box with the rear panel removed, and need bracing for adequate stiffness for ‘sporty’ handling (note Type R usage of OE rear tower brace (so I’m told)). My mate’s EG Civic hatch flexes so much you can actually see the bottom of the rear door moving side to side at the car traverses bumps and undulations. I’ve offered to brace it up for him, but he doesn’t want to sacrifice cargo space…

[55EXX]

stuff cargo space. never had it before with my past obsessions with car audio. i will be fabricating a x brace between a c pillar and tower brace. i will be also making a fibreglass sub enclosure where my spare is. never had the joy of those either. good old tyre weld.

any other places to brace? what would be most advantageous to brace? what about where the rear window strut attaches to instead of the seat belt mount? i could fabricate a mount to that alot easier (being an exposed simple bolt) and instead of buying the c pillar bar for the parts needed to do the seatbelt mount and make a bar to join both of those and x down to the strut towers and ebay rear strut brace (brace is a straight pole so just as good as any in my mind). the rear window strut mount is also higher so a longer pseudo moment arm/fulcrum (don't know the term needed here) and further back from passengers heads (saftey marginally improved)

[55EXX]

may need a new thread if you want. just like this grc one.

[JohnL]

stuff cargo space. never had it before with my past obsessions with car audio.

That can be a rather heavy obsession, as in kilos, which affects acceleration, braking, handling, and grip....

I've heard of cars that needed stiffer springs just because of the added weight of the stereo gear. You do know that very loud doof doof makes you deaf?


any other places to brace? what would be most advantageous to brace? what about where the rear window strut attaches to instead of the seat belt mount? i could fabricate a mount to that alot easier (being an exposed simple bolt) and instead of buying the c pillar bar for the parts needed to do the seatbelt mount and make a bar to join both of those and x down to the strut towers and ebay rear strut brace (brace is a straight pole so just as good as any in my mind). the rear window strut mount is also higher so a longer pseudo moment arm/fulcrum (don't know the term needed here) and further back from passengers heads (saftey marginally improved)

Some thoughts:

With braces, bends in the tubes are bad, they substantially detract from the stiffness of the brace. If bends are required then the OD and wall thickness of the tube ought to be substantially increased because the brace needs beam strength if it's bent (and only compressive / tensile strength if straight). When designing braces always try to allow forces to pass along straight paths in straight tubes / bars, and keep in mind that tri-angles are very strong. Keep in mind that tower braces can only transfer significant forces in compression and tension from tower top to tower top, they do not act as 'girders' and cannot not significantly resist vertical forces, only lateral ones (they do inhibit vertical forces being 're-directed' into lateral motion).

In a hatchback, I suspect the legal requirement will be to keep any added internal bars etc below the top of the rear seat back, or substantially rearward of it and not obscuring rear vision in any way(?).

The thing that needs to be kept in mind is that flex won't exist in all areas of the body, and adding braces where no flex occurs only adds weight to the car. It's torsional flexure between the front and rear axle lines (i.e. chassis twist) under roll loads that are important (i.e detrimental to chassis dynamics), other flexures only really matter insofar as they may contribute to torsional flexure (with the exception of flexures that might affect suspension geometry as below last paragraph).

There are ways to check localised flexure using mock braces made from telescoping plastic tubes with a sliding O ring on the smaller OD tube and butted up against the larger tube. In compression the O ring will act as a 'tell tale', moving along the smaller tube as the two tubes move relative to each other, but in tension it needs to be observed as the chassis is loaded. Movements as small as 1mm can be significant.

Well known weak spots on most unibody chassis are between the front and rear towers, and tri-angulating at least one of the towers (or the tower brace itself) to a robust part of the firewall would be a good thing. An 'X' style brace directly behind the back of the rear seat would be good, as would an X brace that braces to points in the boot floor (these are commercially available, from somewhere...).

Braces (straight bars) that attach to the very rear (inside) of the rear boot panel or below the bottom of the boot lid or hatch door probably do very little as they are a long way from where the loads are imparted into the chassis and where the existing sheet metal braces adequately in any case (forces at this point would tend to be in panel 'shear' in any case, not compression or tension which are the only kind of loads a bar or tube can effectively carry without flexing in bend, unless the bar / tube is very substantial in OD).

Bars that run along the rear subframe probably don't do much, the subframe is already very rigid in compression / tension and a bar here won't be able to add much to it's beam strength. Beam strength in the rear subframe is a good thing because to a limited degree it substitutes for the structure above the subframe being weak (i.e. lack of bracing behind the rear seat back etc), but X beaces and rear tower braces would have more affect.

This is also why front subframes tend to be so beefy, i.e. a lot of the front chassis strength relies upon the brute beam strength of the subframe because above the subframe is so weak (i.e. unbraced).

I have no experience using any of the various in cabin braces (floor bars etc etc), but I doubt any but the most intrusive actually do much. In a hatch the most effective place to laterally brace inside the cabin would be between the rear towers, but a brace between the belt anchors wouldn't hurt (especially if also tri-angulated to the opposite tower top, but then this resembles an X brace, which I've already suggested).

The most important thing is to 'close' the 'open' end of the 'box' that forms the passenger compartment, and this really means tri-angulating the rear of the 'box' that effectively ends at the tower braces.

Sunroofs aren't a good thing for chassis stiffness. To stiffen the cabin section the best things you could do would be to fit a fully tri-angulated cage, weld the doors shut, or seam weld the chassis (least effective of these possibilties), but these things might be a little hard core for most people... At any rate, due to it's very large 'diameter' the cabin section itself isn't the biggest problem if the rear of the box is closed off.

If the front suspension has trailing radius rods attached to a forward subframe, then braces connecting the forward subframe to the 'rear' subframe would I think probably be worthwhile (i.e. not the subframe at the rear of the car, but at the rear of the engine bay to which the LCAs, steering rack etc are attached). This would lessen chassis flex under hard braking that would cause caster angle and wheelbase to lessen as the contact patches 'pull' the subframe backward.

[nd55]

> as would an X brace that braces to points in the boot floor (these are commercially available, from somewhere..

PasswordJDm for one. FBI in New Zealand is a distributor.

> Keep in mind that tower braces can only transfer significant forces in compression and tension

Not too sure about struts being effective in compression.
Especially for the fronts, on our cars.

I think mostly how they work is by closing the big 'U' shape that is the front of a modern passenger vehicle.


> That can be a rather heavy obsession, as in kilos

It's Xmas time, weight is not allowed to be discussed.


> You do know that very loud doof doof makes you deaf?

he, he. John, no thread jacking, this started as a thread on bump steer....

[JohnL]

Not too sure about struts being effective in compression.
Especially for the fronts, on our cars.

Your statement is unclear...


I think mostly how they work is by closing the big 'U' shape that is the front of a modern passenger vehicle.

Yes, the tower brace does that, meaning that with a good brace each of the 'U' uprights (towers) isn't independantly laterally floppy, they re-inforce each other through the brace. Note that with a 100% efficient brace we've only doubled the lateral stiffness (as seen at each tower), so only halved any lateral deflection (this is probably somewhat simplistic, but I think will be ball park).

This is why it's a good thing if we can find a means to tri-angulate to the firewall. It would also be nice if we could tri-angulate the towers to the bottom corners of the cross member, but the pesky engine tends to get in the way.

Without a brace across the top of the U, vertical loading from the top of the strut attempts to push the top of the tower up, but the tower is very stiff in this direction (or rather the metal between the tower and the rest of the chassis structure) so the force follows the path of least resistance and pushes the tower inward (despite the angle of 'push'), in which direction the tower is not nearly so rigid.

The tower brace tube is always very weak as a beam (i.e. flexes in bend fairly easily, put one end in a vice and heave on the other end, you'll see...), so cannot add any significant vertical rigidity as many people seem to think it does, and this is why it makes zero real difference whether the bracing bar is bolted to the tower brace brackets or welded to them. It's only effective as a compressive or tensile member.

Note that it's the lateral deflection of the towers that contributes to the totality of torsional chassis twist. It's hard to understand just why such a relatively small chassis deflection (twist) has such a profound affect on chassis dynamics, especially considering how small it typically is in relation to the suspension deflection, but it does, just ask any designer of racing cars how important chassis rigidity is...

[55EXX]

That can be a rather heavy obsession, as in kilos

my ek had the had maybe 100+kg of mdf subs and amps and accesories laden though it. the boot alone would have been the bulk of it. it affected the under/oversteer balance crazily. the hardest setting on the 22mm white rsb was no longer hard enough!

something you should know about this car. to me it was worth the weight.
http://memimage.cardomain.com/ride_images/3/1847/2821/29616410033_large.jpg
and there is more under the wheels arches and then sound dampening felt too.

yeah not into doof doof but rather quality sound! no use having i lost by stupid road noise.


If the front suspension has trailing radius rods attached to a forward subframe, then braces connecting the forward subframe to the 'rear' subframe would I think probably be worthwhile

like this?
http://i25.tinypic.com/2r7vib7.jpg

its again one of those things people are on both sides of the fence about. many reviews about the floor bar were like this and after installing one i definately recommend it so might give this bar a try. my only concern is ground clearance as i am just legal now at bout 110mm at my headers. that pic from this thread a while back. Click (http://www.ozhonda.com/forum/showthread.php?t=91622&highlight=brace) i repped you for a post you wrote in there.

so in making my x brace how would you do it. this is a drawing i did of the two methods i am deciding on. on bolts to mount other welded to the bar.
http://img211.imageshack.us/img211/1889/barex6.png

i know the x distance of the bar under bend forces is the down side of the welded and the unequal force on the mounts. which would you go?

[JohnL]

yeah not into doof doof but rather quality sound! no use having i lost by stupid road noise.

Then get some headphones!


like this?
http://i25.tinypic.com/2r7vib7.jpg

I was talking about when there are two subframes to which the suspenesion is attached. Connecting the two subframes with braces would be a good thing, but I've yet to figure out how to do it effectively on my CB7 (damn suspension gets in the way...).

That looks to be the same front suspension as in the EG Civic, which I don't think has any of the suspension attached to a seperate front subframe(?), as is the case with other Honda double wishbone front ends (e.g. CB7 / CD5 Accords).

From my recollection of how this suspension is arranged, all the lower suspension components attach to a single subframe located behind the engine, so all of the forces are reacted into a single subframe (excepting those forces reacted into the upper wishbone, which are substantially less than the forces reacted into the lower wishbone), not into two subframes connected by the main chassis structure?

I suspect what might happen with that suspension under hard braking is that the forces will try to pull the LCA attachment points apart from each other, causing the track width to increase slightly and the toe to toe-out to some degree(?). If so then I think the lateral member of the brace is acting in tension to limit this(?).

However the brace works, those bends in the lateral tube won't be good for the compressive or tensile strength of the member. I'm not sure that the longitudinal members are actually doing all that much, but then I can't really see the structure all that clearly (but if this suspension is mounted as I recall then I can't see much need for the longitudinal members).

I'm not saying that brace is a bad thing, but the lateral tube should be very beefy to minimise bending flexure.


so in making my x brace how would you do it. this is a drawing i did of the two methods i am deciding on. on bolts to mount other welded to the bar.

i know the x distance of the bar under bend forces is the down side of the welded and the unequal force on the mounts. which would you go?

In an ideal world all tubes would apex into the same point, but sometimes it's not so easy. You could weld as per your drawing but I'd make the tower brace tube somewhat more substantial than would otherwise need to be the case (to cope with the bend load created by the other tube).

Your bolted up arrangement is better I think as it has better force lines, and keeping the individual tubes seperate will make it easier to fit etc.

Of great importance is the design of the brackets that allow the tubes to be attached to the tops of the towers. They should be as robust as possible, which means they should also be as short as possible, i.e. the bolts that attach the bracing tubes to the brackets should be as close as possible to the top of the tower sheet metal.

My rear tower brace is very simple, only a 35mm tube (2mm wall) with each end flattened and drilled. It fits to only one of the strut studs per side (which is perfectly OK). The flattened ends are bent slightly to match the angles of the tower tops, and forces pass very directly from tower top to tower top with no significant doglegs (other than the slight bends at each end). If I were making / fitting adding additional braces that I wanted to feed forces to / from the tower tops, I'd weld robust tabs to my existing tower brace and bolt the tubes to those tabs (the existing brace is robust enough to cope without significant flexure, so long as the loads are fed into the tube as close as possible to the ends of the tube).

[55EXX]

can i please see some pics

[JohnL]

I don't have any. I could take some, but then I'd have to figure out how to post them, and I'm lazy...