147 Sway bar options

[Cowie]

G'day all,
Just wondering what sway bars were fitted to 147s'.
Apparently, according to overseas Alfa forums, some 147 models came with a 24mm hollow bar on the front. I know that the 147 GTA has a 23mm solid bar.
Does anyone know if some 147s' did indeed come with the 24mm hollow bar, and if so, which models/years would I be best to look at for the 24mm hollow bar that I'm chasing.
I'm at Dalby in Qld if anyone has one, or can recommend a wrecker / supplier up here that might be able to help me.

Many thanks in advance !

Tristan

[johnl]

Tristan,
I have no idea what is fitted to the 147 (but I think the rear ARB might be 14mm). I do know that for FWD cars it's nearly always more beneficial to increase the stiffness of the rear ARB relative to the front one (and not vice versa, which is counter productive, unless you like understeer).

Increasing front roll stiffness (without an at least equal stiffness increase at the rear) will increase lateral weight transfer at the front suspension, while simultaneously decreasing it at the rear suspension (even if no stiffness change is made at the rear). This will induce additional understeer, even if the handling response might be sharpened up to some degree. Note that because of the increased front weight transfer, there will be some loss of traction at the inside wheel when exiting corners.

Fitting a stiffer rear ARB will increase lateral weight transfer at the rear of the car, while reducing it at the front (again, even if no stiffness change has been made to front roll stiffness). This will reduce understeer, as well as improving front traction exiting corners.

This is not always the complete story though. Some cars with poor roll induced camber change characteristics in the front suspension (read MacPherson Struts) can decrease understeer by fitting a stiffer front ARB. This is because the reduction in body roll causes a lesser unwanted front camber change (i.e. less positive camber gain at the more heavily loaded outside wheel), so front grip increases despite the greater front weight transfer.

The 147 has a very good double wishbone front end, with good camber change characteristics, so the front wheels remain more vertical relative to the road as body roll occurs, even if quite a lot occurs. Fitting a stiffer rear ARB will still reduce body roll, so there will be less camber change at the front wheels, plus, there will be less weight transfer at the front. So, understeer will be reduced in two ways for the one modification.

My 147s' previous owner fitted a GTA front ARB, which is a bit stiffer than the non GTA bar. I also then fitted a modified Holden Rodeo front ARB (20mm solid) to the rear, and the result was excellent. Reduced understeer with sharper response, much less body roll and not nearly as harsh as I suspected it could have been (the ride is definitely quite a bit firmer, but not uncomfortably so). This modification required altering the shape of the Rodeo ARB, as well as fabricating two custom brackets to fit it, a bit hard to describe. I was also very lucky in that the Rodeo drop links (that came with the second hand ARB) fitted perfectly.

Modifying the shape of the Rodeo ARB involved a blow torch, some bending, and quite a bit of twisting. Any heat treatment of the bar will be lost, but it's highly likely that the bar may not have been heat treated in the first place. Many ARBs are made from a type of spring steel that does not really require heat treatment since the deflection (i.e. twist) of the bar is within the elastic limit of the non heat treated steel. So far no problems. I also did a similar thing to my last car (CB7 Accord), fitting a similarly modified Magna front ARB to the rear of the Accord, again with no issues at all, the bar worked very well for a number of years, and is still working for that cars' new owner.

I don't expect most people would or could go to this much trouble, but for me it was very worthwhile. If purchasing an aftermarket rear ARB, then my preference would be to get the stiffest one I could find.

Regards,
John.

[Cowie]

Thanks John, I will keep the Rodeo rear bar in mind for tarmac work.
My car is a GTA, so I already have the 23mm front bar ( haven't measured the rear ), and it will be predominately a gravel rally car, and as such the accepted rules of Roll couple generally get thrown out the window.
In case you're interested, the main reasons are the difference in static vs dynamic coefficient of frictions between tar and a loose surface ( and the implications for weight sharing between wheels), and wheel placement on the road.
Generally the inside front wheel will be tucked into the loose stuff by the time the car has transferred all weight to the outside, and the outside front will be on the clean swept line, so transferring all weight to the outside is not normally an issue, even lifting the inside front ( I have pics of a previous front drive rally car doing this) is not a real concern as it would only be running in leaf litter even if it was on the ground.
The other problem with trying to control chassis roll with the rear on a front drive rally car is the lack of weight in the rear. we are always trying to centralise the weight because a rally car has to rotate so much more than a tarmac car, so we can't take easy options like putting the spare tyre behind the rear axle. So the rear ends up having very little ability to control the roll of the entire chassis.
In theory a front drive rally car would have no front bar, and a stiff rear bar, because we want to get the car sliding as we don't reach maximum coefficient of friction until it is sliding profusely. But the reduced steering response of such a setup, partly due to the reduced ability of the rear to control roll, can make it a hairy ride indeed !  Think of a road car with a low roll stiffness, being chucked sideways and then trying to throw it the other way. The springs load up on one side, and then spit all that energy back the other way as the car changes direction, making it very hard to drive, and usually resulting in a spin. So even though a higher roll stiffness in the front will technically provide less traction, the confidence it instils in the driver is worth so much more than a small loss in traction.
Out of interest, I compete with a fellow up here running a VW Polo S2000, and the top line Reigers on that car have "roll control", a very low speed circuit designed to control roll in lieu of an Arb, apparently it works quite well.
I won't be running a rear Arb whilst on gravel, counter intuitive I know !
Anyway, all that is possibly of no interest to yourself or anyone else on the forum! But I will still be looking for a 24 mm hollow bar if they actually exist, for no other reason than it should be nearly as stiff as the Gta bar, while being a good bit lighter   

Thanks again John,
Tristan

[johnl]

Tristan,
You're wrong, I found that quite interesting. 

I've driven on a lot of dirt roads, and have three km of dirt road at my house before I get to the tar. Generally I've found that the things I do to my cars to make them handle better on the tar also make them handle better on the dirt, but then I'm not at 10/10ths hanging it sideways with the inside front wheel in the leaf litter and general roadside rubble and crap (which could include rotting roadkill..). The possibilities of a car coming the other way, or damaging my car on the rough dirt surface, or impacting a kangaroo, or a stray cow, tend to moderate my driving behaviour on the public road...

Centralising weight (mass) is a good thing to do, unless maybe you're drag racing, or racing on a track with lots of exits from very low speed corners. To "centralize the weight", I can't see why you can't do things like placing the spare behind the rear axle line? This would help to move the CG toward the middle of the car, so achieving the stated aim, though would increase polar moment, and thus decrease yaw response. So I assume this is the real issue when moving your mass 'too far' rearward?

A very stiff rear end will still control roll angle of the whole car, at least up to the point that the inside rear becomes near fully unloaded. After that the car is a tricycle (and rear roll stiffness is then a non issue until the inside rear drops back to the ground), and so then the entirety of roll stiffness comes from the front end. You do want a fair degree of roll stiffness, not so much to limit the roll angle (though not unimportant) but more to improve steering and handling response. A higher degree of roll stiffness improves transient response because weight transfers occur more quickly, and the handling is thus more responsive to driver inputs. If the car rolls too much, then weight transfer is slower and the car takes longer to 'settle' into a cornering attitude, and so feels 'soggy'.

My last car was a CB7 Accord (the one I mentioned with the Magna front ARB in the rear, and stiffer springs, and Koni 'Sports', and stiffer bushes, and some added bracing, etc.). The Magna bar wasn't hugely stiff, but much much much stiffer than the stock Accord rear bar (14mm spaghetti, that did almost nothing at all, as with most stock rear ARBs, including the one in the 147). In the end I found the Accord cornered better with the front ARB deleted altogether, better cornering balance (i.e. less understeer), but there was a slight reduction in steering and handling response. Overall the trade off was worth it, with this car, though it should be noted that the Accord's springs were significantly stiffer than stock and a bit lower than stock, so a quite substantial increase in overall roll stiffness came from the springs alone.

I haven't tried this yet (front ARB delete) with the 147, there are more important issues to address first. As yet I have only replaced the front dampers with B6 Bilsteins (all four 'old' dampers were newly installed stock rated, but the front ones were absolute crap), and fitted the modded Rodeo ARB to the rear (both mods being great improvements). It still needs (at least) some stiffer bushes (particularly in the rear end), and better rear dampers (B6 to match the front ones, when cash permits).

Less front roll stiffness also made a significant difference to how well the Accord put the power down on loose or slippery surfaces, less weight transferred from the inside front resulted in less wheelspin exiting tighter corners (no trick diff in the Honda). Accelerate hard coming out of a tight gravel road corner and I could get wheelspin with both front wheels (even sans front ARB, with the ARB I could only spin up the inside front wheel, which was a liability), which could be used to lessen understeer. Say you came into the corner a bit too quickly, and the car was wanting to run wide. You (I) had two choices, either back off in the hope that forward weight transfer would regain some front grip (maybe), or, turn the front wheels in hard and apply lots of throttle (of course you are already in the appropriate gear). The former often worked, the latter often worked. The choice you made has a lot to do with just how loose the surface is, clean surface and you treat it like a tar road and back off, loose gravelly surface and the latter will 'pull' the car tighter into the corner (happy face...). Requires confidence and quick decision though, get it wrong and you'll (I'll) likely be off the road.

I think this works for the same reason that you see powerful speedway cars cornering at a substantial angle to the direction of travel, with lots of wheelspin throwing sprays of dirt from beneath the drive wheels. I suspect that the tyres are 'self sweeping' the road, removing the loose crap and getting down to the less slippery surface beneath...?

I could waffle on for a lot longer, better to stop now...

Regards,
John.

[Cowie]

Damn it, I just wrote about a lot more of our interesting findings, now it's been deleted

[johnl]

Well, write it again damn it...

Regards,
John.

[Cowie]

It's good to find someone on here interested in loose surface vehicle dynamics, can't say that I expected that on an Alfa forum.

Yes it can be frustrating driving a nice gravel road and having to go steady. I competed in my first rally at 19, and not only were we allowed to go as fast as we liked, it was thoroughly encouraged, been hooked ever since. . . .

So it has become apparent amongst our group of friends and competitors, that for a fast, spectacular and sideways rally car, a tendency for an understeering setup is needed. From fast front wheel drives with high front roll stiffness at the front and lower on rear, to the 4wds of multi national champions who always run a locked ( or all but locked) centre diff, tight front diff, and a middle of the road or even looser rear diff, the theme is the same. These setups would be considered to tend towards understeer in a tarmac situation.
But on tarmac the car is driven in a balanced manner, the load steadily increasing on one side of the car while entering the corner. Driving in a balanced manner on gravel will lead to two things, slow times, or a crash and sometimes both. On the loose stuff the car is given a ''chuck'' some ways out from the corner, setting it sideways to the road so that the driver can then make adjustments to the line with throttle and brake as much as with the steering wheel, possibly even more  with the pedals on a fwd car.
So what I think happens is that the high front roll stiffness is needed for the quick response for the initial turn in to set the car sideways, and in a nicely setup car the steering wheel can be brought back to centre-ish, without having to ward off the cars desire to keep rotating. If the car were to have a lower front roll stiffness, even though ultimate traction levels might be increased, the delayed reaction can tend to make the driver put in a bigger or longer steering input to get the front end moving, giving them a feeling of understeer. Then mid corner the car would have lots of front end traction, possibly be yawing faster than it otherwise would be due to the drivers extra input, and if the rear roll stiffness is relatively high, it will try to keep yawing further than desired. This is known as the understeer on turn in, mid corner oversteer dreaded by rally drivers. I serviced for a mate last year at the border Ranges rally, ( fast smooth shire roads), so he dropped the ride height on his evo 6 20mm. He came in to the first service and the car just would not turn  in for him. So we raised it again, didn't worry about aligning it because the evo's have very little bump steer, and it was fixed, back to the delightfully handling car he had in previous rallies. So the lowered roll centre and roll stiffness is the only thing we can attribute the poor turn in to.

I recently pulled the front arb off my brothers Hyundai excel rally car, to help with drive due to the open diff. It helped immensely with the drive, but became much worse to drive, especially in high speed changes of direction. I have bolted it back on, yet to drive it in anger again.

You are right about the polar moment. We have few options to remove weight off the front end without spending $$$, and sticking the spare right out the back makes a noticeable increase to the willingness of the car to keep rotating once started, making for slower direction changes. So fwd rally cars usually end up even more nose heavy than the same road car, limiting the ability of the rear to control roll. Like you say, once the inside rear is lifted, that's it.
The other consideration is road placement of the rear wheels. If being driven correctly, the car will have enough attitude that the inside rear will be following a line similar to the outside front, i.e., the clean line with lots of traction. In a flip of the situation out the front, the outside rear will be out in the loose stuff, so my way of thinking is that it is better to share the weight as equally as possible across the back wheels, to take advantage of the traction available to the inside rear, and stop the car from being an unwieldy oversteering battle.

As for the effect of polar moment, I often describe it this way to rally newcomers. We drive on tarmac headed north 0deg, turn right through 30 deg ( E,030), then left again through 30deg back to North. In the rally car you may have 15 deg or more of attitude relative to the road  as you exit the first right, and will be pointing to E 045 Deg. Then we flick it back in one motion to an angle 15 deg more than the angle of the next left corner, and the car now points to W 345deg. So the tarmac car rotates 30 deg, then another 30 deg.  The rally car rotates through 45 deg in the first corner, then through 60 degrees in the one flicking motion ( 045 deg to 345 Deg)as it turns left.  This usually lets them see how important it is to tuck heavy stuff like spare tyres right up behind the seats.

You are dead right about the tyres sweeping loose material off the top and getting down to the traction if it is there. It's the same reason Abs doesn't work on gravel, the fastest way to stop is with the wheels locked. When coming in too hot in the rally car, drivers will lock all 4 hard, just lifting momentarily as needed to provide steering inputs to set the car up for the corner.

Those speedway boys are the ones that really understand loose surface chassis dynamics, they have cool tools like adjustable roll centres and stagger, but they only have to turn one way on a smooth surface 

I don't actually know what I am talking about with regards to any of the above. I do know that the adjustments described work for a rally car, but it may or may not be for totally different reasons to those I have described, they sound good in my head though.
I do know that my Gta will sound great howling through the forest with 3'' straight though exhaust and throttle bodies, and the sound is the most important thing 

I am mainly looking for the hollow bar because I suspect it will be nearly as stiff as the solid gta bar, but a good bit lighter.

Regards Tristan

[johnl]

It's good to find someone on here interested in loose surface vehicle dynamics, can't say that I expected that on an Alfa forum.

Tristan, it's all grist to the mill of understanding chassis dynamics in general (or trying to...). Thank you for your thoughts.

Yes it can be frustrating driving a nice gravel road and having to go steady. I competed in my first rally at 19, and not only were we allowed to go as fast as we liked, it was thoroughly encouraged, been hooked ever since. . . .

Have you ever driven a proper racing kart? Much cheaper than rally cars, and fewer trees to hit...

For the sort of dosh you'd spend in rallying (and I'd bet with a lot of change leftover) you could easily field a competitive 125cc or even 250cc 'Superkart' to run on the full sized car tracks at speeds up to around 260kmh (and with plenty of runoff area...). I've never driven such a weapon, but have driven a 125cc 'shifterkart' on a sprint kart track, and it certainly gets your full attention. With about 48 hp these gearbox 'sprint karts' have a power to weight of about 270hp per tonne (including driver), but that pales next to a 250cc 'Superkart' (more like 90hp and about 440hp per tonne...).

So it has become apparent amongst our group of friends and competitors, that for a fast, spectacular and sideways rally car, a tendency for an understeering setup is needed. From fast front wheel drives with high front roll stiffness at the front and lower on rear, to the 4wds of multi national champions who always run a locked ( or all but locked) centre diff, tight front diff, and a middle of the road or even looser rear diff, the theme is the same. These setups would be considered to tend towards understeer in a tarmac situation.

I suppose a fundamentally oversteering balance would make the car a real handful on a loose surface? So long as the car wasn't so severe an understeerer that it continually wants to 'snap' back from 'provoked' oversteer into strong understeer, I think I can see why a somewhat 'understeering' balance might be faster, or at least more confidence inspiring (which can amount to the same thing in the end).

But on tarmac the car is driven in a balanced manner, the load steadily increasing on one side of the car while entering the corner.

We should keep it in mind that even cars which feel quite 'neutral' in their balance (even tarmac racing cars) are really understeerers, just less so than most 'normal' production cars. A truly neutral basic balance will tend to continually cross the line back and forth between under and over steer, and be a real handful to drive fast. Better if it understeered, or just consistently oversteered, at least it would be predictable. If it stays on just the right side of understeer most of the time, then it will be a lot easier to drive.

Driving in a balanced manner on gravel will lead to two things, slow times, or a crash and sometimes both. On the loose stuff the car is given a ''chuck'' some ways out from the corner, setting it sideways to the road

That is, the driver 'provokes' the car into oversteer, despite its' basic understeering balance.

With 'sprint' karts the driver can provoke a degree of corner entry oversteer by slightly steering into the corner at the braking point and then 'over-braking' the kart slightly (most sprint karts only have rear wheel braking of course, though any kart with front brakes as well as the rear axle brake will shed speed harder than most people can imagine possible, nearly pulls your face off). This sets up a drift / slide where the steering wheel is pointing more or less at the straight ahead and the front tyres are also resisting rotation, and so can also generate a degree of effective braking force (despite lack of front brakes). This lessens the braking distance slightly, which is a good thing in and of itself.

Get it right and the kart reaches the corner at some attitude to the direction of travel, and so is already 'pointing' more or less in the right direction for the corner exit, even before the theoretical corner apex is reached. Because of this the driver can hit the throttle significantly before the theoretical corner apex, and the kart accelerates into, through and out of the corner (hopefully faster than the other drivers can manage to do). Of course this improves exit speed, which is an advantage carried all the way down the next straight. Hard on the rear tyres though.

Sorry, heading way off topic...

so that the driver can then make adjustments to the line with throttle and brake as much as with the steering wheel, possibly even more  with the pedals on a fwd car.

I do find that a FWD car on a loose surface responds well to throttle inputs, in a different way to RWD cars (no great surprise or insight there...). When the front wheels are pointed well into the turn (even 'oversteered' into the turn), and then provoked into wheelspin, they literally 'pull' the front end across the road into the turn. This affect has kept my car out of ditches more than a few times, and not even driving all that fast (say when you put a wheel on the slippery downward slope at the edge of the road and the car 'wants' to slide down that slope). It doesn't really work on a sealed surface, I think because on such a surface spinning up the front wheels (if you even can) just eats too big a portion of the 'traction circle' pie, leaving too little remaining for sufficient 'lateral' grip.

So what I think happens is that the high front roll stiffness is needed for the quick response for the initial turn in to set the car sideways, and in a nicely setup car the steering wheel can be brought back to centre-ish, without having to ward off the cars desire to keep rotating. If the car were to have a lower front roll stiffness, even though ultimate traction levels might be increased, the delayed reaction can tend to make the driver put in a bigger or longer steering input to get the front end moving, giving them a feeling of understeer. Then mid corner the car would have lots of front end traction, possibly be yawing faster than it otherwise would be due to the drivers extra input, and if the rear roll stiffness is relatively high, it will try to keep yawing further than desired. This is known as the understeer on turn in, mid corner oversteer dreaded by rally drivers.

Sounds plausible to me.

I serviced for a mate last year at the border Ranges rally, ( fast smooth shire roads), so he dropped the ride height on his evo 6 20mm. He came in to the first service and the car just would not turn in for him. So we raised it again, didn't worry about aligning it because the evo's have very little bump steer, and it was fixed, back to the delightfully handling car he had in previous rallies. So the lowered roll centre and roll stiffness is the only thing we can attribute the poor turn in to.

Maybe the lowering was enough to adversely affect the dynamic camber change in bump and roll? Just an alternative theory (as opposed to an alternative fact...).

Mac' struts are bad this way, when the suspension is lowered the desirable camber gain is reduced. In some cases this can even become opposite to what is wanted, i.e. instead of some increasing negative camber with bump motion (which of course is also roll motion at the outside wheel), you can get increasing positive camber gain instead of negative camber gain. If the lowering results in the control arms becoming horizontal or near to it, then it's very likely for this to happen (this is why those kits exist, the ones to lower the location of the ball joint, i.e. the ones that restore the angle of the control arm on lowered Mac strut suspensions).

Lowering a Mac strut suspension can also result in the geometric roll centre (GRC) lowering quite a bit more than you might expect (i.e. more than the degree to which the CG lowers, and maybe substantially more, especially if the lateral control arms are on the shorter side). If so, then there can be a significant reduction in geometric roll stiffness (which would not be the case if the CG and the GRC both lowered to the same degree, though of course lowering the CG will reduce total weight transfer regardless of RC height, whether it be geometric or elastic weight transfer).

Geometric roll stiffness is very important for responsiveness, because the % of the total weight transfer that occurs through the geometric vector occurs instantaneously (i.e. in locked step with the rise, or fall, in lateral force), as opposed to the %age of weight transfer that occurs elastically (i.e. via springs and ARBs), which has an element of 'delay' associated with it.

So, I think you may possibly be thinking along the right track, lowering the CG will very likely also lower the GRC even more, resulting in less geometric roll stiffness, slower weight transfer and 'soggier' response, possibly enough to cause a problem.

Note that rear roll stiffness, and the ratio of geometric to elastic roll stiffness, is very important for steering response. Body roll is not just body roll, it is one of the things that contribute to 'lateral stiffness', and lateral stiffness is important for responsive steering and handling. Lateral stiffness has to do with how much the CG of mass (as a representation of the total sprung mass) moves laterally relative to the contact patches when the chassis mass and suspension is laterally loaded.

The things that comprise this lateral movement are body roll (as the unsprung mass 'rolls' it moves laterally relative to the contact patches, more so the less geometric stiffness), lateral tyre deflection, lateral suspension bushing compliance, wheel and suspension member deflection, all of which allow the CG to move laterally relative to the contact patches.

The CG moves laterally because it is (usually) some distance above the roll centre, so there is a 'lever arm' effect working to 'roll' the chassis around the roll centre, which acts as a virtual pivot (this is very oversimplified, the physics of roll motion are complex). Tyre deflection is self evident, as is bushing compliance and flexure of suspension members etc.

This lateral CG movement takes time to happen, which impacts on the speed with which weight transfer occurs and the rate at which the chassis mass will 'yaw' in response to steering and other inputs. When the front wheels turn in the chassis responds by moving in the direction of the steer, but if rear lateral stiffness is low, then the chassis will tend to yaw before it actually starts to change direction on the road. So low rear lateral stiffness slows steering and handling response. Higher rear geometric roll stiffness decreases this unwanted 'yaw before turn' affect. But, raising the geometric roll centre also makes the suspension less sensitive to elastic roll stiffness, i.e. changes in spring and / or ARB stiffness become less influential. 

Stiffer rear bushes also sharpen up steering and handling because this also increases lateral stiffness, as do stiffer sidewalls and higher tyre pressures (higher psi makes the tyres behave as if the sidewalls were stiffer). I'm sure this is why my 147 steers and handles MUCH more sharply when ambient temperature is low, and is a lot vaguer when ambient temperature is high. That is, in hot weather all the rubber in the bushes is substantially softer than it is when the temperature is cold, so when it's hot the lateral stiffness is badly affected (stiffer bushes are on the wish list).

If geometric roll stiffness is 100% of total axle roll stiffness, then no body roll will occur (at all), and the geometric lateral stiffness will effectively be 'rigid', and all lateral weight transfer will be 'instant'. If geometric roll stiffness is less than 100% then body roll does occur, and a %age of the weight transfer occurs at the rate of the speed of the body roll motion (as the springs and ARBs deflect), which is relatively 'slow'. If geometric roll stiffness is 100% at one end of the car, and less at the other, then some body roll will still be evident at both ends, but the end with 100% geometric roll stiffness will still transfer it's 'weight' instantaneously. This would create a very responsive set up, but also one with a lot of other problems. It could well be too 'twitchy', which will be compounded with the excessive jacking force associated with very high roll centres (think VW beetle, Corvair, and some earlier model Fiats that all used crude swing axle rear suspension with a very high geometric roll centre).

This tendency to increase steering and handling response is at least one reason why so many higher performance and racing cars use a rear geometric roll centre that is significantly higher than the front geometric roll centre, but not so high that it creates other significant problems. A balance in all things, and horses for courses...

Sorry to ramble on, one thought leads to another, and the knee bone is connected to the ankle bone, all the way up to the head bone. Not meaning to write an essay, but after writing X you realise it doesn't make sense unless you explain Y, and then Z, and then...

This is just scratching the surface of the dynamics related to roll stiffness, weight transfer, steering and handling response etc. It's all very interactive, complex, and confusing. It can make your brain bleed if you think about it too much. And, take what I say as being worth what it cost you to read it, I'm no expert on this stuff (if I were then I could probably be paid well for it), just an amateur and sometime confused 'thinker'...

You are right about the polar moment. We have few options to remove weight off the front end without spending $$$, and sticking the spare right out the back makes a noticeable increase to the willingness of the car to keep rotating once started, making for slower direction changes. So fwd rally cars usually end up even more nose heavy than the same road car, limiting the ability of the rear to control roll.

Well, at least that will help to put the power down...

The other consideration is road placement of the rear wheels. If being driven correctly, the car will have enough attitude that the inside rear will be following a line similar to the outside front, i.e., the clean line with lots of traction. In a flip of the situation out the front, the outside rear will be out in the loose stuff, so my way of thinking is that it is better to share the weight as equally as possible across the back wheels, to take advantage of the traction available to the inside rear, and stop the car from being an unwieldy oversteering battle.

I can see why you'd want the rear wheels following in the tracks of the front wheels, already swept as you say. If the inside rear is travelling a swept path, then the outside rear would by default be doing the same thing.

As for the effect of polar moment, I often describe it this way to rally newcomers. We drive on tarmac headed north 0deg, turn right through 30 deg ( E,030), then left again through 30deg back to North. In the rally car you may have 15 deg or more of attitude relative to the road  as you exit the first right, and will be pointing to E 045 Deg. Then we flick it back in one motion to an angle 15 deg more than the angle of the next left corner, and the car now points to W 345deg. So the tarmac car rotates 30 deg, then another 30 deg.  The rally car rotates through 45 deg in the first corner, then through 60 degrees in the one flicking motion ( 045 deg to 345 Deg)as it turns left.  This usually lets them see how important it is to tuck heavy stuff like spare tyres right up behind the seats.

Well high polar moment will slow down all responses, including corrective actions. The car will tend to get into trouble more 'slowly', but be harder to 'rescue' once it has. In my experience with a Lotus 7 style of car (a 'Nota'), cars with a very low polar moment tend to be delightfully precise and responsive to drive, both with reactions to inputs and to corrections. I recall Clarkson describing a Caterham 7 as 'changing direction like a fly", which they do. Karts are the ultimate expression of this.

I don't actually know what I am talking about with regards to any of the above.

Paint me with that brush too. As I said above this sort of thinking is more or less a hobby of sorts for me.

I am mainly looking for the hollow bar because I suspect it will be nearly as stiff as the solid gta bar, but a good bit lighter.
Without doing the mathematics, I would expect a 24mm hollow bar would probably be significantly less stiff than a 23mm solid bar.

There are formulas and calculators on the web that would help you determine the relative stiffness of two ARBs of similar shape but different thickness, and probably hollow vs solid. Here's the first one I found with Google:

http://www.gtsparkplugs.com/Sway-Bar-Calculator.html

It's imperial measurements, so you'd have to convert the metric to inches. Since you are only wanting to determine relative stiffness between solid 23 mm and hollow 24, you could just use arbitrary numbers for dimensions other than the OD and ID.

There is a notion 'out there' (i.e. I've read people writing this) that a hollow tube is more rigid in both bend and torsion than a solid bar of the same OD. This is not correct. The confusion lies in that a hollow bar is weight for weight stiffer than a solid bar, but the hollow bar has to be of a significantly greater OD for this to be true.

I doubt the weight saving with a hollow bar will be significant, considering that the difference in bar mass (whether solid or hollow) is only a very small %age of the totality of unsprung mass. Yes, in theory it is better if the ARB were lighter, but in practice I'd be amazed if it made any material difference in the forest.

It takes only a quite small increase in bar OD to make a very substantial increase in stiffness, all else being equal. Stiffer 'D' bushes also help significantly, especially with initial bar deflection. With a stiff ARB much of the initial deflection is actually 'D' bush deformation. You want initial deflection to be reasonably stiff for initial response.

Also note (slightly off topic) that the steel that the bar is made from, and that steels' (whatever steel it is) state of heat treatment (or not), makes next to zero difference in how stiff the bar will be. This is only determined by the OD and general shape of the ARB, because nearly all steels (excluding some rather exotic alloys that won't be found in an anti roll bar) have the same modulus of elasticity (within a percentage point). All that changes to any significant degree between different grades of steel and with differing heat treatment is the hardness, and the yield point, i.e. whether the steel will snap or permanently bend at whatever loading. Below the individual steels' yield point, all steels are elastically similar.

Regards,
John.

[johnl]

Tristan,
It occurred to me that you probably won't know the wall thickness of the hollow ARB. The best you can do if you can't find out, is to estimate a likely wall thickness. I would expect it to be around 2mm to about 3mm, something like that (I'd be surprised if were to be as thick as 3.5mm, but maybe). Using the ARB stiffness calculator you could play with different minimum IDs' to get some sort of range of expectation. It would only be approximate, but maybe that's all you need?

Out of curiosity I have done it for you, with some arbitrary assumptions that won't affect the relative stiffness numbers, so long as you understand that the resulting stiffness numbers don't really represent anything other than a relative comparison.

Keeping mind that this particular calculator works with imperial measurements, the arbitrary assumptions are that the:

ARB arms = 15"
The lateral section of the ARB = 25"

A 24mm bar = 0.945" OD, a 23mm  bar = 0.906" OD

If the wall thickness of a 24mm hollow bar is:
2mm thick, then the ID = 0.867"
2.5mm thick then the ID = 0.847"
3mm thick then the ID = 0.827"
3.5mm thick then the ID = 0.808"

So the online calculator makes these numbers:

A 23mm solid ARB (with above assumed dimensions) has a stiffness of 106.9 pounds / inch

A 24mm solid bar has a stiffness of 126.53 pounds / inch

A hollow 24mm bar with a 2mm wall thickness has a stiffness of 36.88 pounds / inch

A hollow 24mm bar with a 2.5mm wall thickness has a stiffness of 44.87 pounds / inch

A hollow 24mm bar with a 3mm wall thickness has a stiffness of 52.33 pounds / inch

A hollow 24mm bar with a 3.5mm wall thickness has a stiffness of 58.91 pounds / inch

So you can see that even with an assumed quite thick wall, the hollow 24mm bar will be way less stiff than the solid 23mm bar.

If you had access to an example of the hollow ARB, you could perhaps measure the thickness of one of the end flanges where the tube has been flattened to form it. The tube wall thickness would be likely half the thickness of the flange, sine the flange is most probably formed by flattening the ARB tube. But I doubt you want to use the hollow bar now??

Regards,
John.