virtual swing arm length
Discussion
Is there any formula for working out the "best" virtual swing arm length? I've got 2 books, both of which talk about "choosing" your swing arm length without commenting on how you might like to do this. Both reference other books which I'm going to have to wait 6 weeks for access to when our library re-opens that section 
No, there isn't. It depends on what you are trying to achieve...
Main factors are good camber control in bump, good camber control in roll, and good geometric roll centre location. As a general rule, you'll find that it's quite easy to achieve 2 out of the 3, but pretty much impossible to achieve all three. Since all are equally important, you have to decide on your own compromise. Personally, I lean toward the 'good roll centre location' end of the spectrum, but it does depend on your tyre characteristics and range of suspension movement.
Just to complicate matters further, the VSA length isn't the only factor. Location of pick-up points on chassis and upright can make a big difference: you can have suspension geometries with identical VSA lengths but dramatically difference geometric roll centre/camber characteristics, due to different pick-up locations.
It's probably worth adding that the geometric roll centre is NOT a point about which the chassis rolls, therefore it probably isn't where you think it is when you are trying to plot your geometry. In reality, the chassis will adopt a skewed bump/roll attitude in cornering (ie 'leaning' on one corner of the vehicle), depending on springs and weight transfer. The geometric roll centre is just one of the factors you use to calculate the weight transfer, that's all (but that means that if it is moving about dramatically as the vehicle corners, the weight transfer will be varying at the same time and the car's handling will feel, at best, rather uncertain!).
There was a looooong discussion about roll centres and weight transfer some while ago. It did get a bit farcical, but there's some good discussion on there if you are serious and can be bothered to sift through the nonsense... I'm sure that searching the terms 'roll centre' and 'weight transfer' on this forum would throw it up pretty quickly.
Confused yet? If you're not, you clearly haven't grasped what you're up against!
Main factors are good camber control in bump, good camber control in roll, and good geometric roll centre location. As a general rule, you'll find that it's quite easy to achieve 2 out of the 3, but pretty much impossible to achieve all three. Since all are equally important, you have to decide on your own compromise. Personally, I lean toward the 'good roll centre location' end of the spectrum, but it does depend on your tyre characteristics and range of suspension movement.
Just to complicate matters further, the VSA length isn't the only factor. Location of pick-up points on chassis and upright can make a big difference: you can have suspension geometries with identical VSA lengths but dramatically difference geometric roll centre/camber characteristics, due to different pick-up locations.
It's probably worth adding that the geometric roll centre is NOT a point about which the chassis rolls, therefore it probably isn't where you think it is when you are trying to plot your geometry. In reality, the chassis will adopt a skewed bump/roll attitude in cornering (ie 'leaning' on one corner of the vehicle), depending on springs and weight transfer. The geometric roll centre is just one of the factors you use to calculate the weight transfer, that's all (but that means that if it is moving about dramatically as the vehicle corners, the weight transfer will be varying at the same time and the car's handling will feel, at best, rather uncertain!).
There was a looooong discussion about roll centres and weight transfer some while ago. It did get a bit farcical, but there's some good discussion on there if you are serious and can be bothered to sift through the nonsense... I'm sure that searching the terms 'roll centre' and 'weight transfer' on this forum would throw it up pretty quickly.
Confused yet? If you're not, you clearly haven't grasped what you're up against!
Thanks Sam.
Yes, I most certainly am confused, I've had a headache for days now
However, I feel there is little point me trying to design my own car unless I can at least understand what the suspension on the thing is doing! The plan is to design it as best I can, then pay someone with real-world experience to assess it and tell me if I'm about to build the worlds next shopping-trolley-with-bent-castors...
I came across the post you mentioned while looking for an answer on the VSA length - I'll have to try again as I got a bit frustrated trying to sift. I think I'll need to as, as you say, without fully understanding the roll centre I'm not going to know - or even be able to estimate - what point the car is really going to roll around.
I've learnt pretty quickly that one man's "perfect" set up is another man's shambles, read 2 different books and you'll get 2 completely different answer. Just last night I spent an age trying to find out the pro's and con's of positive or negative scrub radius on a mid engined... after the forth find on the subject it would appear there are no pro's or cons, merely choice of opinion
Yes, I most certainly am confused, I've had a headache for days now
However, I feel there is little point me trying to design my own car unless I can at least understand what the suspension on the thing is doing! The plan is to design it as best I can, then pay someone with real-world experience to assess it and tell me if I'm about to build the worlds next shopping-trolley-with-bent-castors... I came across the post you mentioned while looking for an answer on the VSA length - I'll have to try again as I got a bit frustrated trying to sift. I think I'll need to as, as you say, without fully understanding the roll centre I'm not going to know - or even be able to estimate - what point the car is really going to roll around.
I've learnt pretty quickly that one man's "perfect" set up is another man's shambles, read 2 different books and you'll get 2 completely different answer. Just last night I spent an age trying to find out the pro's and con's of positive or negative scrub radius on a mid engined... after the forth find on the subject it would appear there are no pro's or cons, merely choice of opinion
Davi said:
...without fully understanding the roll centre I'm not going to know - or even be able to estimate - what point the car is really going to roll around.
To be honest, I've come to the conclusion that the traditional way of modelling geometry - by looking at one end of the car in isolation and checking pure bump/pure roll (with roll assumed to be around the geometric roll centre) is the next best thing to worthless. You really need to do a full 3-dimensional analysis of the car, modelling spring/ARB rates and response to cornering acceleration, before you get anything close to the true situation.Unfortunately, so far, I haven't come across any software that can do this properly for less than tens of thousands of pounds, so it's back to paper dollies and educated guesswork...
If you can come up with a layout in which the geometric roll centre doesn't move much relative to the sprung mass, no matter what combination of bump/roll movement you throw at it, though, then it's fairly simple to model a reasonably accurate analysis using nothing more than Microsoft Excel, AutoCAD, and the chapter on weight transfer in Staniforth's 'Competition Car Suspension'. The trick is getting this 'fixed roll centre' in combination with acceptable levels of camber control.
You almost need to reverse-engineer the layout: start by assuming fixed roll centres, then model the weight transfer (with estimated weights, CG height, spring/ARB rates, etc.) to work out what sort of skewed roll/bump attitude the car will adopt when cornering (camber geometry doesn't matter at this stage). Once you've done that, then go back and try to find a geometry that gives you the fixed roll centres that you have assumed, plus reasonable camber control for both dive/squat and the skewed roll cornering state that you calculated from the weight transfer.
If that's not complicated enough for you, you can then muddy the water still further by trying to model transient response and damping...
Davi said:
... I spent an age trying to find out the pro's and con's of positive or negative scrub radius on a mid engined... after the forth find on the subject it would appear there are no pro's or cons, merely choice of opinion
Yep, that sounds about right. For what it's worth, my personal opinion is that whilst negative scrub radius gives better stability (particularly in the unlikely event of a blow-out), it tends to deaden feedback through the steering to an unacceptable degree. I don't know about their current models, but it certainly used to be much favoured by Audi... 'nuff said??!I'd always prefer a small amount of positive scrub radius, as part of the steering weigh/feel equation, but you're right - that's just my personal choice/opinion.
Sam_68 said:
Davi said:
...without fully understanding the roll centre I'm not going to know - or even be able to estimate - what point the car is really going to roll around.
To be honest, I've come to the conclusion that the traditional way of modelling geometry - by looking at one end of the car in isolation and checking pure bump/pure roll (with roll assumed to be around the geometric roll centre) is the next best thing to worthless. You really need to do a full 3-dimensional analysis of the car, modelling spring/ARB rates and response to cornering acceleration, before you get anything close to the true situation.Unfortunately, so far, I haven't come across any software that can do this properly for less than tens of thousands of pounds, so it's back to paper dollies and educated guesswork...
Hmmm slight dampner on my plans there though - Finding it bad enough trying to do this in isolation! Just too much guesswork for me as I have absolutely zero idea about spring / arb rates or cornering acceleration at the moment. Sam_68 said:
If you can come up with a layout in which the geometric roll centre doesn't move much relative to the sprung mass, no matter what combination of bump/roll movement you throw at it, though, then it's fairly simple to model a reasonably accurate analysis using nothing more than Microsoft Excel, AutoCAD, and the chapter on weight transfer in Staniforth's 'Competition Car Suspension'. The trick is getting this 'fixed roll centre' in combination with acceptable levels of camber control.
that sounds a bit more encouraging - unfortunately our library is undergoing renovations - I was in there the day before yesterday asking about Staniforth's book, but they can't access it till the middle of November. Joy. Sam_68 said:
You almost need to reverse-engineer the layout: start by assuming fixed roll centres, then model the weight transfer (with estimated weights, CG height, spring/ARB rates, etc.) to work out what sort of skewed roll/bump attitude the car will adopt when cornering (camber geometry doesn't matter at this stage). Once you've done that, then go back and try to find a geometry that gives you the fixed roll centres that you have assumed, plus reasonable camber control for both dive/squat and the skewed roll cornering state that you calculated from the weight transfer.
weight, CG height I'm reasonably confident on now... as above though, spring / arb rates - not a clue. Sam_68 said:
If that's not complicated enough for you, you can then muddy the water still further by trying to model transient response and damping...
Now you're just trying to upset me Sam_68 said:
Davi said:
... I spent an age trying to find out the pro's and con's of positive or negative scrub radius on a mid engined... after the forth find on the subject it would appear there are no pro's or cons, merely choice of opinion
Yep, that sounds about right. For what it's worth, my personal opinion is that whilst negative scrub radius gives better stability (particularly in the unlikely event of a blow-out), it tends to deaden feedback through the steering to an unacceptable degree. I don't know about their current models, but it certainly used to be much favoured by Audi... 'nuff said??!I'd always prefer a small amount of positive scrub radius, as part of the steering weigh/feel equation, but you're right - that's just my personal choice/opinion.
Davi said:
...unfortunately our library is undergoing renovations - I was in there the day before yesterday asking about Staniforth's book, but they can't access it till the middle of November. Joy.
E-mail me via my profile. It may take me a couple of days to get round to it, but I'm pretty sure I can get the relevant chapter on weight transfer scanned for you before November. It's not actually very many pages long and the maths is refreshingly simple! That particular chapter is written by David Gould. The rest of the book (by Allan Staniforth) doesn't do anything but cover the basics in a pretty conventional form, so whilst it's nice and easy to read, you're not missing anything if you haven't got access to it.Davi said:
Sam_68 said:
If that's not complicated enough for you, you can then muddy the water still further by trying to model transient response and damping...
Now you're just trying to upset me If you tell the damper manufacturer your spring rates, sprung weight, etc., they'll valve the dampers close enough to the mid-range of what you need and you can tune the settings by trial and error once the car is on the road.
Davi said:
having spent more time reading, I'm heading in the same direction - and for good measure have picked a totally arbitrary figure of 20mm positive. One day I might come up with a plausible reason I chose that figure. Right now it's simply "because I want to write at least one thing on this shiny page"
Well, 20mm positive sounds a perfectly plausible and acceptable figure, to me! To be honest, most people building specials just stick with a proprietary front upright and whatever geometry it dictates to them, so you're a step ahead already by actually having made a conscious specification!Cheers Sam, email on route shortly, and 20mm now penned in (though admittedly the pen is one with an eraser on the other end...)
By total fluke I came across a mid engined setup not too dis-similar to mine earlier, chap helpfully told me he went for a VSA length of 2.2 x track with front roll centre of 50mm, rear of 80mm - reckons it's yet to cause him to fall off a track, so as a starting point, it's as good as any in my mind!
By total fluke I came across a mid engined setup not too dis-similar to mine earlier, chap helpfully told me he went for a VSA length of 2.2 x track with front roll centre of 50mm, rear of 80mm - reckons it's yet to cause him to fall off a track, so as a starting point, it's as good as any in my mind!
Gassing Station | Suspension, Brakes & Tyres | Top of Page | What's New | My Stuff