LocostUSA.com

So my suspension was designed around a wheel with a 40mm offset but my new wheels have a 25mm offset. Should try to get it back or will it not affect handing much? It looks like the scrub radius changes but is this an issue from a real life stand point??

Mo Susp Design-1 (18OCT16)

_________________
My build: http://www.locostusa.com/forums/viewtop ... 18&t=13096


You can build the most awesome thing in the world but at some point, an 80yr old man in a crx is probably going kick your butt on the track... don't ask me how I know.

This calls for an opinion answer, and mine is No, not an issue.

I'd sort the comedy roll centre movement before worrying about the wheel offset, if I were you.

how so?

_________________
My build: http://www.locostusa.com/forums/viewtop ... 18&t=13096


You can build the most awesome thing in the world but at some point, an 80yr old man in a crx is probably going kick your butt on the track... don't ask me how I know.

You should aim to ensure that the roll centre stays pretty much fixed (to within a small fraction of an inch, ideally, both laterally and vertically) relative to the sprung mass, no matter what the wheel movement is in bump or roll.

Dynamic weight transfer is partly dependent on the couple between the sprung mass and the geometric roll centre, so if the RC is moving around all over the place, the diagonal weight transfer will be shifting about correspondingly, and will tend to make the car feel squirrelly and uncertain (unless you use the brute force and ignorance approach of simply springing it so stiffly that the suspension barely moves).

I don't entirely trust Vsusp, and I don't have the time to plug your geometry onto a proper analysis, but at the moment your RC seems to be leaping around like a pea in a shell game.

"I don't entirely trust Vsusp, and I don't have the time to plug your geometry onto a proper analysis,.................."

........ but taking the Vsusp result at face value, you're geometry is an abortion.

Sam_68, stuff the attitude. It's just not a good way to educate anyone.

By my look the roll center moves 5 millimeters vertically for 3 degrees of body roll.

Explain why you consider lateral movement of the roll center an issue?

_________________
Marcus Barrow - Car9 an open design community supported sports car for home builders!
SketchUp collection for LocostUSA: "Dream it, Build it, Drive it!"
Car9 Roadster information - models, drawings, resources etc.

Bobber... no more clinic visits.

I think he was talking to me. On that, I don't know how to say this but my buddies and I have some very well calibrated butt gforce sensors and they disagree.

_________________
My build: http://www.locostusa.com/forums/viewtop ... 18&t=13096


You can build the most awesome thing in the world but at some point, an 80yr old man in a crx is probably going kick your butt on the track... don't ask me how I know.

Who says I have a responsibility to educate? I have a long memory, and people who treat me as an a$$hole get treated as assholes themselves.

If mjalaly's friends are the same ones who told him that his simple, curved-tube spaceframe has a torsional stiffness in excess of 38,000lb.ft/degree, then their butts ought to be well calibrated, admittedly, 'cos that's where they keep their brains. But I rather think that it's more likely to be a case of Colin Chapman's assertion that 'any suspension will work if you don't let it' (ie. make it stiff enough, and it won't move enough to get you into trouble).



Short answer to your question, though, is that cars have two ends, but only one sprung mass.

You therefore need to think in three dimensions, and if you need to ask the question, I'm probably going to struggle to explain it to you without resorting to diagrams, but I'll have a go:

Easiest way to visualise it is to imagine the car static, with a magical piece of string attached to the CoG of the sprung mass, which can pass unimpeded through the structure. Pull the string directly to one side, and you're simulating roll force; directly fore or aft and you'd be simulating braking or acceleration. With me so far?

When you pull on this piece of string to simulate roll, the car does NOT try to roll around the front roll centre or the rear roll centre. It tries to roll around the geometric roll AXIS - the line that connects the two. I say 'tries', because it won't actually succeed... the differing F:R roll resistances from the springs will resist it to differing degrees, to force it into an attitude of 'skewed roll', with a component of pitch, but let's not muddy the water with that just yet.

If you give it a bit of thought, you'll appreciate that any deviation of the roll axis from a straight, horizontal fore-and-aft alignment will geometrically try to force the car to try to 'lean' on one corner as it rolls. If you give it a bit more thought in terms of a very large lateral movement of the roll centre at one end, you'll hopefully be able to appreciate that it will introduce a large fore-aft component in the weight transfer.

Not necessarily a problem if the relationship of the roll axis to the sprung mass is fixed (ie. roll centres both ends fixed relative to sprung mass), and indeed you can use it as a tool to tune understeer:oversteer balance. But hopefully you'll be able to appreciate that if the roll axis is swinging around wildly in whatever orientation, it will be playing merry hell with the diagonal weight transfer as it does so.


Anyone who tells you that lateral roll centre location is not important is thinking in only two dimensions.

This post was established purely to address the offset issue and to sum that up, i didn't feel a difference once i switched the wheels but I haven't tracked the car recently to see if there were changes at the "limit"

I keep getting pulled into the trolling, dumb me. Who treated you badly? if it was me I am truly sorry about that and if you feel so strongly about what was said, just ignore post from them. No need to add to the conversation.

Conversations should be more like this...
http://www.fsae.com/forums/showthread.p ... t-question

this is also good
https://docs.google.com/file/d/0B3NuTXn ... JMcFU/edit

and a few notes and I am not trying to be mean.

1) "Comical" is just an ignorant comment, it is way more complicated than what vsusp shows and i agree on that so if it is way more complicated, why even make that statement. horizonjob is right... don't make snarky remarks without saying why. Are you on this forum to learn, teach or just troll? If it is to teach, you could use some refinement (I know I was a professor for 3yrs).
2) Compromises? Car intention? Drive skill? What are we designing here?
3) I don't know where my CG is... nor do most people
4) My suspension was designed with the help from Dennis at Palatov. If it is good enough for his Pikes Peak winning cars its is good enough for me.
5) I am not an F1 driver and I can only judge suspension design based on my experience.
6) I have never been able to drive a car with a non-moving RC to know how it affect's things in real life. To actually do this I would need the same exact car with both designs and pick the one that is best. Just because paper says its is right, doesn't mean it is.
7) My car is actually really comfy to drive and the suspension works as design so i would think Mr Chapman would disagree with you here.
8.) I am also still confused on why with all of this knowledge you haven't built a car that would just kill every car out there?
9) You should also look up Cheapracer... you guys are really similar.

My "buddies" are NASA instructors and racers who have joy in driving my car, then their actual race cars, then my car to figure out why mine is better. When I drive their cars, i think theirs are better. I guess it comes down to the drivers style and maybe what we had for lunch.

now you can have the last word and I will unsubscribe

_________________
My build: http://www.locostusa.com/forums/viewtop ... 18&t=13096


You can build the most awesome thing in the world but at some point, an 80yr old man in a crx is probably going kick your butt on the track... don't ask me how I know.

OK, a couple of serious comments:

I won't speak ill of the dead - he did a lot of great work - but both the way Bill Mitchell's software calculates roll centres, and a lot of what he believed about force based roll centres is questionable in terms of real-world applicability. His software, certainly in terms of the last version I examined closely, assumes that the chassis rolls about a fixed point at ground level, regardless of where the geometric roll centres were, or what the springs were doing to resist, front and rear, which is just plain nonsense. Maybe he fixed it in later versions, I don't know - I'd written it of as a waste of money by then.

3) It in terms of the issues of roll axis movement, it doesn't much matter if you don't know where you CG is. The CG of the sprung mass doesn't move significantly - even the driver bit of it should be strapped firmly into his or her seat. If your roll centres are fixed, the roll axis will be fixed relative to the CG, wherever it is, so the weight transfer will be predictable, even if it's predictably bad.

6) I have (see point 8, below). I have also driven a car (the TVR Griffith 500) where I eventually tracked a particularly lethal and unpredictable handling trait down to a rear roll centre that moved in a way not dissimilar to yours: under very specific circumstances (a combination of squat and roll; usually as you were accelerating and pulling out to overtake across the crest of a camber) the roll centre would leap from low down on one side of the car to somewhat higher up on the other side of the car, instantaneously. It didn't happen often, but when it did, the results were not nice.

8 ) Maybe it's because every other pro race designer I know takes very great pains over roll centre location, too? I'd go so far as to say that it's one of the fundamentals, so you won't come across many competitive race cars with compliant suspension that don't have close control of roll centre location. Doing it right when everyone else is also doing it right won't give you much of an advantage.

Of course, with very stiffly sprung wings-and-slicks single seaters, the Chapman dictum comes into play and other factors can take precedence.

Sorry to bust in here but this is exactly what I'm looking for. What was the geometry of the TVR that resulted in this erratic behaviour? Can I reliably pull something off the internet for this geometry? Actually, locations of instant centers and control arm lengths at ride height should get me what I need to know. Sam, can I impose on you for this info? How close is it to Mo's?
(Sam knows about my "suspension machine" and I'm still on my witch hunt.)

"6) I have (see point 8, below). I have also driven a car (the TVR Griffith 500) where I eventually tracked a particularly lethal and unpredictable handling trait down to a rear roll centre that moved in a way not dissimilar to yours: under very specific circumstances (a combination of squat and roll; usually as you were accelerating and pulling out to overtake across the crest of a camber) the roll centre would leap from low down on one side of the car to somewhat higher up on the other side of the car, instantaneously. It didn't happen often, but when it did, the results were not nice."

I am going to start a thread where people can input their suspension numbers into VSusp so that they can be compared.... granted as Sam said this is a bit more complex but at least it gives us a start point.

_________________
My build: http://www.locostusa.com/forums/viewtop ... 18&t=13096


You can build the most awesome thing in the world but at some point, an 80yr old man in a crx is probably going kick your butt on the track... don't ask me how I know.

I no longer have it for my old car, I'm afraid - I lost it in a hard disc failure, some years ago.

You might try asking the guy who has prepared the CAD model on this thread on Pistonheads

It's something I've been vaguely meaning to do myself, but never got round to it - I wanted to recreate my previous work and see if I came up with the same result; TVR chassis are notoriously inaccurate, geometrically (as you'll see from the discussions on the laser scan data on the link above), so just 'cos it caused problems on my individual car doesn't necessarily mean that it will be to the same degree on all of them.

In my case, the geometry was such that under certain ranges of movement, the roll centre would drift off to one side and downward, but then go 'over centre' and magically leap way from one side of the car to the other, moving several inches vertically as it did so. The practical outcome when it happened was that (if you survived) you needed to buy a new pair of trousers. I sold the car PDQ, once I'd figured out what was going on, as I couldn't be bothered re-engineering it (there were other aspects of the chassis design that made my skin crawl, so I kinda decided that it wasn't the car for me, anyway, and went back to my Lotuses).