Sam_68
|
| posted on 2/2/14 at 10:38 PM |
|
|
quote:
Originally posted by 43655
I'm having a crack at the roll centre thing, which way should it be constrained, up/down or laterally?
Both!!
Ideally, the geometric roll centre shouldn't move more than a few millimetres (I'd say 5mm at worst should be achievable, 2mm is
better...), relative to the sprung mass, either vertically or laterally.
I'm not saying that achieving this level of constraint on your roll centres is easy, mind you - even with software like SusProg, expect
to spend several days tearing your hair out as you try to square the circle of conflicting geometric and practical requirements, and with something as
basic as Vsusp, it may well take weeks of trial and error and turn you prematurely grey...
quote:
Originally posted by 43655
you say relative to the sprung mass, i can only vaguely guess at how that will be laid out. From what i've read elsewhere it sounds as though
the roll centre should be pretty close to the centre of mass. Yet i've also see cars set with roll centres just above the ground. I don't
know...
Your roll centre shouldn't be close to the centre of mass (AKA Centre of Gravity) of the sprung mass - that would place it far too high
and result in excessive jacking.
It will, as you say, be just above ground level and the front RC will need to be slightly lower than the rear, to give slight downward
inclination to your roll axis; you might want to start with figures of, say, 50mm high at the front and 65-75mm high at the rear.
For what it's worth, your centre of gravity is likely to be around 450mm above ground level (obviously design dependent), so your roll centres
will be nothing like as high.
There is a school of thought that says the roll axis should be parallel (or almost parallel) to an imaginary line called the Mass Centroid
Axis, but personally I reckon that's b*ll*cks: with any reasonably stiff chassis, you've only got one centre of gravity and all forces
acting on the sprung mass can be assumed to be acting on that one point. You can attempt to calculate the correct roll axis inclination, if you have
all relevant data on masses, spring/ARB resistances, tyre slip angles, etc., and the sums for doing so have nothing to do with the 'mass
centroid axis' - it's just a coincidence that they usually run not far off parallel - but the calculations are monstrously complicated so
you're better off just sticking to empirically proven rules of thumb using figures around about those I've given above.
...But what I'm trying to say is that as the sprung mass (chassis) moves up and down in pitch and roll, the roll centre should stay in the same
position relative to it... so if you plot pure roll for the chassis, the roll centre doesn't move at all. If you plot pure bump, so that
the chassis moves down relative to the ground by 50mm, then the roll centre should also move down toward the ground by 50mm.
Or in other words if the static RC is on the centreline of the sprung mass, 350mm (or whatever) below the CoG, it should remain on the centreline of
the sprung mass, 350mm below the CoG, no matter how the wheels are moving up and down.
|
|
|
|
|
43655
|
| posted on 14/2/14 at 09:20 PM |
|
|
right i've been spending some time on Vsusp (i got a trial copy of SusProg but got nowhere) and the best i've got so far is 14mm/degree
lateral movement and decent
http://tinyurl.com/q8zy5pp however this isn't really particularly practical due to how narrow I've had to make the chassis
http://tinyurl.com/phtq3mz is a bit more realistic although moves a bit further. Am I at least i the right kind of ballpark with the roll centre /
geometry?
anyway as for the chassis
is V5, widened at the back to fit the huge engine/gearbox combo. hoping to get hold of both before tooo long so i can get proper measurements but
it's not far off
overview drawing, it's pretty big i think
http://i1020.photobucket.com/albums/af322/43655/Chassis5140214.png
all comment welcome, you've been real helpful so far 
|
|
|
Sam_68
|
| posted on 14/2/14 at 10:50 PM |
|
|
Sorry, the following post is a bit long-winded:
quote:
Originally posted by 43655
right i've been spending some time on Vsusp (i got a trial copy of SusProg but got nowhere) and the best i've got so far is 14mm/degree
lateral movement and decent ____
I think you must have got distracted and left this sentence unfinished - decent what??
quote:
Originally posted by 43655
http://tinyurl.com/q8zy5pp however this isn't really particularly practical due to how narrow I've had to make the chassis
http://tinyurl.com/phtq3mz is a bit more realistic although moves a bit further. Am I at least i the right kind of ballpark with the roll centre /
geometry?
The first link (which I guess must be for your rear suspension?) isn't working for me. 
The front suspension, I've only had a very brief look, 'cos I'm really busy working this weekend, but:
* Roll centre control is not brilliant, but not bad either: to give credit where it's due, I've seen a lot worse, and plenty of people
have built cars without even thinking about it, let alone putting in the effort that you have. BUT... I'm sure it could still be improved if you
tried hard enough. 
The roll centre is also a bit higher (at 97mm) than I'd be aiming for, too - I'd be aiming for not higher than 75mm, and preferably
50-55mm.
Like I said, Vsusp is going to be a pain in the ass, 'cos it's all trial and error rather than being able to calculate appropriate
geometry for you, as Susprog can. You won't get anywhere with the trial version of Susprog, though: from memory, you're stuck with
the fixed upright geometry that they give you as an example, so it's no use at all, other than being able to learn the program's
capabilities.
* Your camber recovery in roll is excellent. In fact it is too good!: you're showing -0.168 degrees at 4 degrees roll on the heavily
loaded outside tire (which is the one you need to worry about). In very basic terms (and ignoring a lot of 'ifs' and 'buts' )
this means that your grip when cornering would be excellent. BUT...
* Your problem with this geometry is that the camber control in bump is lousy. It's showing -6.85 degrees at 100mm bump, which is WAY too
much! In very basic terms again (and again ignoring all the 'ifs' and 'buts' ), this means that your grip when braking (when
the car dives onto its nose, of course) will be seriously compromised.
The problem is that camber control is always going to be a compromise: if you have excellent camber control in roll (as you have), then you'll
get terrible camber control in bump, and vice versa. Or, to put it another way, if you aim for excellent grip in corners, you'll get poor grip
when braking, or vice versa.
...So, what you need to be aiming for is average camber control in both. I'd be aiming to adjust the figures so that your camber control
in roll isn't so good, but your camber control in bump is much better than you have at present.
I tend to bias front suspension slightly in favour of better camber control in bump at the expense of some camber control in roll. This
is because the forward weight transfer means that your fronts are doing most of the work in braking, so you need to give them the best chance they
can, and a bit of progressive understeer caused by the camber in roll being less perfect is no bad thing: it’s a safe characteristic. Also, you tend
to get a bit more camber recovery than the geometry calculators suggest, at the front, because diagonal weight transfer means that you get a bit of
negative camber from bump deflection added to the roll camber.
At the rear, I tend to aim for a fairly equal compromise between camber in roll and bump, ‘cos you don’t want the roll camber progression to be
worse than the front (which would risk roll oversteer tendencies), but equally you don’t want to compromise the camber control in bump too
much, because it degrades traction when you’re accelerating.
Try to aim for closer to the figures I gave on this thread (my post of 9/2/14 at 6:17pm), when I tried to offer a step-by-step guide to designing the
geometry. We’re talking about a three wheeler in this instance, but I’d hold to the same target figures for camber gain and roll centre height for
your design:
http://www.locostbuilders.co.uk/forum/3/viewthread.php?tid=188760
quote:
Originally posted by 43655
anyway as for the chassis...
It's difficult to judge what's going on at the front, but the chassis certainly seems to be moving in the right direction, too. Any chance
of a front 3/4 view?
[Edited on 14/2/14 by Sam_68]
|
|
|
43655
|
| posted on 15/2/14 at 02:32 PM |
|
|
first link was just another 'version' of Vsusp setup
I'll keep playing with it to try and get better geometry.
Must admit i'd completely ignored the camber gain aspect, and yeah i understand how crap it would be in a straight line.
It seems though (trial and error yeah) that the best way to constrain the roll centre is by having the instant centres fairly close together.
they're inboard of the uprights anyway. This means that there's a fair bit of camber change through travel becuase the upper arms are
angled down and not very long.
Got no idea how i'll do the back
hate to ask as you've been dead helpful already, but do you have any pictures of your setup?
Amusingly at 100mm bump i'll be grounded on the chassis!
Slightly tangential again, but am i right in thinking that castor can compensate for less camber gain due to the way the wheel cambers when it turns
(being when you actually need the camber)?
Aye i read through that thread recently, bizarre setup but hey
front 3/4 ish view
& rear
wishbone mounting tubes are ~275 and ~325 from centreline here.
Alarmingly, it's pretty heavy! 139kg chassis and 561kg overall (engine 190, box 77) wheel/tyre comes in at 21kg each which seems a bit
excessive, but i've not weighted the actual alloys, but i think they're around 8.5kg.
Just some musings. maybe the huge 2.7tt isn't ideal
[Edited on 15/2/14 by 43655]
|
|
|
Sam_68
|
| posted on 15/2/14 at 03:25 PM |
|
|
quote:
Originally posted by 43655
Slightly tangential again, but am i right in thinking that castor can compensate for less camber gain due to the way the wheel cambers when it turns
(being when you actually need the camber)?
Yes, castor gives beneficial camber gain in cornering. It also causes a degree of weight jacking by physically trying to lift the corner of the car
when you turn the wheel (which is what causes the self-centering effect, of course).
I vary my geometries according to each specific design - there isn't one 'magic' set of numbers that you can apply to any car; it
depends on wheelbase/track dimensions, weight distribution and a number of other factors - so I have a number of different geometry files that
I've worked out on Susprog, but no, I'm afraid placing that sort of information in the public domain would be a step too far - I'm
not willing to share all my secrets!
|
|
|
43655
|
| posted on 15/2/14 at 03:34 PM |
|
|
ha ha that's understandable. out of curiosity, do you do this professionally?
|
|
|
Sam_68
|
| posted on 15/2/14 at 04:05 PM |
|
|
quote:
Originally posted by 43655
out of curiosity, do you do this professionally?
Sometimes, and semi-professionally, you might say.
I'm actually an Architect by training and main profession, but yes, I have made part of my living from chassis and suspension design
& set-up in the dim and distant past.
|
|
|
43655
|
| posted on 2/3/14 at 07:24 PM |
|
|
Geometry is pretty good now.
I have opted for Delrin solid bushes and a stock ball joint in the lower wishbones, the uppers will have rose joints and a ball joint to allow some
tweaking in caber and caster.
Currently lateral roll centre movement is ~17mm/degree, to 52mm at 3 degrees.
Any significant improvement on that is going to be compromising geometry and layout of the chassis.
I'd say that's pretty good?
I'm pretty happy with my upright design, but no FEA or anything done, too advanced for me.
Will use standard bolt-on hubs, heavy but simple.
As for the chassis itself, as you may notice I've 'laddered' the sides. These lengths of box section are what the shell will be
welded to, and the aluminium sideskirts bolted to. This was something that's bugged me for ages, as to how the bodywork will be attached. The
ladder section is a bit higher than i hoped, but it makes for a stronger chassis, as the 100mm narrower chassis (within the cab anyway) means a
structural tunnel is a no-go.
So, opinions please on version 5.4
|
|
|
Sam_68
|
| posted on 2/3/14 at 07:38 PM |
|
|
quote:
Originally posted by 43655Currently lateral roll centre movement is ~17mm/degree, to 52mm at 3 degrees.
Any significant improvement on that is going to be compromising geometry and layout of the chassis.
I'd say that's pretty good?
Meh. It's getting there, but to give you some idea, I've been working on a geometry for a 'Seven' type car this afternoon,
where total roll centre movement over 3 degrees is 0.7mm.
|
|
|
Doctor Derek Doctors
|
| posted on 2/3/14 at 07:48 PM |
|
|
You've probably already answered this but how are you intending to get the engine in or out?
Designer and Supplier of the T89 Designs - Single Seater Locost. Build you own Single Seater Racecar for ~£5k.
Plans and Drawings available, U2U or e-mail for details.
Available Now: The Sports Racer Add-On pack, Build a full bodied Sports Racer for Trackdays, Sprints and Racing.
www.t89.co.uk
www.racecarwings.co.uk
callan@t89.co.uk
|
NOTE:This user is registered as a LocostBuilders trader and may offer commercial services to other users
|
43655
|
| posted on 2/3/14 at 08:12 PM |
|
|
goddammit i thought i was doing well
ha ha well, the top brace bit will be bolt-in, just couldn't be bothered to model it so.
having second thoughts about the brace over the gearbox at the moment.
hmm. arse
|
|
|
Sam_68
|
| posted on 2/3/14 at 08:22 PM |
|
|
quote:
Originally posted by 43655
goddammit i thought i was doing well
Don't worry - I'm only teasing!
Lateral roll centre movement of that amount is fine, and much better than many designs achieve (in fact vertical movement of that amount would be
tolerable, and it's vertical movement that's more important).
|
|
|
43655
|
| posted on 2/3/14 at 08:53 PM |
|
|
Good, I may well leave it as is then. vertically the roll centre is pretty much fixed at 60mm
How much higher should the rear be?
And is it sinful to have wishbones sloping down to the wheel (to raise the RC?) not looked at the rear end geometry yet
|
|
|
Sam_68
|
| posted on 3/3/14 at 10:53 PM |
|
|
quote:
Originally posted by 43655
How much higher should the rear be?
That's very much open to debate, but as a starting point I'd suggest maybe 50-60mm higher than the front.
quote:
Originally posted by 43655
And is it sinful to have wishbones sloping down to the wheel (to raise the RC?) not looked at the rear end geometry yet
It's not usually necessary, and high roll centres promote jacking. You'll see plenty of modern F1 cars with downward sloping wishbones,
though, so it's not an unbreakable rule.
|
|
|
43655
|
| posted on 4/3/14 at 05:53 PM |
|
|
quote:
Originally posted by Sam_68
quote:
Originally posted by 43655
How much higher should the rear be?
That's very much open to debate, but as a starting point I'd suggest maybe 50-60mm higher than the front.
quote:
Originally posted by 43655
And is it sinful to have wishbones sloping down to the wheel (to raise the RC?) not looked at the rear end geometry yet
It's not usually necessary, and high roll centres promote jacking. You'll see plenty of modern F1 cars with downward sloping wishbones,
though, so it's not an unbreakable rule.
That's pretty much what I've seen elsewere.
Only considering it for the rear end to get the higher RC. F1 barely use suspension tough so geometry isn't the highest priority from what
i've read
|
|
|
Sam_68
|
| posted on 4/3/14 at 07:37 PM |
|
|
quote:
Only considering it for the rear end to get the higher RC. F1 barely use suspension tough so geometry isn't the highest priority from what
i've read
True... Colin Chapman once said something like "Any suspension, no matter how poorly designed, can be made to work reasonably well if you just
stop it from moving". Though interestingly, if you compare the on-board camera footage from a modern F1 car with one from 15 or 20 years aho,
you'll see that they are a lot more compliant now than they used to be - although the compliance all seems to be in bump, at the front,
with a very high degree of stiffness in roll.
Just because F1 cars have downward sloping wishbones doesn't necessarily mean they have a high roll centre, of course (they don't, as best
we can make out)!
...and at risk of stating the bleedin' obvious (though it's a point that many people apparently don't appreciate), you can't
get the jacking effect that is the principal objection to high roll centres if you have nil-droop or very limited droop on your suspension.
But we digress...
|
|
|
43655
|
| posted on 12/3/14 at 05:24 PM |
|
|
quote:
Originally posted by Sam_68
quote:
Only considering it for the rear end to get the higher RC. F1 barely use suspension tough so geometry isn't the highest priority from what
i've read
True... Colin Chapman once said something like "Any suspension, no matter how poorly designed, can be made to work reasonably well if you just
stop it from moving". Though interestingly, if you compare the on-board camera footage from a modern F1 car with one from 15 or 20 years aho,
you'll see that they are a lot more compliant now than they used to be - although the compliance all seems to be in bump, at the front,
with a very high degree of stiffness in roll.
Just because F1 cars have downward sloping wishbones doesn't necessarily mean they have a high roll centre, of course (they don't, as best
we can make out)!
...and at risk of stating the bleedin' obvious (though it's a point that many people apparently don't appreciate), you can't
get the jacking effect that is the principal objection to high roll centres if you have nil-droop or very limited droop on your suspension.
But we digress...
I have read about people limiting or eliminating droop, but i can't get my head around jacking theory at the moment.
HOWEVER, got a f***ing sweet roll centre migration, +-4mm lateral, +0.5mm vertical, that's over 3 degrees of roll. I just hope the rest of the
geometry is still healthy!
still a long way to go, but i'm chuffed
|
|
|
43655
|
| posted on 24/7/14 at 12:37 PM |
|
|
Updated version, steel ordered. still heavier than I want.
No rear suspension yet, and the supporting tubes to the roof line will probably depend on that too.
The weird chassis on the roof is because I think I'll have to extend the doors to include part of the roof, almost Ford-GT-esque due to the
'side skirts' being like 360mm (for reference the doors used to sit 30mm from the bottom of the chassis) but that will depend on how bad
access is during the build.
Hopefully nothing I'm doing seriously wrong though!
|
|
|
