Kinematics - Peer Review Plea

[markocosic]

Hi all,



A plea for help here, or some peer-review at least! We're an entirely newbie team and the kinematics are all on my head – a quick lookover by some more experienced eyes would be much appreciated. Most of this mid-section is the numbers – my main questions/concerns are at the bottom.


Vehicle in general:
230kg and 70kg driver, distributed 45-55 front-rear
20x6.5-13 and 20x8.0-13 "FSAE Goodyears"� front/rear
1600mm wheelbase, 1200mm front track, 1100mm rear track
Steel tube spaceframe, 600cc '05 R6 engine, spool rear-end

Suspension:
Double-wishbones (a-arms) all-round
Pushrod/rocker spring/damper setup all round
Target 1 deg/g roll

Steering:
210 degrees lock-lock with full lock being 30 degrees on the outer wheel
Lower-rear mounted rack, steering arms on uprights pointing at rear axle centreline
(whatever % Ackerman this comes to under your definitions)

Unknowns/Uncertains:
CoG height – assumed to be around the 300mm mark
Dynamics – springs, dampers and ARBs TBC once kinematics confirmed.

Overall:
Our manufacturing capabilities are limited (domestic garage hand/power tools plus elderly manual lathe/mill and TIG-for-backhanders outside), and it'd be fair to say we're cribbing a lot from Pat's technical tips for FSG and the '05 Brisbane/Queensland machine. (not particularly pretty, but then it's not a manufacturing monster and is damn capable too) Input from past and present members of that team particularly welcome!

Kinematics:
There seem to be four schools of thought here: 1) relatively soft setup, short VSALs and drive it cautiously; 2) relatively-hard-as-nails setup, long VSALs, big wedge of static camber and keep it on the edge of 2/3 wheels; 3) so inaccurately manufactured and so compliant that there are no kinematics worth discussing; 4) ˜the specials' along the lines of UWA.

(1) is my philosophy, thinking it should be the least sensitive to driver error, manufacturing error and track conditions even if not the ˜optimum'.



Positions of stuff:

Co-ordinate system: 0,0,0 is the ground at the front axle centreline, positives are right, backwards and upwards. X, Y, Z are lateral, vertical, longitudinal respectively. (SolidWorks convention) Longitudinal locations of the wishbones are arbitrary for now until chassis is further detailed, but should not affect geometry.

Front (XYZ co-ords: RHS shown, LHS mirrored)
LBJ: 565, 130, -12
UBJ: 565, 340, 25
Front Lower: 175, 116, -100
Rear Lower: 175, 116, 100
Front Upper: 270, 258, -100
Rear Upper: 270, 258, 100

Rear (XYZ co-ords: RHS shown, LHS mirrored)
LBJ: 500, 130, 1600
UBJ: 500, 340, 1600
Front Lower: 175, 119.5, 1500
Rear Lower: 175, 119.5, 1700
Front Upper: 270, 269.5, 1500
Rear Upper: 270, 269.5, 1700

Steering Rack (for straight-ahead position, RHS shown, LHS mirrored)
BJ on Upright: 550, 155, 30
BJ on Rack: 180, 130.6, 27.5



Named numbers for stuff:

Front:
KPI = 0
Caster = 10 degrees
Scrub Radius (KPI Trail?) = 35mm
Trail (Caster Trail?) = 35mm

Upper wishbone ~305mm long, lower wishbone ~390mm long

VSALs & RCs (geometric method)
*Neutral: 908mm long, 99mm off ground (RC = 66mm off ground)
*-15mm ride: 856mm long, 70mm off ground (RC = 49mm off ground)
*-15mm ride, 1 deg roll: 892&822mm long, 74&66mm off ground (RC = 49mm off ground)
*-25mm ride: 824mm long, 52mm off ground (RC = 38mm off ground)
*-25mm ride: 1.5 deg roll: 875&775mm long, 57&47mm off ground (RC = 38mm off ground)
*+15mm ride: 962mm long, 131mm off ground (RC = 83mm off ground)
*+15mm ride: 1 deg roll: 1002&924mm long, 139&124mm off ground (RC = 83mm off ground)

Rollcentres stay "nailed"� – lateral movement within 30mm, vertical within 10mm (relative to chassis) The shorter/lower of the two VSALs given when rolling is the loaded side.

Rear:
No steering.

VSALs & RCs (geometric method)
*Neutral: 823mm long, 105mm off ground (RC = 71mm off ground)
*-15mm ride: 762mm long, 76mm off ground (RC = 54mm off ground)
*-15mm ride, 1 deg roll: 800&726mm long, 80&72mm off ground (RC = 55mm off ground)
*-25mm ride: 724mm long, 58mm off ground (RC = 45mm off ground)
*-25mm ride, 1.5 deg roll: 779&674mm long, 64&54mm off ground (RC = 45mm off ground)
*+15mm ride: 888mm long, 140mm off ground (RC = 87mm off ground)
*+15mm ride, 1 deg roll: 933&845mm long, 149&131mm off ground (RC = 87mm off ground)

Much the same as the front, though slightly higher throughout and proportionately shorter VSALs due to narrower track and no camber-adding steering angles.


My questions/concerns:
For the spool rear-end to work we need to unload that inside rear – do you think that the 0/10/35/35 on the steering geometry will be sufficient to do this? Too much/airborne rear wheel? Too high a steering wheel force? It is our intention to adjust only toe and static camber on the vehicle from a kinematics point of view – any errors elsewhere will have to be tuned out/masked by the springing and damping so it's imperative this one turns out approximately right.

The short VSALs and "more than a gnat's bollock above the ground"� rollcentres: I'd like to get some geometric weight-transfer in there (and build some camber-thrust when braking for a bend), but does this jump out as asking for trouble with jacking and cresting bumpy tracks?

Steering – done based on lock requirements given by others in the group and a nagging feeling says too much lock&applied too fast, but I've nothing to base this against. Degrees lock-lock, angle of the outer wheel and wheelbase of a few of your cars would be much appreciated!



In a nutshell: Could you take a look please, and if you see anything hideously wrong please flag it. "80% right"� would be more than good enough for our purposes this year – we need as much time as possible on detailing the beast and provided nothing is too wrong with the thinking would like to crack on!


Cheers,

[markocosic]

Hi all,



A plea for help here, or some peer-review at least! We're an entirely newbie team and the kinematics are all on my head – a quick lookover by some more experienced eyes would be much appreciated. Most of this mid-section is the numbers – my main questions/concerns are at the bottom.


Vehicle in general:
230kg and 70kg driver, distributed 45-55 front-rear
20x6.5-13 and 20x8.0-13 "FSAE Goodyears"� front/rear
1600mm wheelbase, 1200mm front track, 1100mm rear track
Steel tube spaceframe, 600cc '05 R6 engine, spool rear-end

Suspension:
Double-wishbones (a-arms) all-round
Pushrod/rocker spring/damper setup all round
Target 1 deg/g roll

Steering:
210 degrees lock-lock with full lock being 30 degrees on the outer wheel
Lower-rear mounted rack, steering arms on uprights pointing at rear axle centreline
(whatever % Ackerman this comes to under your definitions)

Unknowns/Uncertains:
CoG height – assumed to be around the 300mm mark
Dynamics – springs, dampers and ARBs TBC once kinematics confirmed.

Overall:
Our manufacturing capabilities are limited (domestic garage hand/power tools plus elderly manual lathe/mill and TIG-for-backhanders outside), and it'd be fair to say we're cribbing a lot from Pat's technical tips for FSG and the '05 Brisbane/Queensland machine. (not particularly pretty, but then it's not a manufacturing monster and is damn capable too) Input from past and present members of that team particularly welcome!

Kinematics:
There seem to be four schools of thought here: 1) relatively soft setup, short VSALs and drive it cautiously; 2) relatively-hard-as-nails setup, long VSALs, big wedge of static camber and keep it on the edge of 2/3 wheels; 3) so inaccurately manufactured and so compliant that there are no kinematics worth discussing; 4) ˜the specials' along the lines of UWA.

(1) is my philosophy, thinking it should be the least sensitive to driver error, manufacturing error and track conditions even if not the ˜optimum'.



Positions of stuff:

Co-ordinate system: 0,0,0 is the ground at the front axle centreline, positives are right, backwards and upwards. X, Y, Z are lateral, vertical, longitudinal respectively. (SolidWorks convention) Longitudinal locations of the wishbones are arbitrary for now until chassis is further detailed, but should not affect geometry.

Front (XYZ co-ords: RHS shown, LHS mirrored)
LBJ: 565, 130, -12
UBJ: 565, 340, 25
Front Lower: 175, 116, -100
Rear Lower: 175, 116, 100
Front Upper: 270, 258, -100
Rear Upper: 270, 258, 100

Rear (XYZ co-ords: RHS shown, LHS mirrored)
LBJ: 500, 130, 1600
UBJ: 500, 340, 1600
Front Lower: 175, 119.5, 1500
Rear Lower: 175, 119.5, 1700
Front Upper: 270, 269.5, 1500
Rear Upper: 270, 269.5, 1700

Steering Rack (for straight-ahead position, RHS shown, LHS mirrored)
BJ on Upright: 550, 155, 30
BJ on Rack: 180, 130.6, 27.5



Named numbers for stuff:

Front:
KPI = 0
Caster = 10 degrees
Scrub Radius (KPI Trail?) = 35mm
Trail (Caster Trail?) = 35mm

Upper wishbone ~305mm long, lower wishbone ~390mm long

VSALs & RCs (geometric method)
*Neutral: 908mm long, 99mm off ground (RC = 66mm off ground)
*-15mm ride: 856mm long, 70mm off ground (RC = 49mm off ground)
*-15mm ride, 1 deg roll: 892&822mm long, 74&66mm off ground (RC = 49mm off ground)
*-25mm ride: 824mm long, 52mm off ground (RC = 38mm off ground)
*-25mm ride: 1.5 deg roll: 875&775mm long, 57&47mm off ground (RC = 38mm off ground)
*+15mm ride: 962mm long, 131mm off ground (RC = 83mm off ground)
*+15mm ride: 1 deg roll: 1002&924mm long, 139&124mm off ground (RC = 83mm off ground)

Rollcentres stay "nailed"� – lateral movement within 30mm, vertical within 10mm (relative to chassis) The shorter/lower of the two VSALs given when rolling is the loaded side.

Rear:
No steering.

VSALs & RCs (geometric method)
*Neutral: 823mm long, 105mm off ground (RC = 71mm off ground)
*-15mm ride: 762mm long, 76mm off ground (RC = 54mm off ground)
*-15mm ride, 1 deg roll: 800&726mm long, 80&72mm off ground (RC = 55mm off ground)
*-25mm ride: 724mm long, 58mm off ground (RC = 45mm off ground)
*-25mm ride, 1.5 deg roll: 779&674mm long, 64&54mm off ground (RC = 45mm off ground)
*+15mm ride: 888mm long, 140mm off ground (RC = 87mm off ground)
*+15mm ride, 1 deg roll: 933&845mm long, 149&131mm off ground (RC = 87mm off ground)

Much the same as the front, though slightly higher throughout and proportionately shorter VSALs due to narrower track and no camber-adding steering angles.


My questions/concerns:
For the spool rear-end to work we need to unload that inside rear – do you think that the 0/10/35/35 on the steering geometry will be sufficient to do this? Too much/airborne rear wheel? Too high a steering wheel force? It is our intention to adjust only toe and static camber on the vehicle from a kinematics point of view – any errors elsewhere will have to be tuned out/masked by the springing and damping so it's imperative this one turns out approximately right.

The short VSALs and "more than a gnat's bollock above the ground"� rollcentres: I'd like to get some geometric weight-transfer in there (and build some camber-thrust when braking for a bend), but does this jump out as asking for trouble with jacking and cresting bumpy tracks?

Steering – done based on lock requirements given by others in the group and a nagging feeling says too much lock&applied too fast, but I've nothing to base this against. Degrees lock-lock, angle of the outer wheel and wheelbase of a few of your cars would be much appreciated!



In a nutshell: Could you take a look please, and if you see anything hideously wrong please flag it. "80% right"� would be more than good enough for our purposes this year – we need as much time as possible on detailing the beast and provided nothing is too wrong with the thinking would like to crack on!


Cheers,

[J. Vinella]

I didnt have time to run the geometry, but I have one question.

So with short VSALs and 10 deg of caster, on front outside tire what do you think you are going to see in terms of camber at 1.7g @ oh lets model a hair pin, 20-35 degrees of steering lock?

Also to play the design judge devils advocate can you justify having to two different sizes front to back on your tires?

Spools definitely get the job done but boy they are a bitch to push around.

Cheers,

[flavorPacket]

you may want to put your hardpoints in SAE coords...

[js10coastr]

I don't think "manufacturing error" should be a reason you take into the design review.

Take some time to plan it out (don't "just build it"), and you'll be able to accurately manufacture it.

[ben]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by js10coastr:
I don't think "manufacturing error" should be a reason you take into the design review.

Take some time to plan it out (don't "just build it"), and you'll be able to accurately manufacture it. </div></BLOCKQUOTE>

I think it's interesting in the sense that you will always have manufacturing tolerance. If someone came to me with a geometry and then said it was still within performance targets with extremes of an acceptable tolerance I'd be impressed.

I believe the Lotus kinematic software has a tolerance function to automate this type of analysis.

Ben

[Mike Cook]

Marko, everything looks good to me. Most teams are running a lower roll center I think and I don't think you need to run them that high since our cars have high roll stiffness/low roll inertia and will transfer weight pretty quickly elastically. But I also don't think that they will be a problem.


My only advice is to pay attention to your wheel rates. Sure, its better to have small amounts of roll center movement rather than lots- but in the end your always looking to see how these things will affect whats happening at the tire-ground contact patch. If your wheel rates are so non linear that they affect your LLTD a lot, it doesn't matter that your roll center location only moves .01 inch.



If your stuck running a spool, (which I don't think is great), I would check out more what go karts are doing as far as caster. Your going to want to figure out how to get that inside tire to pull up definitly. The thing is that at fsae competitions the course is very tight, and there are many hard braking zones then tight corner then straight away. With this configuration, i'm still a pretty firm believer that the two most important things are to a) be able to manuever tight turns without sacrificing a lot of speed, and b) being able to put a lot of power down at about 15 mph. A spool doesn't help either one of these things out a whole lot.

[Marc Jaxa-Rozen]

You might want to run the calcs for steering effort and diagonal load transfer to make sure you unload the inside rear enough without killing your endurance drivers, I think Erik Zapletal's papers on steering geometry (posted here some time ago) included some simplified equations, otherwise it's not too long to work out. You might be able to get away with using a little less caster/mechanical trail.

Marc Jaxa-Rozen
École Polytechnique de Montréal
http://www.fsae.polymtl.ca

[Ashley Denmead]

The major problem i see is the steering effort.....10 degrees castor is a handful! we tried this on our spool car and could barely steer it around a track let alone a complete endurance (we did only have 180deg steering lock though). i havnt checked your hardpoints but make sure your steering balljoint is inline with your steering rack to get the best mech advantage. ide also add some kpi and reduce the caster by a fair bit....Your car will probably be very sensitive to rear bar and this will be your biggest tuning tool so concentrate on getting the thing drivable with the springs. Agressive driving is also the key.

[Frank]

probably bring the front track into 1150mm

10deg castor is big, 8deg would suffice

VSAL's are good, RC's a bit high, geometric WT is not good

RC's @ 30mm F / 50mm R would suffice

trail and scrub is good, perhaps a little less trail, but good

keep kingpin = camber = about 1.5-2.0 deg equal front and rear

unsprung nat freq's of 2.7-3.0 front 3.6-4.0 rear (assuming no ARB's)

I'd use the same tyres front and rear.. 6.5" GY's or 7" hoosier r25A c2000's

30deg outer wheel and 100% ackerman will be very close to steering inversion.

I'd go for 25deg outer / 35 deg inner (app 100% ackerman) with the 210 deg L-L

keep the front steering arms long!! ~100mm (means a big pinion gear)

jig the pickup joints straight, it's the only parts of the chassis that need be straight

as a newbie team you'll find it hard to get 230kg...

make sure you're dampers can work at the rates you need, adjust rocker ratio so that you can use the minimum damping rates your shocks can provide (~50% bump / 90% rebound).. the classic mistake is damping rates too high because you didn't determine what the shocks can do firstly. Most shocks can't get under 2Ns/mm.

your wishbones are only spread 200mm? that's too little!! 300mm minimum IMO

the hardest manufacturing problem is physically achieving the rocker ratio you wanted, and keeping the shock/rocker/pushrod IN PLANE (at neutral ride height). Jig,jig,jig, and jig some more. Once they're on, it is incredibly hard to measure that you have what you wanted. If you get them wrong by 10%, then you get the rates wrong by 20% (squared ratio remember)

don't sweat anti-dive or squat, i think you'll find you've got a little "pro-squat" there, because loads reacted by chassis (see milikin) don't worry about it, its all good

my best advice is to achieve the 10 goals I ranted here http://fsae.com/eve/forums/a/tpc/f/825607348/m/72410757...10859731#86010859731 (http://fsae.com/eve/forums/a/tpc/f/825607348/m/72410757731?r=86010859731#86010859731)

best of luck!

[Matt N]

It might sound pedantic, but I would suggest this to all suspension designers out there: find the lateral weight transfer equations in RCVD (or elsewhere) and construct a simple spreadsheet where for given static wheel loads (f/r weight dist, car weight) and given general parameters (RC heights, front/rear track widths, anti-roll from the springs and ARBs) you can see the dynamic wheel loads at a projected lateral acceleration, say 1.4 g's.

You will quickly see that staggered tire sizes are really only a benefit (assuming you want the car to be neutral steer, which you do in FSAE) if you run a car which is a lot closer to 40/60. Most FSAE cars are significantly closer to 50/50 which is why I'm assuming most run same size tires.

A continuation: if you are getting an advantage from the staggered size tires, you should realize that a lot of cornering force would be coming from the inside rear tire. Now think about the spool again, and what that does for inside rear cornering force.

I'll take a look at your geometry as soon as I'm working at the kinematics again (later in the week). With all my analysis of the TTC data and Hoosiers specifically, with our relatively limited suspension travel and car roll, the actual kinematics (if they are within the ballpark) are of little importance TO MAXIMUM GRIP, because the tire sensitivity is hardly variable from 0-2 degrees negative camber. I'm starting to think that reducing compliances is much more important than 'ultimate' kinematics. Do make sure that you examine your kinematics in terms of steering effort and feel.

Matt

[Matt N]

Ditto Frank on RC height. Lower them.

Matt

[Frank]

Also..

1/ Your uprights are very Tall.. it's not necc.

2/ Your rear uprights have no "castor" (if you call rear "castor".. "castor"), which means you might have a hard time getting a pushrod past the driveshaft. There is two strategies for this.

A: Give the rear uprights a touch of castor, and perhaps a pinch of driveshaft misalignment, and sneak the pushrod/pullrod past the driveshaft. With this strategy, you are best to put the toelinks about level with the axle. You can achieve no bumpsteer doing this, simply set the toelink projection to intersect the instantaneous center, and to make it perfect put the toelink inboard and outboard nodes on the lines, in front view, projected by the wishbone pickups.
This strategy works to give you a bit more rear weight (valuable), and shocks to the engine (pushrod suspension), with positive castor.

B: The "Triangle Rear Upright". Uses bags of castor(pos or negative), and a toelink in plane with one of the wishbones

3/ The front upright steering arm height is also best near the axle height. But this usually creates problems with steering rack VS drivers lower leg.

4/ You can't analyse your geometry without the axle height. And you can't design good suspension without knowing this accurately. For 7" c2000s it's about 532mm diameter unladen, 4mm static tyre compression, giving axle height of 532/2-4 = 262mm

5/ Back to the "wishbone spread" problem.. probably want more like 350-400 on the rear (if symetric about the axle in top view, the front is more likely to use "sweep" in one direction, to enable ackerman, if you're using sweep in one direction, keep the front nodes at least 250mm in front of the front axle.

6/ Ackerman is not just the stering arms in top view, it's the for aft position of the steering rack also that gives you "progression" and CAN mean no ackerman if incorrect. Warning that Caroll Smith's book has a mistake on this matter... So analyse it yourself and understand it. It's much easier with the rack behind the front axle. With 1600 WB, and an R6 engine, you should be able to get the steering (column)shaft fairly straight with "rack behind the front axle".

[markocosic]

A flurry of cracking posts - the next round might rival Z or Geoff P in scrollbar length, so grab that cup of tea! Guess I'll start with what I'm reasonably certain about and work backwards from there...



Tyre dia

Varies with camber, loading, and decrepitude but I'm working on 250mm loaded radius for the 20x6.5-13 Goodyears and 245mm loaded radius for the 20x8.0-13s.



Manufacturing tolerances

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">I don't think "manufacturing error" should be a reason you take into the design review. </div></BLOCKQUOTE>

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content"> If someone came to me with a geometry and then said it was still within performance targets with extremes of an acceptable tolerance I'd be impressed. </div></BLOCKQUOTE>

No tolerances on designs (or error bars on experimental data) will see you lynched by a grumpy academic or lab tech around here! They have good reason too: everything has a finite accuracy, and a datapoint without bounds is as good as having no datapoint?

There'll be some uncertainty about where your kinematic points are – whether it be a steady-state error from when it was manufactured, a steady-state error from when it was bumped over that pothole, a dynamic error as the slack is taken up in different directions or a dynamic error as components flex doesn't particularly matter.

What you care about in the end is the effect is has on the behaviour of the vehicle and it's therefore legit to bring ˜all' sources of error into your design considerations IMO. I'm not putting money on being able to better +/-1mm manufacturing tolerance over the entire vehicle (in the first year/with limited tooling at least) and that has a real affect on the kinematics. http://fsae.com/groupee_common/emoticons/icon_redface.gif

Aside:
IIRC the Elise has it's suspension pickups accurate to +/- 0.5mm so wouldn't surprise me if the Lotus package was one of the more advanced when it came to manufacturing the beast.

For what it's worth, I found the magic-black-box packages hindered more than they helped when trying to understand suspension from scratch – an equivalent effect on me of handing caveman a 3G phone. Pencil, paper, string and playing with a four-bar-linkage rather than ˜suspension'; then pencil, paper, suspension books and playing with instantaneous centres; then pencil, paper, suspension books and Excel models. (current state of play) Going through the crude methods first helped understand (haha!) what the blankety-blank that magic-black-box package is doing. (something I'd use once established/with at least one car behind us) I'd read ˜academic' Milliken & Gillespie last and the ˜clubman' Smith etc first for similar reasons.



Wishbones:

The +/- 100mm was arbitrary (just to keep the Excel sheet happy) as I was primarily interested in the 2D kinematics, but there's some great feedback there!

Hear you on 300mm+ - the plan was for as close as practical to equilateral on the lower wishbones and keep the uppers approximately the same width. I'm not sure I follow the "250mm rule"� on sweeping the fronts though – why do you say this? Will things not be grand provided the spacing is wide enough and you're not exceeding 90 degrees on any of the lower wishbone angles?



Front uprights:

~90mm from axle centreline to the tops, ~120mm from axle centreline to the bottoms. Drop the LBJ as far as possible to reduce lever arms from the tyre contact patch. Lift the UBJ from the minimum required to miss the wheelbearing because the load-paths into the spaceframe for the wishbone mounts aren't as nice as those for the lower wishbone on the front. (bastard driver's legs!)

I'd wanted to whack the steering arm at axle-level with the tie-rods pointing towards the IC of the suspension (and the inboard pivots on the line between the two wishbone mountings) but this would require an amputee driver at the moment. Going back to uprights I guess ˜shorter' uprights (the top bit at least) and correspondingly ˜shorter' distance between the inboard mountings on the chassis would let me drop the rack enough to get the drivers legs over the top – will have a play, but not /too/ fussed for reasons given in a sec:


Steering:

Good call on the steering arm lengths, the steering being close to inversion and rack position affect on Ackerman Frank – you've gone straight past the questions I asked and into something I wasn't so keen on myself... (the length of the steering arms and the shuffle the rack forward to prevent inversion kludge - hope you're not a design judge! http://fsae.com/groupee_common/emoticons/icon_razz.gif )

The rack I'm designing around is the highest-ratio Titan rack that they do (gives me 71.63mm rack movement per revolution, or 20.9mm movement from centre-full lock for 210 degrees L-L. It ain't cheap, and I'd like a higher ratio, but as of yet I've not come across any other racks that I "like"� the look of. (after something that will not require bodging or "end-supports"� – sources?) Comments on tyre widths, front track, RC height and steering lock are confirming suspicions that I was erring on the side of too tight/too low speed (more later), but for now less lock required will buy me longer steering arms for that given rack.

Where and when do you need the stiffness? On the rear there are big loads and we don't want any toe movement to upset that (fairly small) slip-angle. On the front there's this driver as the most flexible link in the steering system, a tyre/road surface that isn't the most rigid thing in the world, plus steering movement that's an order of magnitude larger than the tyre slip-angle. Hard braking is where you'll load the steering (arms/rack) to the max, but here the arms/rack are fairly perpendicular, only the differences side-side are being transferred to the driver rather than the full force and we're fussing about longitudinal slip rather than lateral/slip angle slip and so a little wavering won't be the end of the world/noticed. Cornering is where the driver will care the most, but the loads are going to be small compared to the loading under braking and the driver/tyre/road still less rigid links.

100mm steering arm length is calling for roughly as much stiffness in toe as in camber and braking? (based on the lever arm from the axle alone as a measure) Steering arm length with the previous setup would have been 42mm. Reduce lock requirements and it ends up ~60mm. The Titan rack would be supported as far as is possible given +/- 21mm of rack movement and is ~22mm in diameter. (ie – though not the prettiest solution, the 22mm rack 0 (inside wheel) or 42mm (outside wheel) from it's supports will take some bending moment at both ends) Unless I'm barking up completely the wrong tree (happens a lot!) I reckon 60mm with that kind of setup should be plenty stiff enough – comments?



Rear uprights:

Frank type "A"� (with the driveshafts slightly below and to the rear of the axle centreline in our instance - doesn't need much in the way of angle to fit the pushrods past) Camber/toe adjustment interlinked.

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content"> You can achieve no bumpsteer doing this, simply set the toelink projection to intersect the instantaneous center, and to make it perfect put the toelink inboard and outboard nodes on the lines, in front view, projected by the wishbone pickups.
</div></BLOCKQUOTE>

Can you really eliminate bump-steer? Sure, there'll be no bump-steer for the default toe/camber setting, but as soon as I play with that toe-link (or upper wishbone) the outboard node will move out of line with the line projected by the wishbone pickups and I'll have bump-steer again? Personally I don't think it's the end of the world given the likely range of toe-settings (yes, there's bump steer, but sufficiently small enough not to worry me given the size of the slip-angles the crossply tyres like to run at – "zero enough for government work"�), but would like to check that I'm not imagining things when I say you will get bump steer if you tweak that toe link.

Frank type "B"� (with the toe-link on the upper wishbone) plane. Can eliminate a chassis hardpoint, can separate camber/toe adjustment, can arrange zero bump-steer.

How (aside from an oversized rod-end in bending http://fsae.com/groupee_common/emoticons/icon_biggrin.gif ) do you provide for camber adjustment (at the upright, so as to avoid altering the kinematic points) whilst maintaining rigidity with both types though?

The Brookes machine shows the only two I can think of:
http://www.cosic.org.uk/galleryv2/main.php?g2_itemId=10018
http://www.cosic.org.uk/galleryv2/main.php?g2_itemId=10009
(I'd be adjusting camber at the upper wishbone at both ends)

The alternative is adjustment with rod-ends at the inboard ends of the wishbones – definitely rigid, but will influence the kinematics and isn't the easiest thing to set repeatably. Thoughts?



Kinematics!

Too much caster. (shade over 3.5 degrees in answer to UoW John) RCs and geometric WT too high. Front track too wide. Too much lock. Unanimous with it too, which says I'm exhibiting muppet-like tendencies. (or making the wrong assumptions depending on who is asking! http://fsae.com/groupee_common/emoticons/icon_wink.gif )

Thanks for the honest critique – numbers can only go so far without info from those who've been there before to bring them into line. I'll adjust the kinematics to suit.

Steering effort would be an effort, but bodies are surprisingly adaptable and seem to manage karts & old land-rovers easily enough – point taken though.

KPI... ...forever associated with all varieties of pleasant anglo-saxon expletives! http://fsae.com/groupee_common/emoticons/icon_redface.gif Zero was on the basis that it won't hurt me, but drowning the effect of a little KPI with caster is ok too. Will ease packaging and manufacture too.



Tyre sizes

Differing sizes front-rear - differing (45-55) mass distribution front-rear, put rubber on the ground to reflect this. Depending on driver style/capability (can they ˜drive' it around hairpins or do they ˜stop-point-scarper') we might even entertain slightly more mass to the rear through playing silly buggers with fuel/electrical/cooling packages.

Could run the same tyres front & rear and compensate for the different loadings/friction coefficients/lateral force slopes (however you want to phrase the front and rear behaving differently for a given % lateral weight transfer) with the springing/damping and ARBs.

Fairly marginal either way (and might not work for a number of reasons such as heat etc) so the plan was to try the "default"� of matching mass to rubber distribution and see where it goes from there (it being straightforward to switch to narrower rims/tyres) in terms of data.

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content"> A continuation: if you are getting an advantage from the staggered size tires, you should realize that a lot of cornering force would be coming from the inside rear tire. Now think about the spool again, and what that does for inside rear cornering force. </div></BLOCKQUOTE>

That's an interesting one! http://fsae.com/groupee_common/emoticons/icon_redface.gif

The differing ˜corner radius' the front and rear wheels track around (fronts tighter) is something that provoked the suggestion of a lots-nothing front-rear distributed front-driver FSAE car (AutoTest/Gymkhana spec Mini...) with fiddle brakes on the front and just enough weight on the rear not to jump on it's nosecone under braking. Difficult to fit within the rules though and in all likelihood a what-if for the pub rather than a viable car LOL.



More tyres:

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content"> With all my analysis of the TTC data and Hoosiers specifically, with our relatively limited suspension travel and car roll, the actual kinematics (if they are within the ballpark) are of little importance TO MAXIMUM GRIP, because the tire sensitivity is hardly variable from 0-2 degrees negative camber. I'm starting to think that reducing compliances is much more important than 'ultimate' kinematics. Do make sure that you examine your kinematics in terms of steering effort and feel. </div></BLOCKQUOTE>

Reassuring – and explains why the wedge of static negative camber and not move much machines seem to work well too.



Dynamics

Noted and will return to these later no doubt!



Cheers all for the input – off to revise the kinematics and look forward to another round of comments on the design of steering/uprights/wishbones! http://fsae.com/groupee_common/emoticons/icon_smile.gif

[Frank]

It's the height of the pickups that matter most (for RC's accuracy) , for-aft positional accuracy makes it easy to install wishbones.

Why sweep the fronts forward?, my preference because it gives you room for the rack to be mounted. Why 250mm?, I don't really have numbers for these arguments for you.

ASCII art required

Front

&lt;-250-&gt;
\_____|
_\____|
__\___|
___\__|
____\_|
_____\|

Rear

&lt;------350-----&gt;
\___________/
_\_________/_
__\_______/__
___\_____/___
____\___/____
_____\_/_____

Make your own rack, it's not hard, better still, have a design that you can change rate with.. (good idea if you're 230kg 47% front weight car miraculously becomes a 272kg 57% front weight car)

Need stiffness to retain toe control at rear, you can heave on the rear wheel on many FSAE cars and watch 'em flex. Compliance means you'll have wandering erratic behaviour on power-down

once you've built a car, strain gauge it, and see what sort of forces you get in your links, resolve the forces and determine loads / moments imparted at the contact patch

Need stiffness at front, because some idiot will eventually hit the brakes with the steering at full lock and skid into something, and try and invert the linkage for you http://fsae.com/groupee_common/emoticons/icon_frown.gif

Need stiffness, at front to keep that lovely Ackerman stuff, instead of compliance turning your Ackerman into parallel steering

Can you eliminate bump steer without the toe link "in-plane" with the wishbones? YES absolutely.

you can also do some pretty whack stuff with toe links and achieve "almost no bump steer"�, remember you're only going to need 25mm of droop, and 25mm of bump in your suspension...which is about typical at the front, with the rear suspension blatantly about half of this, to which everyone turns a blind eye.

Just make absolutely sure you know the tyre is a certain size, have seen many nasty surprises occur with respect to this.

[markocosic]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Why sweep the fronts forward?, my preference because it gives you room for the rack to be mounted. Why 250mm?, I don't really have numbers for these arguments for you. </div></BLOCKQUOTE>

Sure - just checking that there was nothing particularly magical about the 250mm! As-is, the rack will be sitting just above the lower chassis rail, so was planning on a "super-node" of sorts in that area and the wishbone mount/rack mount being on the same plane running vertically across width of the car.

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Make your own rack, it's not hard, better still, have a design that you can change rate with.. (good idea if you're 230kg 47% front weight car miraculously becomes a 272kg 57% front weight car)
</div></BLOCKQUOTE>

Hear you, but I really don't fancy any precision machining requiring multiple setups on the mahcine tools we've access to - it'll just eat time and end in tears...

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Need stiffness at front, because some idiot will eventually hit the brakes with the steering at full lock and skid into something, and try and invert the linkage for you Frown </div></BLOCKQUOTE>

Good call, though whether any stiffness will save you from a kerb or not is debateable.

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Can you eliminate bump steer without the toe link "in-plane" with the wishbones? YES absolutely. </div></BLOCKQUOTE>

Yes - by having one end point to the IC and putting the other at a height such that if you were to project it to the plane of the wishbones it'd be at the "right height" - but this only works perfectly for one toe setting?

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">you can also do some pretty whack stuff with toe links and achieve "almost no bump steer"�, remember you're only going to need 25mm of droop, and 25mm of bump in your suspension...which is about typical at the front, with the rear suspension blatantly about half of this, to which everyone turns a blind eye. </div></BLOCKQUOTE>

Just for the record, what is considered "almost no bump steer" (or "bump steer an FSAEer won't notice") 0.05 degrees total? 0.1 degrees total?

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Just make absolutely sure you know the tyre is a certain size, have seen many nasty surprises occur with respect to this. </div></BLOCKQUOTE>

Ok

Cheers Frank!