Vehicle Dynamics starting points and design process

Jan - Thanks again! You don't have to sugar coat the build quality, we know it's awful and it's one of our big things to improve, the reason it's so bad was because time simply ran out :( For the pedal box reservoirs I was thinking if a banjo was just used to turn the outlet 90 degrees then the reservoirs could be off to the side shrouded by a small plate/floor access 'hump'. I did look into vertical mounted master cylinders too but I think this method has the ability to push the bulkhead even closer to the pedal face.

z - Ahh ok I understand that now :) I did a bit of thinking on the beams (drew a few sketches, going to build a physical model later) and something that seemed to be the case was this:

Beam = best camber control roll/bump, worst in single wheel bump, most un-sprung mass

LLink = worse version of beam (poor camber control in bump) less un-sprung mass, slightly better camber control in one wheel bump (unaffected wheel stays flat)

Swing arms = similar to LLink really, higher RC if parallel to ground.

Double-wish (with average FVSAL lengths) = pretty much the 'jack of all trades master of none' largest number of hardpoints, some (possibly lots) camber deficiency in bump and roll, un-sprung similar to that of LLink but likely the best at dealing with one wheel bump situations.

I guess the interesting point is that, we run on tracks with no kerbs that are completely flat so why do we care about one wheel bump situations at all? and if they do happen they're likely to be infrequent.

So I'm actually leaning towards beam axles... but I do feel like I've been slightly pushed that way, certainly makes rear hardpoints easier.

The idea of a custom casting is something we have thought about and it's possible it will happen in the future depending on if one of our team members gets his placement at the casting company.

I get the 60/40 argument, it's just if we can actually achieve it, we're still trying to shuffle everything as far back as we can get them to go whilst being reasonable.

915mm is the leg length. See the MC's under floor doesn't make the frame higher in my mind, if the MC's are sat just above the lower edge of the frame rail then there only needs to be a slight height increase in the floor above them. If this point is under the bend in the drivers legs then why does it matter that the floor is higher there?

Fun bit?!!!!! It would be fun if I had years to think about it yeah, it's the time issue that makes it substantially less fun!

Matt - Thanks I'll have a look for that and pass it on to the driver environment guy too. Well to be honest 'big picture VD' is effectively my project title but it goes under the guise of 'suspension' here, that's partly due to a lack of understand about the project but I wouldn't like to discuss that any further for obvious reasons.

Mitchell - That makes sense now, I just couldn't see the vertical slot, if we were going down the beam axle route we'd probably have to go with two shocks since I doubt one of our small shocks could take the load on its own. I see what you're saying about the MR too, seems simple enough :) The pedal box will likely be sliding on the frame rails with a simple pin dropping through the pedal box and frame (with a split pin to stop the pin falling out), easy to adjust and simple to make. I too will be surprised but I think it is possible providing everyone really pushes it.

Jay - Interesting thoughts, I'll pass them on to the engine/dry sump guy.

MCoach - Nice idea with the mock up, we actually had a full set of scales but they've just been sent off for repair as the load cell gave up in one of them. I see the benefit with the high roll centers (more kinematic weight transfer = faster than elastic transfer) but how high is really too high? I have no experience of what I'm aiming for here, 50% elastic, 50% kinematic transfer? 90% elastic, 10% kinematic? I have very little reference points to go off, likewise with how much steering lock do I want? All I know is that last year 28 degrees at the outside wheel wasn't enough but how much does that need to increase by to be useful? or was it because of the awful ackerman that made it so clumsy around corners? All questions, all tricky to answer when starting from zero.

---------

Just a quick thought (I have a lecture in a minute) but if I was to go to a front beam axle, where would I mount the steering rack? on the beam seems easy but then it will be rotating and moving WITH the beam, wont this lead to strange steering characteristics?

Thanks all of you for your input, it's really helping us all out here! :)

There are also pull-type MC's, we use them since 2010 and work reasonably well...plus being short they are not affected by the footwell template (less than 100mm length). Regarding the steering on the beam argument, take a look at sprint cars and their way of doing things.

Christian,

I do not yet have an answer as to how high is too high, however, in layman's terms: the lower limit of 'too high' is where your car is reacting faster than you can keep up, "twitchy, or darty" as it is done with a maneuver before you have time to catch up, the upper limit is where your car rolls over. I wish I had more time to explore the concept and calculations myself, but it's crunch time for me. Adding an extra hole for the control arm to mount to can give you some roll center adjustability which may/may not add value to testing your car.

Steering, you could always use more steering lock. Drift cars have a nice set up for high lock to lock steering. If you're finding that you don't have enough steering lock, you can always add some toe out in the rear. :) ....not that that is a bad thing anyway. It'll help the car yaw more to navigate around tighter corners.

Since I'm actually doing our kinematics this year, I've been designing beyond what the car is 'designed' to roll to just to make sure that what ever compliance from some poorly designed part that pops up won't affect the system a whole lot, so thowing points in areas with low sensitivity to change. On that note, it's really something to watch a car go out on track with -4 degrees of camber....and still watch it at a positive angle in every single corner. It's baffling.

A beam is best for roll control, but that's not to say you can't get similar performance out of a double wishbone. You may end up sacrificing some bump conditions for actually getting accurate control over roll conditions, but I think it sounds worth it. Especially when the god awful camber gain in brake dive everyone complains about can be nearly mitigated through stiffer springs, a heave spring, or some other funny things.

As for your scales, find some mates who have bathroom scales. Amateur hour can still work in these conditions and sometimes a low accuracy tool is better than no tool at all.

Good luck. Hopefully nothing I've said in this thread seems like hearsay.

Getting back to the brake MC setup: The configuration you mentioned can be achieved very conveniently with very tiny master cylinders like these for example: http://www.sportsbike.com.au/images/...ake_master.jpg

For comparison: The rod ends are mm if I remember correctly...

Mech (Harry) - I'll have a look, thanks for the tip (on both counts) :)

MCoach - That makes sense to me but seems tricky to quantify without a physical car and roll center adjustment points to see the back to back effects. Steering, yep I agree, providing that your arm angles aren't limiting your steering lock! :p I understand the compliance factoring (or at least I think I do) but it's something easier quantified with physical testing to my mind.

I can't say I agree with you on running stiffer springs to counter the dive/traction effects because if I understand it correctly you're crippling your performance in traction zones to do it. Bathroom scales is an option, we did actually talk about that.

Doesn't sound like hearsay to me but then again I don't know much!

Luniz - That looks interesting, I'm assuming you were going to say 8 or 6mm? we are looking at getting new master cylinders since our current ones are not spherical mounted at their ends, what kind of prices are we talking?

Onto my next thing, I get the impression I'm going to make myself look rather stupid over the next few sentences but here goes!

(I assume all camber increases to mean negative camber gains!)

With a double-wish with FVSAL's bigger than track/2 camber compensation in roll is sub 100%, now to my mind, any steer input will impart a roll event so the camber winds off as that begins, turning the wheel (assuming some caster angle) will increase the camber again so the camber is roughly increased back to its original static setting (or more, whatever you fancy!) but the caster gives you a jacking effect which is bad as it moves the CofG higher correct?

So... with a beam axle, there is no camber change in roll so there's no need to dynamically add camber to the system so why run any caster at all? if you want some steering 'weight' (again another topic on how much is enough/too much/etc) and self centering just add some mechanical trail (with spindle offset). With no caster there will be no jacking so no ill effects... seems a bit have your cake and eat it but what do you guys think?

(Note I have assumed the tyre doesn't pull under and reduce its own effective camber angle which I understand to be something quite possible?, I have also assumed infinite stiffness uprights and beams which again are likely not true).

I'd also like to add that we have a 2D plasma cutter at our disposal so I'm going to draw up some 1/2 scale beams tonight and hopefully cut and get them welded tomorrow, I do much better with proper visual aids than CAD/theory so I think this will help me, plus it will give me an idea of how long it takes to make the beams.

For a beam I was thinking of going for a inverted Watt's linkage (center on the frame, pickups on the beam) for starters as RCVD talks about peg and slot designs being tricky so for a starter I'm going to try Watt's link. I'm still struggling on how I would package a front beam but I'm thinking 4 arms leading to the front bulkhead?! not sure how I'd constrain lateral motion though and I still don't have the steering issue figured out but I'll get there!

More sketches tomorrow! (sorry I never got round to posting a new one yesterday)

Quote:

---

Originally Posted by Jay Lawrence

On the subject of sumps: a well baffled custom wet-sump can be around 40-50mm thick

---

12.5 for my old team's last dry sump, without modifying anything of the engine crankcase ;)

Quote:

---

Originally Posted by Jay Lawrence

extra load on engine to drive pump

---

But then again, you will get some power back (crankcase vacuuming)...

Jay is of course correct with the rest of what he said, dry sump does mean extra issues, he does mention a few.
On the other hand, it will also take someone fairly dedicated to develop well enough baffles (or rotary/swinging pickups, or whatever) that actually work good enough on a shallow sump, without "overfilling" the engine with oil. I have seen at least as many teams which destroyed their engines with poorly (or overly ambitious) designed baffled sumps than with dry sumps.

Well, you have to weigh the arguments for yourself :) Both options have their pros and cons.

Jan,

I think our dry sump was a similar thickness, but at the time we were running a CBR600F4 and the shifter lever had to point downwards. We were running a single gear at the time so not having a shift lever didn't matter (otherwise the shift lever would have gone below the chassis...). Something to possibly be aware of, but I think the geometry of the RR engine negates this anyway.

And yes, don't underestimate the complexities of either system. Another anecdote: after I left the team there was a wet sump 'design' that didn't have baffles at all. The solution? "Just add more oil." I think it ended up with something like 6 litres in there to stop it surging. It ran just fine, but was certainly not desirable.

Christian,

TIMELINE - Given that you have to the end of this month (say?) before "final concept" is frozen, it is good that you are investigating all these different possibilities now. The final concept drawings are only hand sketches, or maybe even just your tapelines-on-floor, so they only have to be "frozen" on Sunday night, 30th November.

On Monday 1st December, you have a big Team meeting and say, "These are the major dimensions. Now let's get detail drawings finished before Xmas...". Or something like that... :)
~o0o~

SUMP - Preferable is lower engine CG via a shallow sump, either dry or baffled-wet.

But ABSOLUTELY ESSENTIAL is an engine that works reliably.

So maybe set your overall concept dimensions based on a lowered engine fitting in a minimum wheelbase chassis. Then, worst case, if this engine is not ready for early testing, you fit your regular engine a bit further back, which probably means stretching the wheelbase to 1.6 m (or even 1.7 m?), and fit wheelspacers to widen the rear-track a bit.

The higher CG will still allow the car to accelerate hard (ie. more rear weight transfer). The longer wheelbase and wider track will slightly slow-down lap times. But lots of driver testing will speed-up lap times much more. Then maybe in May (???) the lowered engine is ready, and you can go back to the shorter and narrower car. Wow, faster still...

Expect to take quite a few of these detours. Nothing much is linear in the real world...
~o0o~

CAMBER CONTROL - "Beam = best camber control roll/bump, worst in single wheel bump, most un-sprung mass."

NOooooo!!! This very common mistake comes from a complete lack of clear definitions, or clear statements of assumptions, that is all too typical today (ie. failed education system...).

Namely, what is a "single wheel bump"???

I covered this before, so won't go into detail again. But any passive suspension has ABSOLUTELY NO CONTROL of the "camber change" that comes from short wavelength bumps. (And a possible good definition of a "single wheel bump" is one that "has a wavelength less than the distance between nearest wheels".)

Anyway, think about the relative front-view angle between the undistorted tyre-footprint and the road-surface as a car drives over a very heavily rutted road. This angle is constantly changing from +~45 to -~45 degrees, regardless of wheel-camber wrt chassis.

For FS purposes, the road surface can be considered a straight-line passing through the two tyreprints (ie. only VERY long wavelength bumps). So, whenever one or both wheels move up or down, the best camber control comes from a beam-axle. On this flat road, there are ALWAYS situations where independent suspensions give "bad" camber.

Yes, the beam does have a bit more unsprung mass. But see the last few pages of the "Beam-Axles..." thread for Ralph's pics and links to beam-axled cars racing on what amount to ploughed paddocks. And quite fast they are...

For reference, the "best" suspension kinematics for off-road racing (IMO, and seen in practice) is NO LATERAL-SCRUB of the wheelprint (ie. so horizontal n-lines, or ground level "RCs"), and NO CAMBER-CHANGE (ie. infinite FVSALs, for minimum gyroscopic forces). These kinematics are easiest achieved with pure Leading-or-Trailing-Arms (ie. a lateral-pivot-axis swing-arm).
~o0o~

ROLL-CENTRE HEIGHTS - Warning!!! Alarm bells!!!!!

DO NOT HAVE "HIGH" RCs WITH INDEPENDENT SUSPENSION!!!!!!

Well, you can, but if you do, then you MUST take appropriate measures.

For example, fit a stiff anti-droop mechanism to the Axle-Bounce-mode (eg. use a "droop-limited third-spring"). Or just use the standard solution when people back themselves into this corner, and fit really stiff springs everywhere, so as to prevent any suspension movement at all. "Any suspension will work, if you don't let it..."

"High RC" in FS conditions (ie. track ~1.2 m) is anything above about ~0.1 m. That gives a Jacking-Force ~1/6 x the horizontal-cornering-force, which can start to give significant chassis lift. Then bad things happen... As noted above, there are ways to cope with this, but please be aware of the potential pitfalls first.

Also note that Beam-Axles DO NOT JACK. Well, not in the same way. So you can run highish RCs with beams. On rough tracks, lower RCs are usually better (see "lateral scrub", last paragraph, last section). On FS type tracks, and with beams, you could have the RC at CG height, with very soft springs all around, and no body roll through corners. Of course, then all LLTD tuning would have to come through F:R RC height. Easiest to start with RC-height at a convenient 5 - 10 cm.
~o0o~

BRAKE-PEDALS - I think there are enough "prior art" solutions out there that, at the big-picture level, you can consider the front-bulkhead to be just in front of the forward-most pedal position (= Percy's 915 mm legs).

Design guidelines for Brake-Guy:

* The brakes MUST be able to lock all four wheels at Scrutineering.

* They SHOULD have a "hard" pedal, and an easy-to-modulate feel.

* They DO NOT have to be "optimised" to the N'th degree. (An excess kilo or two will NOT make any measurable difference to lap times.)
~o0o~

I better get going on these Beam-Axle sketches. (Steering can be done via a telescopic, ball-bearing-splined shaft, to be shown on sketch...)

But meanwhile, make sure you get lots of good sleep now. Your brain does its best big-picture thinking then. Well, mine does... :)

Z

Only thing I really have to add at this point is that Z's definition of 'high' roll centers is higher than what I was considering high.
I was considering a high roll center to be anything really above about .75m so I guess Z has a higher limit on jacking forces than I was considering.

While I believe a beam axle rear is a fantastic solution for fsae I am not really sold on beam axle fronts. We completed a design with a front beam axle but scrapped it due to many issues.

The whole nose of the car needs to be higher: 20-30mm beam clearance to ground + ~60mm high beam + 30-40mm clearance to chassis for suspension travel and you are looking at the chassis floor starting 100mm off the ground. Add in template height and a top to the chassis and you can't lie the driver down anymore because they can't see over it all.

The chassis structure gets messier: to get the driver low and the nose high (for beam clearance) you have to have either a large step up in the chassis or lift the driver, can't have a flat floor (at least with a 4cyl).

The steering system is significantly more complicated and has more sources of compliance/play. With double wishbone you can easily run a straight shaft from wheel to steering box. No unis, no bevel gears, no special slip joints or bearings. It's cheap, effective, light and has little play. If you're smart about it you can even have good ergonomics. Steering lock with a beam was also more difficult to achieve with the trailing arms required for longitudinal control. Finally with steering, due to the low mounted steering box, low clearance to chassis/steering link in roll and packaging a steering pickup in 13" wheels - the angle of the steering links becomes quite steep in front view, creating bending loads in the rack. Hello compliance.

The front of the car already has a pretty great structure to bolt some wishbones to. A rear beam removes a large amount of unnecessary chassis structure, the front does not. I am sure with more effort these issues can be solved. The ECU steering solution is a great start and i wish I had thought of it, although we still probably would have ended up with a wishbone front.