Dax/Peter Walker

[markocosic]

http://v3.espacenet.com/textdoc?DB=EPODOC&IDX=GB2341362

http://v3.espacenet.com/textdoc?DB=EPODOC&IDX=US2002125674

http://v3.espacenet.com/textdoc?DB=EPODOC&IDX=US2003205880

Views?

[nathan s]

I am not sure what views you are asking about.

I do think it is odd that the patents are dated only a couple of years ago. I must have missed something in the descriptions.

[markocosic]

"Camber Correction" suspension - it gives camber change in "roll", but not in "pitch".

The camber changes as the body rolls around corners to keep the tyres perpendicular to the road. The camber does NOT change when both wheels move up/down together under braking/power.

Basiclly means you don't get shedloads of negative camber when the nose dives under braking. Dax use it on their 'top end' se7ens and Cobra replicas - wondered what folk's views on it are?

[Z]

Marko,

I only had a quick look at the sites above, but I think I know the system. Don't want to sound too cynical http://fsae.com/groupee_common/emoticons/icon_wink.gif, but the system is basically a complicated version of a beam axle! The same benefits (ie. wheels always at constant camber relative to road) can be had much more easily with a simple beam, and they both suffer the same disadvantages (one wheel bump causes both wheels to change camber relative to road).

To be fair to Dax, their system can be tuned so that its behaviour is somewhere between a beam and a double wishbone, and it might make for easier packaging if the engine is between the wheels.

Similar systems for "independent suspensions" have been popping up every ten years for at least the last fifty years. The Mercedes GP cars of the 1930's had a De Dion rear axle with an axial sleeve in the middle of the tubular beam that allowed the ends of the beam to rotate relative to each other. Each end of the beam was attached to a longitudinal arm that carried the wheel "upright" and extended forward to a ball-joint on the chassis. Lateral location of the beam was via a peg-in-slot on the back of the diff. Because the beam was behind the axle line, whenever the car body leant outwards in a corner the tops of both wheels leant in towards the corner! With body bounce and pitch the wheels stayed upright. (Think about a "torsion beam" rear axle, as on small cars, but with the beam behind the axle line.)

The Dax system is claimed to have very good performance. To me this is just further confirmation that beam axles are a good thing, especially on relatively smooth racetracks (less so off-road).

Z

[Kevin Hayward]

Does look like it would act very much like a beam axle and suffers from the same problem of the inside wheel hitting a sizeable bump.

Probably could be done lighter than a beam axle setup given that the connecting links are just compression / tension members. Seems unneccessarily complex though and still doesn't solve the whole problem.

If you want good camber under pitch and roll go beam axle as Z says. If you want good camber control under either pitch and one wheel bump, or roll and one wheel bump go double a-arm.

But how to get ideal camber control for pitch, roll and one wheel bump?

Kev

[ben]

Originally posted by Kevin Hayward:
But how to get ideal camber control for pitch, roll and one wheel bump?

Kev

Make the roll mode v-stiff and the warp mode soft, but Kevin and Z will be able to expand more on that...

Marko: As the others have pointed out the Dax system is basically a beam axle with more components.

Ben

[markocosic]

Originally posted by Z:
Don't want to sound too cynical http://fsae.com/groupee_common/emoticons/icon_wink.gif

I wouldn't assume /everybody/ disagrees with /everything/ you have to say; but then again my mates are always trying to nominate me for that BBC "Grumpy Old Men" programme. http://fsae.com/groupee_common/emoticons/icon_wink.gif


but the system is basically a complicated version of a beam axle!
<snip>
To me this is just further confirmation that beam axles are a good thing, especially on relatively smooth racetracks


If you want good camber under pitch and roll go beam axle as Z says. If you want good camber control under either pitch and one wheel bump, or roll and one wheel bump go double a-arm.

But how to get ideal camber control for pitch, roll and one wheel bump?


If I fill in some background then:

I saw this in one of the kitcar mags and the first thought was "that's a bloody complicated way of building a beam axle". Thoughts then switched to "most FSAE cars seem a bloody complicated way of solving an apparently simple problem" and "remind me what's wrong with beam axles front and rear again"...

Kev, "ideal" could be:

Double a-arm with good camber control under pitch and one wheel bump - and no body roll. (what ben says/what UWA were aiming for?)

or: Double a-arm with good camber control under roll and one wheel bump - and no body pitch. (what most seem to be going for?)

or: Beam axle with good camber control under roll and pitch - and no one wheel bump. (what am I missing here?)

From walking it and watching on the UK track the only bumps were the transition from tarmac to concrete and the joining strips between the concrete slabs. These aren't /one/ wheel bumps...

What's the US track like - how significant would one wheel bump be there?

Are parallel rear ends always as tail-happy as the Land-Rover, or is that more a function of the one-wheel bump and generally ropey axle location?

[Kevin Hayward]

Ben & Marko,

I was not implying that the UWA system was a way around a fundamental geometrical problem. There is no real way to get rid of one of either pitch, roll, or one wheel bump.

Even in the decoupling systems the geometric problem remains even if its consequences are reduced. For example the Kinetics style system allows you to increase roll stiffness while maintaining pitch, and reducing warp stiffness. With a double a-arm system that means you can get away with less camber gain as the roll angles aren't as high.

Practically you cannot eliminate roll using the system, nor would you want to.

The same goes with the other modes. You would not want to eliminate pitch completely and eliminating one wheel bump is in the hands of the track makers.

I would say that the one wheel bump problem is significant in the US Detroit carpark. We left our linear pots on the car in Detroit to see how bad it was. It was definitely more of an issue than the track in Oz last year. And significantly more of than a problem than the billiard table smooth surface we test on.

So each of the "ideals' you mentioned Marko is a compromise. As of yet I have not seen a passive mechanism that solves the problem without just being a different compromise in itself. There was a Mercedes prototype a few years ago that had actively controlled geometry that looks pretty exciting ... but that is a whole other level of complexity.

I would also like to note that even on our super smooth test track (got to be seen to be believed) the warp mode still has significant effects on performance. On the Detroit surface you would not find me putting a beam axle anywhere near one of these cars. Australia ... hmmm ... wish we had the linear pots attached to the car during that comp.

Kev

p.s. I leave my question in the previous post still open ... cause I sure don't have an answer for it.

[ben]

Kev: I wasn't implying the system could solve any geometric problems either, I was too lazy to write fully but what I intended was to say that if you can run a very high roll stiffness in conjunction with a soft or zero warp stiffness you can focus the suspension design much more on camber compensation for bump rather than roll.

Zero displacement/infinite stiffness in any of the modes is not a good idea as Kevin points out, predominantly because you then have a mode's stiffness and damping dictated largely by the tyres, not good from a tyre loading standpoint. Or were you thinking more from a human factors point of view Kevin? I.e. the drivers perception of what the car is doing and roll being an important part of this?

Ben

[Kevin Hayward]

Ben,

I think yes on both the tyre loading issue and the driver perception issue. Both are important and both influence performance.

Kev

[nathan s]

Ok, I think I see it now. Suspension is something I have been trying to learn. What do you mean by one wheel bump? I am guessing it is where one wheel moves (like hitting a bump or hole) independent of the other wheel. Also, I gather that these setups, like the beam axle setups, would cause the other wheel to move, and compounding the problem.

Does this sound right?

[TG]

Originally posted by nathan s:
Ok, I think I see it now. Suspension is something I have been trying to learn. What do you mean by one wheel bump? I am guessing it is where one wheel moves (like hitting a bump or hole) independent of the other wheel. Also, I gather that these setups, like the beam axle setups, would cause the other wheel to move, and compounding the problem.

Does this sound right?

Yes you are correct on both accounts.

Kevin, the Mercedes system you were talking about was the Mercedes F400 concept with active camber control. I was sitting next Csaba Csere, editor-in-chief of Car and Driver, and we were discussing this concept. He was wondering how this improved grip (Mercedes claims a 20% or so gain in lateral grip) due to the increased outside wheel camber (0 - 20 degrees). (Mercedes also has the active body control system on the concept so there isn't much roll). The best thing I could come up with was that the tire would deform like the rubber disc in between two metal plate model described in Racecar Vehicle Dynamics, which he was familiar with. I was wondering if anybody had a better perspective on the subject and could elaborate on it?

[Kevin Hayward]

Nathan,

You are right in saying what affects one wheel bump. However you still get the condition with undulations in the road etc. It does not have to be as extreme as a pothole.

Individual wheel movements are not in themselves bad, it is the act of the wheels keeping in contact with the ground.

However when you have any ineterconnection the disturbance from one wheel will have affects on the other wheels. This is easy to understand using simple examples. For example if a car has a rollbar and while rolling the inside wheel travels over a bump it will transfer some of that effect to the outside wheel and vice versa.

However in reality the disturbance from one wheel in this case affects all four of the wheels. The bump effectively lifts one corner of the car slightly. This means that the diagonal to that corner will be compressed and the other diagonal (comprising the other two corners) will extend slightly.

This is why you can see very stiffly sprung cars effectively bounce over road disturbances. As far as I know there are no systems that can completely decouple what is applied to the vehicle. There are systems that decouple the major motions:

Heave
Pitch
Roll
Warp

But even examples such as a one wheel disturbance end up being a combination of those effects. It can all get pretty confusing and any way you frame the problem you will end up with some compromise.

Zapletal (Hope the spelling is right), who posts on these forums, wrote a really easy to read paper on decoupled suspension systems. (Balanced Suspension - SAE No. 2000-01-3572) It is really worth the read as most people tend to think in the four modes above rather than individual wheel movements.

Studying individual wheel movements is also worth doing. Programming a simple quarter car model as a 2-DOF is pretty straight forward. Plenty of papers can show you how. The info found looking at this is pretty useful in understanding what happens to the tyre with different spring and damper settings. This sort of stuff can be extended but in terms of best return for your time the simple examples get you a long way in understanding some of the issues involved with the suspension systems.

Cheers,

Kev

[markocosic]

Originally posted by Kevin Hayward:
I was not implying that the UWA system was a way around a fundamental geometrical problem. There is no real way to get rid of one of either pitch, roll, or one wheel bump.

What you and Ben say - its not a solution to the geometrym, but does help negate the effects.


We left our linear pots on the car in Detroit to see how bad it was.

Don't suppose we can have a rough description of that data can we? A 'how bad was it compared to xxx' type answer?


There was a Mercedes prototype a few years ago that had actively controlled geometry that looks pretty exciting ... but that is a whole other level of complexity.

"Parallel" double a-arm (good one wheel bump and pitch) with the upper chassis mounts on a rack for when the chassis rolls?


On the Detroit surface you not find me putting a beam axle anywhere near one of these cars.

Ok. And its "not beam" (or "not a-arm beam") because of the one wheel bump issue; not unprung weight, lack of camber change in pitch/roll or how practical achieveing low warp stiffnes is?


p.s. I leave my question in the previous post still open ... cause I sure don't have an answer for it.

Me neither. http://fsae.com/groupee_common/emoticons/icon_smile.gif

I guess whether you're running relatively tall/skinny/high profile 13s or wide/low profile 10s also factors into how significant the camber change at the 'good' wheel when the other undergoes 'one wheel bump' also factors into this a lot?

[Kevin Hayward]

Tim,

Unless I am mistaken the tyres on that concept were much more like a motorcycle tyre than a car tyre in the way they were constructed.

I wonder how they make the claim of 20% improvement in grip. If the tyre is specially made for the car (as I am pretty sure it was) how do they do an accurate benchmark.

Would they use the same tyre without the camber system? Surely that is a pretty crappy test if the tyre is specially designed for the camber conditions of the car.

Do they compare to a normal car of simialr size with the same tyre?

Different car with different tyre?

I am struggling to see how they can make the 20% claim accurately given the amounts of changes made.

Cheers,

Kev

p.s. I'm not saying the system isn't better as I can believe that it would be. I'm just curious as to how they can come with a number like that.

[markocosic]

(Balanced Suspension - SAE No. 2000-01-3572) It is really worth the read as most people tend to think in the four modes above rather than individual wheel movements.


That SAE bunch know how to charge don't they! http://fsae.com/groupee_common/emoticons/icon_frown.gif

SID was all decoupled, but probably "cheating", heh!

http://www.lotusespritworld.co.uk/EOtherstuff/sid2.html
http://www.lotusespritworld.co.uk/EOtherstuff/sid.html

[TG]

I might have worded the 20% improvement wrong. They got 1.2 g's so they claimed it was a 20% improvement over current high performance road cars.

With regards to how the system works: Camber is controlled at the upright. It's basically an upright within an upright with an attach hydraulic actuator that adjust the angle of the uprights with respect to each other.

The tire is a whole different story. It's 19" diameter on on outside and 17" diameter on the inside. The difference in the diameter takes into account the rounded inside corner of the tire, which is sort of like a motorcycle tire. But that's only a little more than an inch in radius. I don't know how that little bit can make such an improvement, if that is the reason for the improved grip levels.

[Kevin Hayward]

Marko,

I have none of the US data at home (where I am working at the moment trying to write a thesis). If I get a hold of it and remember I'll try and post some sort of comparison. By memory the increase in peak damper velocities before changing any damper settings from home was quite significant. By memory the peaks were up to 75% higher and the histograms were much fatter. The time around low velocities was decreased by about a third. I am talking about when we first arrived in Detroit testing on a surface very similar in roughness to the actual event carpark. By the time we reached the carpark we were running different tyres, different springs, different damper settings.

That Lotus is cool. It would be great to get a hold of a testbed car like that for a day or two. I would imagine the things you could learn would be amazing when you can change characteristics so quickly.

Tim,

I would say the tyre would be the first place I would look for that magnitude of improvement. A motorcycle tyre is a wonderous thing. Most on road driving for a car is done on the same rubber that does the cornering. However a motorcycle style tyre essentially corners on different rubber than it drives along at. This gives you the ability to alter the construction of your tyre to allow for areas of different grip and wear characterisitics. Also for a prototype car designed to show as much grip as possible it may be in the interests of the company to use a higher grip / lower durability tyre to show off.

Still an amazing car.

Kev

[TG]

Kevin,

thanks for the response. I figured so much as to the tire being extra soft but I wasn't sure if there was anything more to it than that at all. Here's a couple of pics that show what we are talking about:

Here's the high camber angle when turning (http://www.seriouswheels.com/2000-2003/2002-Mercedes-Benz-F400-Carving-Concept-Front-1600x1200.htm)
The upright (http://www.seriouswheels.com/2000-2003/2002-Mercedes-Benz-F400-Carving-Concept-Construction-1600x1200.htm)
You can see the rounded inside shoulder of the tire in this one (http://www.seriouswheels.com/2000-2003/2002-Mercedes-Benz-F400-Carving-Concept-Rear-1280x960.htm)

[markocosic]

Originally posted by Kevin Hayward:
(where I am working at the moment trying to write a thesis)

Get back to work then! http://fsae.com/groupee_common/emoticons/icon_razz.gif http://fsae.com/groupee_common/emoticons/icon_wink.gif


By memory the peaks were up to 75% higher and the histograms were much fatter. The time around low velocities was decreased by about a third.

http://fsae.com/groupee_common/emoticons/icon_redface.gif


By the time we reached the carpark we were running different tyres, different springs, different damper settings.

Wow - that'll be a team that take setup SERIOUSLY then!


That Lotus is cool. It would be great to get a hold of a testbed car like that for a day or two. I would imagine the things you could learn would be amazing when you can change characteristics so quickly.

I'm wondering if it still exists and whether Lotus would even consider entertaining such a proposition... http://fsae.com/groupee_common/emoticons/icon_wink.gif "Please sir, may we use your one-of-a-kind test car to learn about suspension..." http://fsae.com/groupee_common/emoticons/icon_biggrin.gif

This one is another that might be about to go commercial?
http://www.google.com/search?sourceid=navclient&ie=UTF-...en&q=bose+suspension (http://www.google.com/search?sourceid=navclient&ie=UTF-8&rls=GGLG,GGLG:2005-19,GGLG:en&q=bose+suspension)

[Z]

First of all, you guys should get back to some real work! 15+ posts a day on one thread is too much! http://fsae.com/groupee_common/emoticons/icon_smile.gif


Originally posted by Kevin Hayward:
But how to get ideal camber control for pitch, roll and one wheel bump?

Well, the obvious way is with some active controls, ie. electronics + hydraulics + etc., as on above posts. But I reckon it can be done passively (ie. all mechanical, with no energy input), but I'll have to post on that later... (other things to do now...).

As far as practical simple suspensions are concerned, I see it this way.

If you have a really bumpy track - ie. rough off-road, not that millpond US FSAE track http://fsae.com/groupee_common/emoticons/icon_smile.gif - then you want no camber or track change of the wheels, relative to the car body, as they move in bump (so no gyro or scrub forces). So VW double trailing links at front and trailing arms (pure, not semi-) at rear. Not surprisingly this is still very popular on off-roaders. Or leading and trailing arms as per the Citroen 2CV. The excessive body roll during hard cornering that comes from this system can be either ignored (off-road), or countered with appropriate springing (stiff roll mode with soft twist (warp) mode). Using an active roll mode spring gives pretty much a complete solution for all conditions (albeit with the cost of the "active", but this can be done "passively"...).

If the track is not so bumpy, say "undulating", then beam axles make a very simple and effective solution. Other than having the consistent camber control relative to the road mentioned above (regardless of body bounce, pitch, or roll), they can also be used with very soft springs. This means that there are lower twist mode induced changes in wheelprint loads. It also means that soft dampers can be used, hence again less change in wheelprint loads. The main disadvantage is higher inertial forces over bumps because of larger "unsprung" mass (hate that term, because it is sprung!). This used to be a problem with "olden day cars" with large diameter, heavy wheels, on a narrow track axle, travelling on very rough roads - hence the bad image. Not so much of a problem with small light wheels on a wider axle and smoother track. And even in rough off-road racing, beam axled "truggies" (truck/buggy) do win races.

This is a big subject, but gotta get other stuff done!...

Z

PS. Maybe someone can find a reference to the Renault/Michelin (??) "passive camber control at the upright via a below ground virtual pivot"???, tested only a few years ago.

[ben]

I have copy of that patent at home. I'll dig out the number at some point this week.

Ben

[clausen]

We were looking at systems like this to incorporate into our 2004 car, in a desperate attempt to try and remove at least some of the many compromises we have to deal with. We also came to the conclusion that such a system acts just like a beam axle, and that that's a good thing. So that's we why ran a De Dion.

Did you see our car Z? (Uni of Adelaide in the australian comp)

The Dax system probably gives you the ability to have zero camber in two wheel bump, but pull a little bit of negative in in roll.

Our car ended up having perfect tyre temp distribution on the skid pan with about half a degree of negative static camber.

We didnt do any serious calcs or simulations to see just how big the compromises were, which we should really have done.

Bob from Monash (another beam axle fan) and I were talking about how we'd quantify the effect of the lack of independance on single wheel bump, but we couldn't really come up with much.

Kevin - are you still at Uni? If not where are you working?

[markocosic]

What mass were you looking at for the 'unsprung' bit of the Di-Dion setup?

[Rickard]

The Dax system probably gives you the ability to have zero camber in two wheel bump, but pull a little bit of negative in in roll.


The DAX system has no camber change in roll. atleast not if you trust this picture:
http://locost7.info/files/suspension/dax/daxcamb3.jpg

The downsides of system that I heard of is track change (good in fsae to warm tyres?) and alot of jacking.

[Z]

Originally posted by clausen:
Did you see our car Z? (Uni of Adelaide in the australian comp)


Paul,

Eventually found a small picture of your De Dion in RaceTech magazine (Feb/Mar 2005). Some questions, if I may;

What weight for the beam's main tube? Is it aluminium?

Did you use the flanges between the beam and the "uprights" for camber and toe adjustment? If so, did you make many adjustments there?

Looks like a four-link longitudinal location? Can't quite make out the lateral location, but it looks like a Mumford link? Any reasons for using these?

What did the Design judges say? http://fsae.com/groupee_common/emoticons/icon_eek.gif

Most importantly, what did you think of its performance? http://fsae.com/groupee_common/emoticons/icon_smile.gif

Z

[markocosic]

Z:

http://fsae.com/eve/ubb.x/a/tpc/f/8356059423/m/30810020911

http://fsae.com/eve/ubb.x/a/tpc/f/8356059423/m/54410868711

Welds and flange sizes would suggest the beam is Al and its 4 links for longitudianl location, but I'm not spectacularly clear on the lateral location either! http://fsae.com/groupee_common/emoticons/icon_confused.gif

Very neat packaging though

[Denny Trimble]

Yes, they used a mumford link. And the design judges said "You Win!".

[Greg H]

They've got a clear model of the whole setup here (http://www.uarc.sethdesigns.com/).

[clausen]

Hi,



The rear unsprung weights were around 14kg per corner if I remember correctly. They were about 1kg more than the fronts.



The "uprights" were 0.6mm 4130, and the main tube was Aluminium. I think it was 2kg. I was hoping for 1kg but ended up not easily being able to get tube thin enough. Its 3 inch dia by 4.6mm wall. I was hoping to use 3 inch steel exhaust tube of 0.8 or 1.2mm wall, but they I dont think they could bend it that thin. Part of the reason for a seperate tube part was to replace it with a carbon one, but we obviously never got to that bit http://fsae.com/groupee_common/emoticons/icon_smile.gif



We used 4 trailing links, which was a bit crap. We got binding in roll - I did have plans to run the 4 link plus a rotational sleeve in the middle (was excited to hear that someone else did that with a dedion years ago elsewhere i nthis thread). I didnt want an asymetrical 3 link because i wanted the main tube to have to deal with as little load as possible.

We used the mumford for the same reason. Both the mumford and the watts link are the only commonly used ones where the load paths are directly from the chassis to the hub area on both sides of the car. A panhard rod is nicer because its simpler, but depending on which way its mounted, there will always be a wheel whose loads are fed into the chassis by first going through the main tube. And we needed a pretty low roll centre, so a watts link wasnt going to work.

Z, we thought abuot putting wheel aligntment adjustment down there but it would involve a lot of driveshaft plunge when changing camber. On the outboard ends of the sheet steel uprights are aluminium hub carrier's (arranged like UWA's hubrights (and all of the people that used that arrangement before them but didnt think to give them a catchy name)) which had a really crap shim adjustment for camber and toe.

I dont recall the judges saying a real lot about the rear end, but we did win. I think Ron Tauranac liked it. I thought we'd proabably do ok when in an informal conversation before hand he started a sentence with "when I did my de dion..."

Claude Rouelle thought it was a crap idea when i first showed him the concept models I had, and after seeing the car he still doesnt like it. He thinks the roll centre is under ground, which is wrong. he also thinks it would have different spring rates in ride and roll, which it doesnt. A lot of effort went into the positioing of the dampers and pushrods so that it does have the same wheel rate in single or two wheel bump.

[clausen]

If I was the design judges I would want to see some type of evidence of professional engineering decision making that would have to involve at least an attempt at quantifying the effect of the lack of independance on a single wheel bump input, and also (much easier) some curves showing the effect of the different cambers in the different scenarios from a beam and independant suspension. Neither of which we did.

My thinking was that we could have a relatively soft, compliant grippy suspension, that might roll a fair bit, but roll didnt hurt us becuase we had no camber compromise. Problem is that if you're designing for nice bump compliance then ruining the car's suspension independance is a step in the wrong direction.

the big question is - is this backwards step a bigger one than the forward step of nice camber control?

You can say that a beam is a nice idea on a smooth track, but if the track is smooth, then why not just stiffen up the roll so there's no camber loss or gain to worry about?

Given all this, when i get around to designing a formula ford, im still going to make sure that i can bolt on bolt off a dedion to try it. (ala 70's ferrari F1 dedion experiement)

[Z]

Well, here I go again - a critique of Adelaide's De Dion rear suspension! http://fsae.com/groupee_common/emoticons/icon_eek.gif Seriously though, I hope anyone reading this takes it as constructive criticism, rather than merciless student bashing. http://fsae.com/groupee_common/emoticons/icon_smile.gif

One of the strengths of beam-axles is their relative simplicity. It would be good to capitalize on this advantage by keeping it simple. I think Adelaide's location of the beam, and their pushrod and rocker springing, is unnecessarily complicated. Otherwise I commend Adelaide for being brave enough to take this rarely trodden path, and doing a neat job of it.

RE: LOCATION OF THE BEAM
========================
Using four longitudinal links and the (complicated) Mumford lateral location has overconstrained the beam, hence the "binding" Paul mentioned. A beam axle needs 2 degrees of freedom relative the body, ie. one for each wheel, or one for axle-bounce and another for axle-roll, depending on how you want to think about it. This only requires 4 ball-ended links to constrain 4 out of the 6 available degrees of freedom. There are many neat and structurally efficient ways of doing this. If you use the even simpler Ford Model T style of beam then there is only 1 ball-joint required, together with lateral location, which can be, for example, a simple "peg-in-slot" (for the short travel required in FSAE). The 1930's Mercedes GP system is also very simple. (If someone tells me how to post pics on these threads I can show it.)

RE: ROLL-CENTRE HEIGHT
======================
If the beam's RC is a large distance from the ground (above or below), then any roll motion of the axle, as from a single wheel bump, causes a sideways movement of either the body or the wheelprints, or both in opposite directions. This is bad for both comfort and grip. It is easily felt on 4WDs that have Panhard bars, and hence RCs, about 0.5m above ground. But a RC about 0.1m away from the ground is less of a problem.

On a beam axle it is possible to use adjustable RC height (say 0-15cm) as a powerful tuning aid for handling balance (via "kinematic load transfer"). This actually works better than tuning via springs/ARBs/dampers because it acts faster. It is more difficult to do this on independently suspended cars because of "jacking" (fewer jacking problems on beam axles).

RE: "INDEPENDENCE" OF SUSPENSION
================================
It is usually taken for granted that "independence" is a good thing. Why? Would a football team with players that each acted independently, with no communication or cooperation, be a better team? It doesn't seem to work with the "fully independent" players in Under 6 year old teams! http://fsae.com/groupee_common/emoticons/icon_smile.gif It is generally accepted that a fully active-everything suspension would work the best, if not for the extra complication/cost/unreliability. This is because the central controller - the "brains" - would know everything that is happening with the body and each wheel, and could thus coordinate all four wheels' actions accordingly, ie. "fully dependent suspension".

A beam axle, and the DAX system, does this coordination to a small extent by keeping both wheels perpendicular to a line through their wheelprints, and hence perpendicular to the "average" road surface between the wheels. Interconnected springing is also advantageous for similar reasons. There are too many of these reasons to cover fully here, but briefly... It is easy on a beam-axle to interconnect the springing of the two wheels. For example, rather than an "independent" spring acting on each wheelprint, you can have an "axle-bounce-mode" spring (eg. a vertical coil at the beam centre), and an "axle-roll-mode" spring (eg. an ARB). Each of these "axle-(or two-wheel)-mode" springs gives a force at each wheelprint that depends on the positions of BOTH wheelprints, ie. the force on one wheelprint is "dependent" on the position of the other.

Paul says that they tried hard to make the spring rates equal in axle-bounce and axle-roll. When you consider the behaviour of all four wheels, then this system (with a similar one at the front) means that the four wheels have equal stiffnesses in four-wheel-bounce/pitch/roll/twist-modes. This is not good because the twist-mode is (usually) best left very soft. Mounting approximately vertical coils on a beam inboard of the wheels (as is normally done) makes the axle-roll-mode softer than the axle-bounce-mode. From a four wheel perspective this gives a soft four-wheel-roll-mode, so the body will roll more during cornering, though this doesn't effect camber angles as noted before. But the four-wheel-twist-mode is also soft, which means less adverse wheelprint load changes on undulating ground. This is a big advantage. Wheels that cooperate and work together can beat a motley rabble of "independent" wheels - just watch those Under 6 teams! http://fsae.com/groupee_common/emoticons/icon_smile.gif

Again, congratulations to Adelaide for some original thinking. (And BTW, I like the 0.6mm 4130 uprights, rather than CNC'd aluminium.)

Z

[Psychosis]

I don't know a great deal about suspension, but a couple of thoughts. could you not replace the peter walker cross links with some form of shock/hydraulic ram that responds to the high velocity impact of a bump, but not the low velocity of vehicle roll?

shoot me down if im wrong!

[Z]

Originally posted by Psychosis:
... some form of shock/hydraulic ram that responds to the high velocity impact of a bump, but not the low velocity of vehicle roll?

I think it would have to be the other way around. The "shock/hydraulic ram" should NOT pass the camber change information from one wheel to the other during high velocity bumps, but it should do the camber changes during slower body roll situations. This would mean that rather than a "hydraulic damper" type linkage, a spring type linkage would be better. But that means that wheel camber angles are controlled by a spring and hence somewhat floppy.

I think (memory???) Wolfgang Weiss presented an SAE paper, sometime in the late 1990's, explaining his hydraulically interconnected Dax-type system. The hydraulic approach would make for easier packaging, eg. if there was an engine between the wheels, and would also make for easier implementation of active controls. It would also, like all hydraulic systems, eventually leak and make a mess.http://fsae.com/groupee_common/emoticons/icon_wink.gif

Z