Suspension designing from scratch

[dtype38]

Goran, sorry to ask stupid questions, but what angle do you call "Sai"?

[ubrben]

Originally Posted by ss_collins

It was used on the Deep Sandersons in the 1960's and is a fascinating concept, possibly obselete - possibly not, we featured it in our May issue V15N5. It proved to be far more effective than the more conventional suspension on the Lotus 18.

http://www.lawrence-tune.co.uk/lawrence_link_1.html

"Ive never liked this business of anti roll bars lifting the inside wheel to make the outside wheels work properly. I call that building a four wheel motorbike. As you get into a corner you are only running on two wheels - Somebody once said an anti roll bar is quick way to prop up an otherwise inadequate suspension"

"a low roll centre that was completely fixed and would not migrate no matter what the suspension did"

I read that edition. Couple of points:

1 - a completely fixed KINEMATIC roll centre is possible with a conventional double wishbone suspension system so this isn't an advantage of the Lawrence link.

2 - The kinematic roll centre is not a motion centre so controlling it within fractions of mm is not necessarily worth bothering with anyway.

3 - You don't need a Lawrence link system to run no ARBs - just ask Lee Stohr who doesn't run them on his D-sports runoff winning car.

4 - Who told you it was better than a Lotus 18? Non other than the guy who invented it. Did you know gullable's been removed from the latest edition of the dictionary...

Ben

[Goran Malmberg]

Originally Posted by dtype38

Goran, sorry to ask stupid questions, but what angle do you call "Sai"?

SAI stand for Steering Axis Inclanation, the same as the older term KPI.
Goran Malmberg

[dtype38]

Originally Posted by Goran Malmberg

SAI stand for Steering Axis Inclanation, the same as the older term KPI.

Ahaaa. KPI. Errrr, ummm, not actually sure! The plane of my upper ball joint is inclined at 5 deg from my wheel axis and the plane of my lower ball joint is inclined at 2 deg from my wheel axis (in the opposite direction). That should make the kpi 7 deg, but only if the ball joints have the same centre line. If their centre lines are offset (very hard to measure) then I really don't know.

I do know that the lower ball is set further into the rim than the upper one, so I guess I must have some kpi

At this point I have to admit that there is the distinct possibility that I have no idea what you're talking about, and in fact my car goes round corners better than virtually anything I've raced against by complete fluke! This is born out by the fact that I seem to be the only car on track with no anti-roll bars. (Oh, except for my team mate who watched how my car corners and decided to throw his away too)

[Goran Malmberg]

Originally Posted by dtype38

Ahaaa. KPI. Errrr, ummm, some kpi
)

No problems, If we drw a line throught the centre of the upper balljoint and down throught the centre of even the lower balljoint, this line will have an angle to a vertical line. This is Sai, the axis around which the wheel turn when steering. So, if changing the camber, even Sai is altered. Where the projection of this line meets the ground we eill have a distance to the centre of the wheel, called scrub distance. This distance arm will creat a twisting motion att the steering wheel if front wheel is subject to diferent foces.
Goran Malmberg

[Goran Malmberg]

For those interested i put the "wheelrate" images out on my site, fare down the side. You will be amazed what it shows.
http://hem.passagen.se/hemipanter/
Goran Malmberg

[dtype38]

Yeh, ok sorry about sounding a bit dense. I have actually read about kpi and at the time I understood what it did and why. It's just that as I get older it becomes harder to retain all this sort of stuff in my head at the same time. I just start to think I have a good visualisation of what's happening, then some one reminds me about a bit I've forgotten and it all falls apart again

FWIW, this is how I went about setting up my geometry:

For no particular reason other than my own intuition, I reasoned that a good place to start to find good cornering was to keep all of the tyre tread flat on the ground. Sooo....

I plotted out my front lower wishbones, uprights, stub axles and tyres on a cad package. I set the ride height with the lower wishbone pivots horizontal then postioned both tyres upright. Ignoring the upper wishbones for the mo, I rolled the car about its (approximate) roll centre by about 3 degrees in each direction keeping the tyres upright. I then introduced the upper wishbones and plotted locus of the inner pivots around the top ball joints for all three positions to see if there was a mounting position which would satisfy the vertical tyre requirement. There wasn't such a position, so I experimented with different wishbone lengths to see if I could find one. It turned out that if I lengthened my existing wishbones by 40mm and moved the pivot down and in towards the centre line of the car, I had a system which would keep the tyres vertical regardless of car roll. Oh, and to keep everything simple I have all my wishbone inner pivots pretty much parallel to the centre line of the car (within my welding ability that is ).

From that point I just added a few degrees of caster to help the outer tyre walls during turn in. After a bit of track testing, I also added half a degree of static negative camber to even out my tyre wear.

This may all seem a bit Heath Robinson to you, but in strictly practical terms, working straight from the tyre tread to the inner wishbone pivots and simplifying everthing else as much as possible, allows me to picture what's going on without having much mathematical understanding. Guess its just as well that I'm the driver as well as the race engineer and designer. No translation problems

[Goran Malmberg]

Originally Posted by dtype38

FWIW, this is how I went about setting up my geometry:
1
For no particular reason other than my own intuition, I reasoned that a good place to start to find good cornering was to keep all of the tyre tread flat on the ground. Sooo....
2
I plotted out my front lower wishbones, uprights, stub axles and tyres on ...........
3
This may all seem a bit Heath Robinson to you........... without having much mathematical understanding.
4
Guess its just as well that I'm the driver as well as the race engineer and designer. No translation problems

1
Right, tire µ for the highest total friction is the key factor.
2
Logical way of working, great.
3
Work in a siilar way myself. Even if experimenting, I use real life models in ored to involve every factor there is. And if studying a physical model one discover things that might else not have been detected.
Math is a great thing when one know what to calculate, and to shorten work time.
4
If working in the way you suggest, you automatically involved yourself as a driver-factor in the prosess when judging the car.
Goran Malmberg

[dtype38]

Thanks for that Goran

[ss_collins]

ben the results show deep sandersons being quicker in the wet.

You have strong opinions - maybe you should write a column for us?

[ubrben]

Originally Posted by ss_collins

ben the results show deep sandersons being quicker in the wet.

You have strong opinions - maybe you should write a column for us?

Which Lotus 18s? Are you sure that the drivers were equally talented, the tyres were the same, etc. A few ancient race results does not prove the system is fundamentally better.

However my take on this is that Lawrence had an obsession with not running ARBs. What's the first thing you're told to do in the wet? Disconnect your ARBs because the roll resistance requirement reduces due to the lower lateral accelerations. All of sudden you get better (i.e. less) load variation on each tyre and therefore more grip and more performance. Why did the Lotus's not do this? I couldn't answer that.

A more plausable explanation I think you'll agree, and nothing to do with this ridiculously over-hyped Lawrence link system.

Don't know about a column though. Would that be under my assumed name as per the letters page this month :-)

Yes I guess I have some strong opinions. The thing that bugs me is the obsession with things like the Lawrence link. You need more articles about how to deal with the multiple interactions that constitute a race car rather than panaceas that don't exist. If the Lawrence link was so good why did Lawrence not design it into the Morgan Aero 8, and for that matter why was the Aero 8 so damn slow at Le Mans last year if he's any good at what he does?

Ben

[dtype38]

Originally Posted by ubrben

If the Lawrence link was so good why did Lawrence not design it into the Morgan Aero 8, and for that matter why was the Aero 8 so damn slow at Le Mans last year if he's any good at what he does?

I think that one might be just a little too far from this thread to delve into too far but on a related idea, did anyone see the new front suspension on the Westfield at Autosport a couple of years ago. I can't remember what they called it (x-link?) but they claimed it solved the anti-roll issue once and for all. Certainly got my vote for most complicated linkage system ever! Hmmm, don't see it on all new race cars so maybe it wasn't all they claimed

[Goran Malmberg]

Where is the best description of the Lawrence and the X-sytem to be found?
"solved the anti-roll issue once and for all" sounds a bit strange as a statement. Really must see what they mean by that.
Goran Malmberg

[TEAM78]

It was the DAX suspension I think your on about. it wasnt good over bumps,

[dtype38]

Sorry TEAM78, I knew it was a Caterham look-alike but was guessing which one. I remember it having two sets of diagonal links across the front wishbones, plus some complicated linkages at each end. Do they have anything about it up on the net anywhere?

[TEAM78]

Originally Posted by dtype38

Sorry TEAM78, I knew it was a Caterham look-alike but was guessing which one. I remember it having two sets of diagonal links across the front wishbones, plus some complicated linkages at each end. Do they have anything about it up on the net anywhere?

http://locost7.info/files/suspension...CamberComp.jpg
http://images.google.co.uk/imgres?im...lr%3D%26sa%3DN

I found this, it was in racecar eng ages ago. some guy was using it in the Caterham championship. There doesnt seem to be much on the net.
think how the system works in one wheel bump then you relise what its down fall is.

[dtype38]

Aha... that's the baby! But I ain't even gonna try to get my head round it. You up to that Goran? :-)

err. ok the daxcamb3.jpg makes it fairly clear, but yes, that does make one wheel bump a bit tricky.

It's late, 3/4 bottle of red wine, sorry!

[ubrben]

It's kinematically similar to a beam axle. We're back to the fact that if you want to control each wheel with a single spring/damper an individual wheel will always have 1dof. Therefore the pair of wheels together have 2dof. You can have synchronous or oppositional movement (heave or roll) one wheel bump is equal amounts of both and therefore always a compromise.

Ben

[ss_collins]

So now lets go loony again - Creuat Suspension - as used by Racing For Holland - complex but seemingly effective RfH insist it is better and the drivers prefer it for sure. Has anyone considered using it?

Ben - you clearly dislike the LL system not quite sure why so strongly - students I think need to explore what may be blind alleys to fully get around the problem of suspension design - so they can justify why the did it the way they did.

Am I right in thinking that you would like more Ortiz and McBeath and less Valkeburgh

[ubrben]

Originally Posted by ss_collins

So now lets go loony again - Creuat Suspension - as used by Racing For Holland - complex but seemingly effective RfH insist it is better and the drivers prefer it for sure. Has anyone considered using it?

Ben - you clearly dislike the LL system not quite sure why so strongly - students I think need to explore what may be blind alleys to fully get around the problem of suspension design - so they can justify why the did it the way they did.

Am I right in thinking that you would like more Ortiz and McBeath and less Valkeburgh

More of all three. Van Valkenburgh has a proven record of applying new technology and optimising race cars. That kind of blue sky thinking is good, because the guy's won stuff at the highest level and his opinion is worth listening to.

I dislike the hype things like the LL create because on a kinematic basis, anything it can do a double wishbone could get close enough that it would make no difference. It's not innovative in that sense.

We contemplated designing our Bikesports car from scratch with a mechanical version of the Crueat system. I corresponded with Josep from Crueat about four years ago and have been a big fan of the principles behind the system since then. We have decided not to do it because we've never designed a full blown circuit racer before and felt that we'd have enough on our plates as it was.

Ben

Edit: the difficulty they've had at RfH is apparently related to the relatively free warp movement allowing too much movement of the aero platform. I suspect the damping curve for the warp mode is crucial to getting it to work.

[GordonG]

While we're discussing unusual suspensions:

Problems (compromise) with double wishbones include the need for camber compensation in roll to compensate for the upper chassis mount moving outboard in roll. The more camber compensation you have for the roll the more camber changes in bump/dive/squat.

Why does nobody run a swing arm suspension with very long and low mounted swing arms? I'm thinking along the lines of swing arms mounted on the bottom of the chassis, either on the centreline or even on the opposite side of a wide chassis (so the LH wheel has a swing arm mounted just inside the RH wheel, low down.) I think this puts the roll centre fixed at the bottom of the chassis height, so less jacking than, say, a VW beetle rear swing arm. Due to the long swing arm length you would get little camber compensation, but would need none as there is little effect on camber in roll (except that the inner pivot will move up in roll, causing the outer wheel to go positive, slightly).

There;s got to be a reason why it doesnt work (other than structural and weight considerations, which can surely be overcome in teh design) otherwise people wouldnt go to the trouble of the Dax system etc.

G

[ss_collins]

RfH seem to have overcme this and are now using it in LMES races - whereas previously they only qualified with / practised with it.

[ubrben]

Originally Posted by GordonG

While we're discussing unusual suspensions:

Problems (compromise) with double wishbones include the need for camber compensation in roll to compensate for the upper chassis mount moving outboard in roll. The more camber compensation you have for the roll the more camber changes in bump/dive/squat.

Why does nobody run a swing arm suspension with very long and low mounted swing arms? I'm thinking along the lines of swing arms mounted on the bottom of the chassis, either on the centreline or even on the opposite side of a wide chassis (so the LH wheel has a swing arm mounted just inside the RH wheel, low down.) I think this puts the roll centre fixed at the bottom of the chassis height, so less jacking than, say, a VW beetle rear swing arm. Due to the long swing arm length you would get little camber compensation, but would need none as there is little effect on camber in roll (except that the inner pivot will move up in roll, causing the outer wheel to go positive, slightly).

There;s got to be a reason why it doesnt work (other than structural and weight considerations, which can surely be overcome in teh design) otherwise people wouldnt go to the trouble of the Dax system etc.

G

Camber compensation is a function of swing arm length. The effective swing arm length of a double wishbone system is kinematically identical to an equivalent single swing arm. I don't see the difference.

What you can't do with a single swing arm is change the length as a function of wheel travel.

Ben

[Goran Malmberg]

Originally Posted by dtype38

Aha... that's the baby! But I ain't even gonna try to get my head round it. You up to that Goran? :-)

Well, this system seem a bit bulky with lots of rods. I see the function but dont think there will be any general advantage. This particular installation on the picture is not very refined, typical kit car divices.

They talk about higher rollcentres, and I dont really get the point here. Forces not rolling the car wil go straight down to the contact patch.



Ben mentioned camber as a function of virtual swingarm length, or should we say the distance to instant centre.
However, the shorter and more angle the upper A-arm, the more the camber compensation. Of course it follows that the instant centre is moving around. This is not the behaviour of a swing arm.
Also, as I show at my site, these sort of cambercompensation mess upp the wheelrate.
Goran Malmberg