33 replies to this topic

[squidbreath]

I would like to put a lawrence link geometry into an old porsche 912 I am rebuilding. I read the recent Race Car Engineering article. Can anyone point me another source. I can't find anything.

[Ben]

Can I ask why you want to use that geometry?

Ben

[squidbreath]

Reasons to use lawrence link geometry:
1. It appears to keep constant camber when the chassis rolls (although not during squat).
2. It uses swing arms that pivot from the lateral axis rather than the longitudinal axis. This means that the suspension pickup points can be positioned at the front and rear firewalls.
3. As a final bizaar twist, I want to use citroen hydropneumatic springing/damping/leveling. I have the parts left over from this project There are three posts in the blog related to the project.

My vision for the car is to replace the rear engine with a subaru turbo and transmission mounted midship. The measurements work perfectly to fit it where the rear seat has been. I envision a new chassis that encompasses the area between the firewalls. If the engine hangs off the back of the rear firewall and the suspension pickups are on the firewall then it simplifies the chassis structure. I won't have to worry about engineering for twisting loads that would occur in harder to stiffen areas distal from the cockpit. Imagine an "H" where the arms and legs of the "H" are the suspension swing arms with wheels at the extreme ends of each. The crosspiece would be the chassis. Although in reality the crosspiece would be more of a box then a line.

[Greg Locock]

http://www.lawrence-...nce_link_1.html

For its day (1960) it is a reasonable design, but it has a few issues, at first sight.

Basically twin pure trailing arms, parallel and equal arms. The hinge axis is then inclined to screw around with the kinematics.

While it has some very good points packaging wise*, it is completely inappropriate for a modern production or road car, as its lateral, toe and camber stiffness will be very poor, and it basically suffers from oversteering toe compliance, at both ends of the car!

To be honest I can't see what is so great about the camber relative to the road either.

Judging by the rest of his comments and career the designer has his head screwed on, there may be subtle advantages that I've missed.

*If you combined it with a torsion bar spring and lever type shock, or ran the shocks/coil overs back to the firewall, you could potentially resolve ALL the suspension forces at the firewall. That is, none of the structure beyond the two firewalls would see ANY suspension forces (except steering). That could reduce your chassis weight/complexity enormously.

[squidbreath]

Greg:

I see your point about lateral stiffness. Toe and camber stiffness are both lateral forces and conventional 'A' arms, being laterally disposed, are naturally better at countering these forces. The article did mention that trailing arm stiffness was an issue.

Let's suspend disbelief for a second about lateral stiffness by supposing that I can come up with a magical lightweight and infinitely stiff material for the swing arms. Is there another geometry issue that I'll be giving up with this design?

[Greg Locock]

Not just your arms, your bearings at the firewall, and the stiffness of the firewall itself will contribute to the compliance.

I don't know what else it'll do, I'd have to model it, and that ain't gonna happen.

if I were building a high articulation suspension for a lightweight low speed vehicle I'd pursue this design. (ie a cross country solar car)

[squidbreath]

Thank you for the clues.

[LMP900]

Since the geometry is a conventional double wishbone set-up rotated 90 degrees about a vertical axis, all the changes to front view virtual swing arm which are apparent in a conventional DWB geometry will be seen in the side view swing arm - i.e. changes to caster (instead of camber) and wheelbase (instead of track).

It's an interesting idea, though the sacrifice in rigidity would worry me in a competition car, particularly at the front where you have to allow for steering lock. Leading arms may be better at the front (of a RWD car) because the cornering force has a rearward component.

IIRC the ROver 2000 had something like this at the front.

[Ben]

The GTM Libra kit car has it at the rear as well. Very neat solution in that context, but the last thing you'd intentionally put on a race car.

Ben

[imaginesix]

Thinking...

What about a hybrid; lower A-arm with upper Lawrence link. That would help the stiffness issue while maintaining the "total camber recovery" of this system. To further hybridise the system, the upper link could have a pivot axis that is neither lateral nor longitudinal, but somewhere in between (also to help improve stiffness).

And that's still only getting funky in 2 dimensions, I don't want to think about a pivot axis that points up/down as well!

[Greg Locock]

What do you guys mean by camber recovery - change in camber per degree of body roll?

If so, is there any reason why you can't get that with boring unequal double wishbones?

[imaginesix]

As far as I can tell, the difference between the ol' boring double A-arm suspension and this 'new' exciting one is that the rate of camber change is more linear with the Lawrence link.

As the man said "I wanted to make the inside wheel do as much work as you possibly can all the time", which I take to mean optimising camber even for a lightly loaded inside wheel, when it is in the far reaches of it's droop travel. My feeling is that regular A-arms could do that only if they were very long, or if they sacrificed tire scrub.

[Powersteer]

Just make sure you are not running bushings. I remember seeing something similar but arranged like a watts linkage. One arm opposing each other. Think it was used by the old Rover.

[Ben]

Originally posted by imaginesix
As far as I can tell, the difference between the ol' boring double A-arm suspension and this 'new' exciting one is that the rate of camber change is more linear with the Lawrence link.

As the man said "I wanted to make the inside wheel do as much work as you possibly can all the time", which I take to mean optimising camber even for a lightly loaded inside wheel, when it is in the far reaches of it's droop travel. My feeling is that regular A-arms could do that only if they were very long, or if they sacrificed tire scrub.

The guy is a good self-publicist, he even tries to claim that it had better roll centre control than an SLA arrangement. I think a marginally better camber curve (I'm not even convinced it has got one) doesn't outweight the other problems.

Is he sure the gains from a camber point of view won't be more than outweighed by the toe and camber compliance?

Ben

[imaginesix]

Originally posted by Ben
Is he sure the gains from a camber point of view won't be more than outweighed by the toe and camber compliance?
Ben

I'd think he is fully aware of the significant tradeoffs, since his declared goal for this setup is very narrowly defined (making the inside wheel work harder) relative to the variety of functions that a vehicle's suspension must be designed for. That goal handily exempts him from having to discuss anything so unimportant as toe control.

I'm fully aware that this linkage is no panacea, but it is an inspiring example of lateral thinking (and it is new to me) and I derive pleasure from examining and analysing these things to death.

In coming years, camber change may have no importance at all in suspension design if some of the interconnected suspension systems that are being explored find success, since they can be set up to eliminate body roll entirely. What form of suspension would become 'de rigeure' then (certainly not the Lawrence link), and what new challenges would present themselves?

[imaginesix]

Originally posted by imaginesix
In coming years, camber change may have no importance at all in suspension design if some of the interconnected suspension systems that are being explored find success, since they can be set up to eliminate body roll entirely. What form of suspension would become 'de rigeure' then (certainly not the Lawrence link), and what new challenges would present themselves?

Please do not reply here, I will start a new thread rather than highjack this one. I am NOT a terrorist!

[Ben]

Originally posted by imaginesix

I'd think he is fully aware of the significant tradeoffs, since his declared goal for this setup is very narrowly defined (making the inside wheel work harder) relative to the variety of functions that a vehicle's suspension must be designed for. That goal handily exempts him from having to discuss anything so unimportant as toe control.

The reason he ended up with the system was to get camber recovery using mounting points close to those found on a VW Beetle trailing arm front end because that's what his existing car was running. Nothing more, nothing less.

It offers nothing over a double wishbone system, except worse toe and camber compliance.

Ben

[imaginesix]

Originally posted by Ben
It offers nothing over a double wishbone system, except worse toe and camber compliance.

From my observation and deduction, conventional A-arms tend to increase negative camber in droop, which is contrary to the optimal camber change desired in that situation. This is because the outer pivots of the A-arms move in an arc (in front view), such that the upper pivot is drawn inboard as it moves down (in droop) causing negative camber gain. The Lawrence link causes linear movement that would promote correct camber change in droop.

So, if the only goal is to optimise camber in droop, Lawrence link is da bomb.

[McGuire]

To me this is merely something one could get away with in the days of skinny tires and limited hp. I don't see the benefit. The geometry advantages are not signficant while the unit loadings are truly scary. Might work slick for a lightweight sand buggy or ATV...leading arms front/trailing arm rear with the loads fed into a central box or cage structure, plenty of travel with minimal rates. But on a real car on pavement with modern tires and engine I suspect the chassis would handle like a box of hinges.

[Ben]

Originally posted by imaginesix



So, if the only goal is to optimise camber in droop, Lawrence link is da bomb.

Well that will never be the only goal, wasn't Lawrence's original goal, and therefore it ain't "da bomb"

Ben

[imaginesix]

Right, so how can we take the qualities of the Lawrence link (that's plural because Greg mentionned something about installation stiffness) and apply them to a sensible suspension system?

That's what prompted my first post.

[JwS]

Hey, in my book anyone that is wacky enough to do what he has to a Citroen has to have something going for him.
What the heck, build it and see what it does!
JWS

[Greg Locock]

Bear in mind that this suspension was developed back in the days when men were men and tyres were crossplies.

A radial tyre has only 25% of the lateral thrust per degree of camber that a crossply has, so in my book toe control is far more important, and camber thrust is almost down in the weeds.

Can somebody define camber recovery?

[imaginesix]

Originally posted by Greg Locock
Can somebody define camber recovery?

I have an idea, but before I choke on my foot I'd like to know why it is that camber seems not to be optimised under ANY driving situation (except FWD acceleration, maybe): So when cornering (both bump and droop), braking, accelerating (RWD) or straight-and-level driving. Why?

[Greg Locock]

I completely fail to understand the question. Camber is optimised, for many vehicles, within the many requirements that it affects. Tire wear being numero uno in most real life scenarios.

[imaginesix]

The optimal camber angle should be the same for the sake of both traction and wear. It is the angle which distributes the loads as evenly and as broadly as possible across tire footprint.

If this premise is correct, then a car with optimal camber should be able to circle the skidpad all day without seeing the outside edges of the tires wear any more than the rest of the tread. If this premise is correct then static camber should always be set at 0 degrees so that when granny drives the Porsche back and forth to bingo every Sunday she doesn't prematurely wear her tires out only on the inside edge.

However tires DO wear the outer edge more when cornering, and they DO wear the inner edge more when driving straight. Therefore if the premise is correct, then conclusion must be that tire camber angles are never optimised. Why?

[Greg Locock]

RCH

[imaginesix]

OK, the conflict between roll centre height and camber change aren't immediately obvious to me, but what is immediately obvious is that the Lawrence link suspension has a fixed RCH, so doesn't this mean that it would allow the optimised camber angles that I was just ranting about?

[imaginesix]

What a minute, it has a fixed IC, not RC.

Thinking...

...ZzZzZzZzZzZzZzZzZzZZ

[McGuire]

Originally posted by imaginesix
The optimal camber angle should be the same for the sake of both traction and wear.

It sure would be nice if they were, but they are not. More camber will increase lateral grip at the expense of increased wear. There is no getting around it.

Tires develop both camber thrust and camber torque. Just as a tire has a slip angle curve, it has a camber curve (and it will be a bit better on the negative side). Obviously the tire will wear faster at its maximum slip angle than when just rolling around free. Camber curve, same thing.

That said, obviously there is a sweet spot in grip vs. wear. It is not that hard to find, by watching the tire temps (inside, middle, outside) and tread wear patterns.

[Supercar]

Originally posted by imaginesix The optimal camber angle should be the same for the sake of both traction and wear.

This is a very theoretical statement, but I think it might be true. Theoretically speaking, if all we do all day long is drive in a circle, then perhaps yes, if we lean all the tires into the turn by a certain angle we will have the best tire wear, uniform pressures and temperatures in the contact patches, and IMHO the best traction too. BUT... if it is an oval, and we have to drive in a straight line for a little bit, then the tires will be running hotter on the inside and will wear more also on the inside. Same is valid for road courses too. This is why folks run more camber than what is good for just tire wear. Perhaps this is what McGuire is saying.

Yes, camber is always optimized for all the vehicles and all the conditions that they experience, but it is rarely optimal for any one of them at a given moment in time.

Philip

[Fat Boy]

Originally posted by McGuire


It sure would be nice if they were, but they are not.

I'm really tired of agreeing with you. Can we get back to the HP/torque thing?

[McGuire]

Originally posted by Fat Boy


I'm really tired of agreeing with you. Can we get back to the HP/torque thing?

Ya know, I don't think we really disagreed about the hp/torque thing; we just insist on different modes of explanation. But don't worry, we can still find plenty of things to argue about.

Along those lines, here is a semi-interesting semantical dispute we could crank back up...look what wandered by (long story, don't ask). This is one of the first 1997 G Force IRL chassis built. Ran only two or three races. I gave my dachshund the camera and told him to go fetch a photo of the undertray. Since the tunnels are 4" or so deep at the escape and extend forward to the engine bulkhead, and the sidepods are raised 2" above the chassis floor, can we call this a "flatbottom" car? I mean, it is sorta "flat" compared to some other cars, but it's not really "flat" in itself. All the features are there, just shallower.

[Greg Locock]

One of the things I found a long while back is that a great way to solve many suspension related puzzles is to get mud on the front of your shirt. Perhaps I should get a dachshund.