Multi-link Suspension

[DannytheRadomski]

I'm trying to learn about multi-link suspension, and everything I've found has really come up short on explaining anything past "you can alter one variable at a time, making it more adjustable than double wishbone". Just any info would be nice. Pros, cons, fun facts, anything

This is purely academic, but I was also wondering about the feasibility of making an FSAE car with multi link suspension. Assuming you could do the maths, would it be difficult to fabricate or weigh too much or something? I haven't heard of any FSAE cars that use this set up and I was wondering why.

[Timmey]

Ah, the good old trusted Multi-link suspension.

We (Eindhoven University of Technology) started using the multi-link suspension in out FSAE/Formula Student car in 2008.
It was developed by a graduate student, http://www.mate.tue.nl/mate/pdfs/9288.pdf

I'm not the responsible suspension guy, so this is my own interpretation of the usefullness of this type. I was responsible for the design of the upright at the time and did the management of the team that year.

Pros:
- more adjustable - you can tweak the suspension to make more efficient use of the tyres.
- no REIB

Cons:
- instantaneous screw axis - force and moment calculations in the suspension are a b*tch
- suspension play - we noticed a strong increase in suspension play during testing due to the close proximity of the hinge points on the upright and the resulting higher forces. You need high quality rod ends to last a FSAE season.
- increase in unsprung mass - instead of 3 points (upper/lower a-arm and toe-link) to connect to, you have to design the upright to connect 6 rod-ends to the hub. You need 48 rod ends for the suspension since it moves in 3D.
- reality - The manufacturing tolerances are not nice for a theoretical finely tuned suspension.
- knowlege transfer - you need a computer and simulation to really understand what happens if you change the rod lengths. Also the graduate student responsible for developing the suspension was only a team member for one more year. After he left, we used the multi-link suspension for additional 3 years and no one really knew how to design it from scratch, so we used a slightly changed adaptation to implement chassis changes.

Fun facts:
- seeing everything move when you steer.
- seeing everything move everywhere untill you connect the last rod-end, after which you hope it starts moving the way you designed it to move.

I hope this helps, the graduate thesis is a good start in understanding the basics of designing a multi-link.

[DannytheRadomski]

If it were just back suspension, would you be more supportive of it? Then you wouldn't have to worry about it being dynamic (on the upright) for steering.

[Timmey]

Well, if it were just the front suspension, i'd be more supportive.

The rear will act like a slightly different double a-arm (if you design it that way), but the dynamic behavior of the front suspension during steering is what makes it better theoretically.
The steering axis is virtual, meaning you can place it anywhere with relation to the wheel, minimizing steering effort and maximizing the use of the tyre (camber/caster/toe) without the mechanical and shape limits of the upright.

[Z]

Danny,

IMO there are two main reasons that Multi-Link (or "5-Link") suspensions are becoming more common on production cars.

1. For the front-wheels they make it a little easier to have a "centre-plane" Steer-Axis, which is generally better than having a lot of Offset (= Scrub-Radius) and Steer-Axis-Inclination (= SAI, = KPI). Of course, Citroen had very elegant centre-plane steering in the 1950s, but I guess modern OEM engineers are a bit lazier when it comes to designing new brake/wheel-hub packages... Also centre-plane steering is quite easy in FSAE, even with 10" wheels (many teams have managed it), so no real advantage there.

2. They can make NVH (noise-vibration-harshness) control a bit easier. This is mainly to reduce the thump-thump of wide, low-profile, radial tyres as they hit lateral cracks/seams in the road. The thumping is reduced by letting the wheel move longitudinally a bit (about +/- 1" wrt chassis, sometimes called "wheel recession"). This is done with 5-Links by fitting the mostly lateral links (~x 3) with stiff, low-compliance ball-joints, for good toe and camber control, while the mostly longitudinal links (~x 2) can have much softer rubber bushes. Again, NVH is not really important in FSAE.

For FSAE the only advantage I can think of is that all the links can be very easy to make. Just a length of tube with a BJ at each end. Maybe you could do it so that only one or two different length links are required for the complete car (4 x 6 = 24 links total, including push/pullrods). This would make carrying spares easier. Of course, the cost here is that you need 8 more BJs than for 4 x Double-Wishbone suspensions, with each upright having 6 connections to BJs compared with 3 or 4 for DW.
~o0o~

To calculate the forces in the links, use the same method as in the "Analysis of Wishbones" thread (ie. easy!).

To calculate the kinematics (ie. location of Instantaneous Screw Axis), use the method given in the Suh (?) paper that is referenced somewhere in this Forum (in last year or so...). This is quite similar to the "link-forces" method above. (Thank you to the Koreans for reminding the rest of the auto-industry that we live in a 3-D world! :)) BTW, either keep the ISA far away from the wheel (at least a few metres), or if it gets close, then make sure it is close to horizontal (to minimize bump steer).
~o0o~

Bottom line is that I do NOT see any performance advantages of Multi-Links at all. They will probably be used again in FSAE by teams looking for ways to complicate the car (maybe to impress the Design Judges!), but who don't really want to win. Much simpler suspensions will get just as much performance out of the tyres, if not more performance (given that MLs can have worse compliance issues...).

Z

[dynatune]

If you want to learn all about multi-link suspension's there is only one book that you must read. This book is widely considered the almighty bible of suspension design and was written by a very famous BMW engineer by the name of MATSCHINSKY. It is called "Radfuehrungen der Strassenfahrzeuge" and as far as I know an English version is nowadays available too. The book will guide you into most of the secrets of modern multi-link suspensions. It is rather expensive and to be very honest very tough reading material until the point that you have understood the vector algebra behind all of it. A term "momentary screw-axis" as mentioned above will have a total new meaning for you, so will be suspension design after reading the book. When I read the book for the first time in 1989 it motivated me to write my own multi-link suspension kinematics program in FORTRAN and it worked... still does ....Go get it, read it and get back to me answering the my question: "How many different suspension types exist for Independent Suspensions" Name them all 3. (This was a hint ! .... and not looking for McPherson our SLA terminologies but the real ones .... those that only expert knows)

Cheers,
Dynatune, www.dynatune-xl.com