Rear Suspension

Welded, So It Must Be Right

Fitting a custom-built rear suspension on my Porsche replica has caused a few anxious moments in design, construction and geometry. Any one of these could be off. Well it is now all welded in place so it must be right.

Adapt or perish

The ‘tricky’ aspects come about because for reasons of mongrel engineering I have a Mercedes 190 multi-link upright - a deep sea creature cast in steel. So first, it needed an adapter plate to carry the coil over. It’s made from 4mm steel and use of precision laser cutting (by Orrsum) means that it interlocks like Lego.

Mercedes Rear Hub Adapter

Despite hours of mental torture I could only figure out how to have two main link points at the Mercedes end of the double wishbones. The lower link is in single shear and picks up on the Mercedes casting while the upper link is in double shear and is fabricated into the adapter. The overall effect is triangular and requires a separate toe arm that fixes to a tie rod ball joint in the Mercedes casting.

The wishbones are ‘by the NZ LVVTA Constructors Car Manual’. The only trouble is they were designed and built before the CCC really got into the Manual, so the plans haven't been sanctioned by LVVTA. Still they are wide spaced, built of 26 diameter 2.5mm mild steel and do not carry spring loads.

So they should be ok???

The inner bushes are 40mm polyurethane front control arm bushes from a 1970s VW Beetle. The outers are 5/8” shank ½” eye Teflon coated rod ends (Speedway Motors #91002967).

A local race car fabricator, John Mines, tigged them together, again as required by the Manual. They sure look pretty. John required a jig - wood was ok. I used a full 327mm length of tube to form the inner eyes and cut out the unwanted section once welded. That way the inner eyes are straight.

Following good advice at club night, I approached a local engineer to do NDT testing. He couldn’t have been friendlier and likes to help hobbyists. He did the job ‘gratis’. I think he find these little jobs a fun diversion from his more complex job of testing high pressure oil and gas pipelines. And of course there were no issues with John’s welding.

Dimensions

Strings n Things

"Just where should the suspension points go?" I hear you ask. Rush to your library and take out Alan Staniforth's, Competition Car Suspension: A Practical Handbook. I also used Jeffrey Daniels’, “Car Suspension At Work”, and many many web articles and other sources.

I scratched together a spreadsheet version of the Staniforth string computer. It's called BonesApart, which is a pun for both sides of the Atlantic and should even make the French happy. It has proved to be very useful for estimating roll centres, camber change, track change etc. The roll centre moves with suspension movement. It can't do roll axis, toe change or bump steer. The 3D maths of this is too hard and makes my head hurt.

The file needs a reasonably recent version of Excel to run. Please enable macros when asked by Excel. Feel free to distribute. BonesApart Download. The first tab contains a list of instructions. The second tab is the main work area. The third tab shows the instant centres and virtual swing arms.

BonesApart Screen Grab

There are useful free tools on the internet such as http://www.vsusp.com/

VSusp Screen Grabs

Static

Roll 2 degrees

Roll Centre Migration Camber Gain

Geometry

All this plotting and planning suggests I’m pretty confident about the final suspension geometry but really this is just informed guesswork.

The car has a static rear roll centre of 50 mm. It tilts down to ground level at the front something intrinsic to the car's front suspension’s trailing arms which have a roll centre at ground level. Daniels says this means the front has a bigger lever arm in roll and will react more weight to the rear and, thus, encourage over-steer.

The chart above shows about one degree of camber change at two degrees of body roll. Thus the wheel cambers at about half the rate of body roll and thus stays relatively vertical which discourages over-steer.

Staniforth says this is a medium-arm design. A long-arm design has poor control of roll but good track and roll centre control; and short-arm has good roll control but extreme camber change in droop and squat.

In theeooorrryy this will - - nah it’s all horse pucky. There are many many other effects: weight; centre of gravity height at both ends; tyre size (F=195, R=215); front suspension geometry - which includes useful camber gain when the steering wheel is turned; springs and dampers; anti-roll bars and driver stupidity (large amounts of that in my car). So, ‘pop’, theory goes out the window.

All I can say is that it went around Manfeild pretty well in wet and dry. It steered from the front and followed from the back! I could feel the presence of the mid engine but it remained tidy.

Laser show

At the risk of boring you with diagrams... the other geometry concern was rear bump steer thanks to the three point - toe arm rear suspension set up.

I cobbled together a laser pointer and mirror bump steer tool . This is easy. The laser is fixed to the wheel, the beam travels to a mirror 5.0m away and back to display as a dot on sheet of paper on the wheel. As the wheel is moved up and down any toe movement is exaggerated 10X at the laser dot. Possible fixes were helped by a handy little chart found on the WonderWebWorld. The laser will also show up movement and compression in bushes and mounts.

I clamped the toe arm to the chassis and moved it up and down, in and out, repeat, measure, repeat… The toe arm inner mount was 10mm out in one plane and I had to pack it by 4..5 mm in the other. 1.0mm made a lot of difference. Around Manfeild with everything clamp-bolted tight it seemmmmeed to work.

The chart below shows how the laser dot moves. It comes from a very useful LVVTA guide to bump steer, found here

Toe Jam

Ok, so last move - weld in the clamped pieces. One of my club mates bravely volunteered to load his MIG into his van and weld the toe mounts in place on site. I plied him with endless cups of Earl Grey while he squeezed upside down to weld up-hand, back-hand, cack-hand. Sod’s law prevailed. His welder worked perfectly down-hand on scrap steel, but soon as he squeezed into place it spluttered and fouled. He is normally phlegmatic but was reduced to muttering about dodgy overpriced wire. More tea ensued.

He also welded in plates to box the lower suspension mounts. I had previously gone mad with the angle grinder to ‘make clearance’.

It is all welded in, so it must be right.

Next Intercooler