CanadianTercel New Reader
7/3/11 11:56 p.m.

In my fit of rage and excitement at possibly buying an FB 82 RX-7, I looked further into the fact that they have a live axle rear end (was verifying since I had told my friend this).. On wikipedia they mentioned a "watt's linkage" rear suspension, so I further looked into it. Now I'm intrigued into one day converting my tercel from a panhard rod- torsion setup into a watt's linkage setup.. Talk about cool! Does anyone have links of DIY conversions, or super informative pages book marked that they could share?

Thanks guys!

JoeyM SuperDork
7/4/11 12:12 a.m.

staniforth and puhn's books discuss it. several people here have built locosts that use it. I'm sure they'll chime in.

turboswede SuperDork
7/4/11 3:04 a.m.

There's also Mumford and DeDion for rear suspension.

The FB suspension suffers from similar binding issues that the Mustang suspension does, so it isn't much better than a standard leaf-spring solution.

11110000 Reader
7/4/11 6:30 a.m.

A Watt's linkage doesn't seem to have much real-world advantage over a well-engineered Panhard rod setup. The complexity/performance ratio seems a little high for my tastes, unless I was starting from scratch.

Curmudgeon SuperDork
7/4/11 8:02 a.m.

I've had 1st gens, there are some cheap and easy tricks to 'unbinding' the rear suspension. As far as building a Watt's link, sure it can be done but honestly the only real difference between a Watt's and a Panhard is a Panhard will move the axle slightly side to side at the ends of its travel. Come to think of it, a poorly engineered Watt's will too.

When I built the Jensenator, I built a custom 3 link using an RX3 rear axle housing, 1st gen RX7 rear disc brakes and limited slip and coilover shocks. I went with a Panhard instead of a Watt's due to space and complexity issues. The secret to a Panhard is to make it as long as possible and keep it parallel with the axle in the vertical and horizontal planes. My rear axle moves about 1/4" to the right at full bump travel.

CanadianTercel New Reader
7/4/11 8:16 a.m.

So you boys think that a torsion-panhard setup will do just as well as the watts linkage in my application then? Mostly lapping day and street driving.. With heavier spring rates and an adjustable panhard to center the axle after i've lowered it, I won't see much lateral movement.. Even now I see that the OEM panhard actually curves a bit to make up large amounts of height.

erohslc Reader
7/4/11 9:29 a.m.

Curved Panhard is very common in racing, see 'J-bar' in any roundy-round catalog.
That said, Panhard is popular with them because they don't have to turn right, and can thus fine tune for (or even take advantage of) the lateral movement.
Watt's linkage is really not complex, but as noted above, can be difficult to package.
An intereresting application worth scrutiny is the rear suspension on PT Cruiser. The link length and bellcrank radius are asymetrical left/right. The reason it works is that the ratio of link length to bellcrank length are the same on both sides (remembering high school geometry, think congruent vs similar triangles). The links are also rather steeply angled, rather than parallel to the ground, but both have the same angle. This allowed them to neatly package the Watt's llinkage within the envelope of the rear beam movement, which the body structure already had clearance for.
Another packaging trick is to make the Watt's link horizontal. The axis of the bellcrank is vertical, and the side links are ahead of and behind the axle centerline, instead of above/below. With horizontal Watt's on RWD live axle cars, a U shaped bellcrank pivoted from top or bottom of pumpkin allows the height of the effective roll center to be set wherever desired.

(Not that I've given this any thought)

iceracer SuperDork
7/4/11 9:46 a.m.

Mascar uses an adjustable panhard bar. Dirt track modifieds use a "Jacobs Ladder" lijkage tha t works like a panhard bar. Being adjustable means that you can adjust the rear roll center. So CT, design an ajustable full length panhard bar and have fun.

Mazdax605 Dork
7/4/11 11:11 a.m.

I think I remember reading that part of the issue with the Mazda Watts linkage is that the pivot point is not in the centre of the axle like it should be. Mazda has it offset to one side,and that makes for some funky geometry when it starts moving. Also the binding that is inherent with SA/FB cars doesn't help either. I know a lot of guys who race their 7's take the rear sway bar off to help with this,but I can't say if it works the way they think it does,


triumph5 Dork
7/4/11 1:08 p.m.

Not worth the time, energy and bucks just to do it for the novelty of it. I suspect the fabrication to install a Watts Linkage would be a bit involved, and simply not worth any gains.

stafford1500 New Reader
7/4/11 2:00 p.m.

The geometry of the Watt's link allows for the roll center of the rear end of the car to stay at the central pivot point. A panhard bar has the roll center located at the midpoint of the bar. In extreme travel cases the Panhard bar can get the roll center lower than a Watts link, but in most cases, the Watts link will result in a lower roll center (assuming the bellcrank arms are not too long, causing the center to be raised). The Saleen company did a Watts Link conversion for some of it's models a few years ago and the geometry was slightly assymetric, casusing almost as much lateral axle shift as the original Panhrad bar. In addition, all of the lateral load on the rear end of the car hasa to go thru the Watts link center pivot. Since this is usually a single shear attachment, the design has to be pretty beefy. Start thinking about the rear weight of the car and use a 1G acceraltion to approximate your loads. It car be done, the Mustang is roughly 3400 pounds (45% on the rear) and we did abuse testing on the Watts link assembly. Granted we had a new rear end cover designed/built with the center pivot strengthened to take the loads. I had one of the junior engineers use the opportunity to find out how much the rear end shifted laterally by build two models of the system (a CAD model and a math model) to verify the 'perfect' operation of the system.

Our Preferred Partners