LocostUSA.com

The traditional Locost style 4-link-and-panhard binds a bit in roll, essentially turning the rear axle (or more likely, chassis flex) into a rear anti-sway bar. One way to reduce that is to put a flexible (rubber, urethane) bushing end in place of one of the eight rod ends in the four parallel tubes that make a 4-link system. I did that and it worked fine, but the more I thought about it, the more I realized that left the other three tubes doing all the work, so I tried simply removing the fourth tube, which made the rear suspension a 3-link-and-panhard with a waaay offset upper link. Seems to me, since all three links are parallel, and since a live axle's outer tube is seriously stiff in torque, it doesn't matter where the top link is mounted. My car drives fine with this offset 3 link, and I can't tell which side I remove the 4th link from...but I have 32 horsepower and high mileage tires (clearly, wheel hop is not an issue for me) and maybe things would be different with a turbo* and slicks.

I want to know what I'm missing. I'm sure I'm not the first person to discover the 4th link is redundant in a Locost, so I'm probably overlooking something. Feel free to start your reply with "Well obviously..." or similar snarkyness but I really don't get it.

*Okay, I already have a turbo, which is why I have 32 horsepower instead of 24; I mean like a turboed Miata or something else in the 200+ hp range.

_________________
Locost builder and adventurer, and founder (but no longer owner) of Kinetic Vehicles

The key to your acceptable performance you've already qualified as being due to the amount of torque the housing is being stressed with. I've been involved in a few rear suspension design situations where the experience with a 3-link was that the housing/pickup tabs were inadequate for the torque. So the solutions are either beef up the center sections with thicker metal and/or add channel beam gussets to the tubes or switch to a heavier OEM housing or switch to either a 4-link or an IRS so that the torque doesn't have to go through the housing.
So, you're not missing anything (except the grin that comes from several hundred ft-lbs of torque).

Makes sense. But I wonder how much torque is transmitted to the housing during bind...in my car at least, I'm sure the torque from dropping one wheel in a pothole is greater than the torque from stomping on the throttle (which in my car is more of an exhaust noise volume control than an accelerator). The torque that the axle housing experiences from power is the torque necessary to keep the diff housing from walking up the ring, and a one sided 3-link would (at worst) have that half of the housing experience double the torque of a well balanced 4-link (which would share the torque equally between R and L sides of the axle housing). As far as adding channel beam gussets to the housing (is that something that is done on high powered cars...drag cars perhaps?) at least I'd only need to add them on one side.

Maybe stick with the 4-link, but change one rod end on each side to a bushing?

_________________
Locost builder and adventurer, and founder (but no longer owner) of Kinetic Vehicles

The housing gussets are common as a retrofit or incorporated into the design (such as Fab9 housings from Chris Alston) on drag race cars, but it was common on Lotus 7s as a way to reinforce the housing when cracks appeared in the stamped steel housing near the pickup for the lower A-bar.
Yes, a couple of rubber-bushed ends would do.

While you are at it you could replace all the rod ends with rubber bushings. Lower maintenance cost. Leaving them on the Panhard rod might be good though.

Vroom, I suspect the rod-ends-are-high-maintenance thing is a believe left over from the previous millennium, from the days when rod ends were metal-on-metal. Nowadays most rod ends have composite liners, don't need lubrication, and last a long time (I have over 100k miles on mine and they are showing no wear or looseness that I can detect). My preference is the minimum suspension smooshiness that works, which in this case is one smooshy bushing on one end of one tube on each end of the axle, which softens the suspension enough not to bind* but is 1/4 as compliant as replacing all four rod ends. Anyway, I'll try it for another 100k and report back to y'all.

*Indeed one rubber bushing on one end of one trailing arm tube softens the suspension enough not to bind, but puts asymmetrical torque loads on the axle.

_________________
Locost builder and adventurer, and founder (but no longer owner) of Kinetic Vehicles

The biggest issue with any 3-link will be the overall torsional strength of the axle. With the Locost 4-link the housing sees minimal torsional loading (well until you have binding). That loading will always be there with a 3-link.

It's not enough of a concern to keepe from doing it - just something to keep in mind.

_________________
-Andrew
Build Log
Youtube

Jack,

A Des Hamill book "How to build and modify sportscar & kit car Suspension and Brakes for Road & Track" (whew) Veloce Publishing, London, describes removing the left upper link of the four-bar set up and replacing it with a shock absorber. The right link forward mount is modified to allow the link to be inclined to balance right-left weight transfer on acceleration. To quote Des (p. 64.) "Unfortunately, such suspension arrangement is not good for circuit work and , more specifically under hard braking (axle tramp), but for drag racing it's ideal."

_________________
Warren
Isuzu Pickup/SR20DE, +401 COLD frame
Build Log: viewtopic.php?f=35&t=11601

If you look at the off-road and rock crawling community you will see alot of parallel 3 links with track bars. They work well because of the lack of binding and simplicity . They are mostly used for the front suspension because the top link can be positioned anywhere along the axle without affecting the geometry allowing for clearance around the engine and trans. The only thing that is an issue with it is that the mounts, both axle and frame need to be strong enough to handle the torque.

Don't forget the rotational torque along the longitudinal axis of the car, which is what load the rear tires unequally under acceleration. This is what having the 3rd link offset to the right side of the car can balance out. But more importantly, don't forget about braking. Acceleration may be limited by the engine, but braking is (or at least, should be) only limited by the tires. And you're now leveraging that braking from the left side of the car through the entire length of the axle. As with any torsion spring, the longer it is without increasing its diameter, the softer its rate becomes and the more total deflection you'll have. So, stiffening the axle with gussets might still be best done on both sides.



Why switch to a bushing on both sides? The 4th link is the only one over-constraining the system. Give it some compliance and the rest should work as intended, while still providing some extra support to the axle when needed...And if not, then stick a bushing in the link on the opposite side too.



Under acceleration, the same amount of torque is being transmitted through the housing, regardless of 3-link or 4-link. The main difference is that each side of the housing is loaded with half the torque in a 4-link, and just one side is loaded with all of the torque in the offset 3-link. So it's not minimal, it's simply half. Therefore, the housing must basically be strong enough to support 2x the load in the heavily offset 3-link vs a 4-link. Braking is another story, but a 4th link that is compliant under acceleration but rigid under braking, would help resolve that.



I don't think Des is giving credit where credit is due. The Jaguar C-Type used a substantially offset third link that fully balanced the torque inputs, to provide equal loading of the tires under acceleration, and was not only successful...But is still considered by many to arguably be one of the best live axle designs. For 'circuit' use the possibility of 'axle tramp' at the onset of tire-lockup provided fewer drawbacks than advantages on acceleration. It was actually for street use, where panic braking was a larger concern, that they stuck with the 4-link. Again, the compliant/locking 4th link would be great....But more realistically, as noted above, I think that simply adding some resistive compliance to the 4th link via a bushing could be an easy compromise to start with.

_________________
-Justin

"Orville Wright did not have a pilots license." - Gordon MacKenzie

It'll depend on the engine and rubber but braking* may be creating a greater moment about the axle than acceleration.

Ultimately I guess it depends on how heavy duty of a rear end you use. It was my understanding that original 7s had an issue since the axle wasn't over built.

*In Jack's case braking is definitely generating more torque than accelerating

_________________
-Andrew
Build Log
Youtube

Bill Shope designed the 3-link used by the pre-IRS Jaguars and another for Mopar's drag racing team. His web site includes a calculator for figuring the links, including torque reactions from asymmetrical 3-links.

His old site from racetec.cc is here:
https://web.archive.org/web/20100203184 ... :80/shope/

His new site is here:
http://www.shopeshop.org/contentsDrag.htm

It looks like some of the stuff from the old site didn't make it to the new one, which is why there are two URLs.

Since there is always more grip on the rear tires under acceleration than under braking; Any car that is capable of at least breaking traction on the rear tires when launching from 1st gear* will likewise experience a greater peak moment about the axle under acceleration than under braking. Fortunately, most modernish rear axles are at least somewhat overbuilt for the Locosts they get used on.



* So most Locosts, but probably not Jack's.

_________________
-Justin

"Orville Wright did not have a pilots license." - Gordon MacKenzie

So where are we on the discussions we had a few years ago with cheapracer who claimed and then demonstrated with life size models that the links should not be parallel on a 4-link, but should instead converge slightly away from the axle? I think this was further supported by examples of Mallocs (hmmm, spelled that wrong? )

_________________
Marcus Barrow - Car9 an open design community supported sports car for home builders!
SketchUp collection for LocostUSA: "Dream it, Build it, Drive it!"
Car9 Roadster information - models, drawings, resources etc.

In my (Jack's) case, I think parking on a hill generates more torque than accelerating.

_________________
Locost builder and adventurer, and founder (but no longer owner) of Kinetic Vehicles