Caseymacgyver

IMHO, if your setup can go to 60* and you get chatter at 45*, you have some room for improvement in your setup. I do not purport to be an expert here, but I have no steering angle limiters and have my EPA's set to push/pull maybe a hair or 2 before the shaft hits the rim of the CVD cup. Because I have so much front arm travel, I had to very carefully set my front roll center and bump steer to very closely maintain that CVD angle, even as the wheel moves up and down, and the outdrives stay fixed. This reduces the steering angle by design, but this is necessarily true for all cars running non-zero caster. Unless something has come loose, there is no perceptible chatter on smooth concrete, textured concrete or asphalt, but this was not without considerable effort. I haven't had a chance to try carpet, so I can't speak to that. I still can't post links, but I'll try to get a video at some point.

Chatter is not a fact of life. It's true that all single joint CVD's will exhibit velocity and torque variation while rotating whenever there is an angle between the input and output shafts. This fundamental variation will exploit any slop in the car's steering and suspension linkages, possibly causing chatter. As the steering angle approaches 60*, those variations become quite significant (~43% increase going from 45* to 60*) and take an increasing level of attention to every single interface in the front end to mitigate chatter. Needs lots of rotational stiffness at the hub, no slop throughout the steering, optimization of steering geometry, strong servo and minimal flex throughout. Also, to the point of why I posted sexy pics, a posican setup has several key benefits WRT chatter. More on that below.

Not all single joint CVD's are created equal though. Most of the cup style CVD's are self limiting, meaning the max possible angle is where the shaft hits the cup rim. This is very useful as it precludes shaft locking. Generally, the shaft is able to complete a full revolution while in contact with the cup rim, with only the unavoidable velocity/torque variation evident. Specifically for the D3, the stock CVD's do not allow this. The groove on the dogbone (ball side) is cut at a shallow angle, so it hits the cross pin and straightens the CVD joint by 5* or more, twice per revolution. With the steering angle set anywhere beyond that angle, this "Bump" will cause chatter period. No matter how tight the front linkages are, these will rip it apart. The D3 UWA's do not hit the cross pins and have a larger cup. This allows a wider unobstructed steering angle, but also comes with the fundamental increase in velocity/torque variation.

Double cardan joint CVD's, like the SSK's do not exhibit velocity/torque variation as they rotate. At this time, the available options are not capable of achieving the same steering angles as the available single joint CVD's. That said, the SSK's will compensate for a whole manner of setup sins and eliminate chatter with even the slopiest of front ends.

You're right about the geometry, but who says the geometry can't be changed? If the steering tie rod forms a straight line through the leading wheel's steering axis at lock, the effective moment arm goes to zero. But what happens if you move the steering axis inward, then tilt it in such a way that the steering arm actually moves rearward as the wheel turns in? This is one of the side-benefits of KPI, pushed even further with the offset adapter.

In a normal setup (modest caster, no KPI), the driver will generally increase their negative camber to achieve the biggest contact patch possible on the leading wheel at lock. Basically to try and avoid any positive camber through the steering range. But that caster still wants to turn that wheel over, and the sudden change in traction from a full patch to a thin strip as it goes up on the tread edge at lock, coupled with the velocity/torque variation of the CVD's, will, as you say, amplify any steering/suspension weaknesses into a cyclic chatter.

I posted details and pics of my setup because I personally own a D3 and have first hand experience with its idiosyncrasies. I gained a great deal of insight from the information, ideas and experiences of other D3 owners who have shared on this site. Having built upon some of those ideas, I came here to give credit to those who have shared by showing them what they have inspired, while at the same time hoping to inspire others to take this chassis even further.

Finally to the point and why I thought it was relevant here: KPI is very much related posican, which still seems to be a very misunderstood topic. Although the concept is simple, it was quite difficult to accomplish and I tried many, many different variations of front/rear suspensions and steering linkages before settling in on an approach to optimize. In application, each front wheel behaves more like motorcycle tires, rolling smoothly across the tread as it steers. Even though the left-right contact patches are not identical as it steers, traction is continual and changes are minor and gradual. People tend to think that this wheel motion would be harder on the CVD's, but with extreme caster and KPI, it's actually the opposite. By further rotating the hub rearward on the leading wheel, and rotating the hub forward on the trailing wheel, the CVD 'sees' less of the effective steering angle than without KPI.

For me, a posican setup made high speed drifting on asphalt MUCH easier and more controllable. It wasn't about sexy pictures or angle specs. I couldn't get it to do what I wanted it to do when I built it stock. So, I set out to 'Fix' it and this is where it is at the moment. There's still a lot of things I want to try, so I don't think it will ever be done.