Made amazing progress today. I played with bump steer most of the morning
in software simulation. I took the drawings from a newer TR4 shop manual
and enlarged them with a copier, playing with the enlargement ratios
(copying enlarged copies) until they were pretty much life size. I
measured everything very carefully and came up with a small correction
factor for the drawing that I could apply to every dimension of parts and
get a number very close to actual. That gave me a good notion of the
trunnion centerline for steering input as well as more accurate centerline
of the ball joint. with the revised number the bump steer graph looked
quite a bit more like the numbers I was getting from the actual
measurements.
I played with a lot of variations and got nowhere until I tried raising
the rack a little--then I saw some improvement. Tried moving it back and
forth and that didn't do much. So I raised the rack on the actual car
about 1/8th" and got a big improvement. Another cycle of trying to match
actual to simulation and I got a near exact match when I used the
intersection of the trunnion centerline and the steering arm as the pivot
point (I was using the intersection of the trunnion centerline and the
pivot point of the trunnion axle). Then I played with the simulation again
and found 3/16" up and 3/8" forward on the rack gave me a near zero bump
steer except where the suspension rebounds to near the lower bump stop.
Tried it on the car, and Eureka!
I'll keep playing, but I'm amazed at how good the graph looks now,
especially in the range of typical travel. Might be a tiny bit squirrely
when it rebounds from a big bump and extends to near the stop, but it
probably won't feel bad since the car will be light on it's wheels at that
point. Even at it's worse it's a tiny fraction of the original hideous
curve.
Small changes, big difference. It's nice to actually have a good result
from all this work.
Kas, I'll write all this up in the stuff I'm working on. There won't be
any directly applicable learning for typical Triumphs--not many of you
have some strange Japanese rack in your TR3. But I think I can generalize
the process. At the very least I've figured out how to represent Triumph
trunnions in suspension software.
|