OK, I did the homebrew alignment last night, after also replacing two
idler arms and putting the best two tires on front. BTW, thank-you Dan
Saitz for emailing to me the alignment specs for my van!

First, I used my 4' box-beam level to locate a level parking lot. Had
to go to a church about a block from the house. On the way home, I
took the van up to about 50 MPH on several smooth roads. It pulled
very strongly to the left, and I thought I could hear some scrubbing
from the tires.

On the driveway at home, I removed the right front wheel and loosened
the two mounting bolts on the upper A-frame. These bolts ride in
slotted holes in the vehicle frame, so that the front and rear mounting
points can be separately adjusted by pulling/pushing toward or away
from the vehicle centerline. So, you adjust caster and camber
simultaneously.

I pulled the upper A-frame all the way to the outside stops. This
would result in maximum positive camber (top of tire leaning outward)
and an unknown (but hopefully small) amount of caster. Then I slightly
snugged the front A-frame bolt to hold it in place, and pushed the rear
mount toward the vehicle centerline by a "calibrated eyeball" amount to
increase the caster. Then I slightly snugged the rear bolt, loosened
the front bolt, and pushed the front mount of the A-frame toward the
centerline by HALF of the "calibrated eyeball" amount. My idea here
was to achieve a net reduction from max camber while preserving half of
the added caster.

I then repeated the procedure on the left side, re-installed the
wheels, grabbed a flashlight, and headed to my level parking lot for
some measurements. I noticed an immediate improvement in handling.
The pulling seemed to be gone, though it was hard to tell at 30 MPH on
bad roads. There did seem to be a lot of "wandering", though.

With the help of my girlfriend, I made camber measurements straight
forward and turning to the stops (maybe around 45 degrees) on both
sides. By having her steady the vertical 4' level against the fender
while I used the flashlight to watch the bubble and scoot the other end
around on the ground, we were able to come up with a reference that was
"plumb" (at least in the plane perpendicular to the vehicle) and within
about two inches of the wheel. I made distance measurements from the
vertical to the rim at top and bottom, to calculate actual camber.
I calculated camber of +1.6 degrees on the right and +1.2 degrees on
the left. The tops of my tires were still leaning slightly "out".

The caster calculations were more problematic because of measurement
uncertainty. When the wheel was turned, this resulted in fender
obscuration of the point where I wanted to set the level. So, this
measurement distance was several inches further away, probably adding
error. I was, however, consistent in seeing a decrease in positive
camber (top of wheel tilting relatively inward) of both wheels when
they were on the outside of a turn. This decrease was "maybe" 0.6
degree on one side and 1.0 degree on the other side. So, I felt I at
least had achieved positive caster. Everyone says that caster is a
handling factor but not a wear factor, so I figured I'd fine-tune this
by road testing.

Next we took several toe-in measurements, and found the wheels to be
toed out by about 1/4".

I drove back home (tools and jack were still on the driveway). I
adjusted the toe to zero by shortening the tie rod on the left side
(remembering the earlier drastic left pull, I was thinking about
centering the steering wheel).

Time for a road test - and the van drove like a dream. No pull, all
the way to 65 MPH, and the wandering (probably due to the toe-out) was
also gone. You could take your hands off the wheel and it went
perfectly straight. And, the wheel was centered.

Then I made a mistake. Thinking about the tire wear due to the excess
positive camber, particularly on the right, I decided to make a small
change and take another measurement. Since decreasing right camber
would supposedly make the vehicle pull a little to the left, I thought
I'd simultaneously decrease the caster on that same wheel (several
references indicate that a vehicle pulls to the side with increased
camber or decreased caster) to compensate.

Well, that was just too many variables to deal with when it's after
midnight and the upper A-frame is free to move in two dimensions at
once. My next road test had the vehicle drifting noticeably to the
left. You just can't make a small, precise, "delta" change without
some sort of fixed reference or simultaneous measurement.

Anyway, it took another round of wrenching (and was about 1:15 am)
before I had the driveability back to where I was satisfied.

This weekend I'll try another idea for fine-tuning the camber, which is
still excessive by anyone's account. I'll use a magnetic-base stand
from a dial indicator to position a pointer directly above the grease
fitting of the upper ball joint. That way, I will be able to tell
exactly how much relative shift I've made in the A-frame before
re-tightening the bolts.

Yes, it was a lot of work, but I really feel good about the learning
experience.

best regards,

Martin

Martin wrote:
> I recently replaced upper ball joints on a '96 Dodge 3500 (1-ton) van.
> This vehicle uses pressed-in ball joints, and I replaced them with a
> standard C-frame tool. After a few months, one of the ball joints
> backed itself out (this is a poor design, IMO). I pressed it back in,
> and tack-welded it in a few spots. So far, it's holding just fine.
>
> In order to do the weld, I removed the upper A-frame. Of course, I
> didn't think to mark it's position (uses no shims, just slotted bolt
> holes to adjust caster/camber) until I'd already loosened the bolts and
> let things shift.
>
> I'd like to get the truck properly in alignment. If that's not
> possible, at least driveably close until I buy new tires and have a
> shop fine-tune it.
>
> I think I can measure camber with a plumb bob, and toe-in's a
> no-brainer, but how in the devil can I measure caster?
>
> Also, is there an online reference where I can find the alignment specs
> for my vehicle?
>
> best regards,
>
> Martin