A3 Stressbar Experiment: Results (long)
Wayne S. Hill
Y'all may remember a thread here about a month back about the need for
stressbars in VW's of various vintages. I've got a 95 Jetta and, as a
compulsive engineer, just had to quantify what was going on here. Several
folks suggested ways of measuring the deflection of one strut tower
relative to the other (good ideas all), which only spurred me to try
something.
What I came up with is pretty simple. I took a steel shelf bracket, about
1/2" wide and 4' long, and bent it up so that it could reach from one strut
tower to the other without getting in the way of anything (like the coolant
overflow tank, throttle linkage, or hood). I drilled a small hole through
the center of the plastic cap on one strut bearing, and bolted one end of
the bracket to the cap. This permits the strut to rotate freely, while
essentially eliminating any slop between the end of the bracket and the
strut tower (like, at most, .001": it's good and tight). I made a very
simple slider bearing at the other plastic cap of duct tape (this project
wouldn't have been complete without duct tape!), permitting the bracket to
slide as necessary in the across-the-car direction while essentially
preventing the bracket from lifting vertically or swinging fore-and-aft.
Does this make sense so far?
I boinked the free end of the bracket with a center punch so that there
was a detent that would tend to rub on the plastic cap at that end. To
produce a recording of the path that the free end of the bracket visited,
I made up a little sandwich of paper, carbon paper, and plastic film
(vu-graph material: where would engineers be without vu-graphs?). I
taped this down to the plastic cap at the free end of the bracket so that
the pip on the bracket bore on the plastic film and thence on the paper
and carbon paper. The plastic film protected the paper and carbon paper
from eventual abrasion (a matter of trial-and-error).
Does this description make any sense at all? I'm no ASCII artist, so I
really can only use words to describe this.
Anyway, the net effect of all of this is that there is now a very stiff
bracket that reaches from one strut bearing to the other, in such a way
that it makes a permanent mark that indicates the relative motion of the
strut bearings. Now, the strut bearings rotate as the wheels steer, so
without driving dynamics one would expect the carbon paper to show an arc
of a circle corresponding to the radius of the free-end strut cap at
which the mark is made on the paper. With the car stationary, I turned
the wheels lock-to-lock to produce this baseline arc.
Oh, BTW, my other equipment: the car has stock springs and shocks (yech),
Autotech front and rear sways, and broken-in but very good BFG Comp T/A
HR4 185-60 R14's (love these tires, but wouldn't mind wider ones). This
is certainly not a competition setup, but is somewhat better than stock.
Wider tires and lower, stiffer springs might make for a better test
platform (please send cash: oh, and I'll want a set of Bilsteins, too,
so send lots of cash).
With this setup, I then waited for a few dry days to go thrash the car
through some hard turns. This took over a week, cuz it's been raining
like hell lately. Interestingly enough, the tops of the strut towers
stayed completely dry through every torrential downpour, including in
highway driving. Finally, my patience paid off, and I had the chance to
really hang the car out in some highway ramps. All that rain also helped
clean the oil off the roads, so traction was better than usual. Although
I'm not an expert driver, I think I can safely say I took the car very
close to its limits (tires just starting to squeal lightly, and
modulating throttle to maintain track). Although most of the really
on-edge turns were on ramps (long sweepers), I have a few spots where I
can really bang the car around a corner, so I think I visited a good
sample of hard turning conditions.
The results: although the carbon paper produced lines that were a bit
wide, making it a little hard to interpret accurately, the most that I
could reasonably say that the strut towers could have deflected relative
to each other is all of 1 mm, 0.040". I have to say that this was a bit
surprising: although I looked at the A3's structural geometry as being
very stiff, I didn't think it would be _that_ stiff. A deflection of
40 mils, divided by a strut height of about 24" (my tape measure estimate)
corresponds to a camber change of 5.7 minutes of arc. To my untrained
eye, this looks negligible, although Randy might consider this the whole
world (the stock suspension spec's call for -35 minutes + or - 20
minutes).
I performed this experiment solely to satisfy my own curiosity, and make
no representations about any car but my own in this experiment: A1's and
A2's might be totally different, and better tires and springs on my car
(or a better driver) might produce a whole lot more deflection in
cornering (but I doubt it, because skid pad tests don't improve quite
that much with sport tuned cars).
Comments, brickbats?
Cheers,
-Wayne
_______________________________________________________________________
Dr. Wayne S. Hill wsh...@world.std.com
Foster-Miller, Inc. ***Disclaimed*** 781-684-4228
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Todd Walke
wshills...@world.std.com (Wayne S. Hill) wrote:
>Y'all may remember a thread here about a month back about the need for
>stressbars in VW's of various vintages. I've got a 95 Jetta and, as a
>compulsive engineer, just had to quantify what was going on here. Several
>folks suggested ways of measuring the deflection of one strut tower
>relative to the other (good ideas all), which only spurred me to try
>something.
<big snip>
>The results: although the carbon paper produced lines that were a bit
>wide, making it a little hard to interpret accurately, the most that I
>could reasonably say that the strut towers could have deflected relative
>to each other is all of 1 mm, 0.040". I have to say that this was a bit
>surprising: although I looked at the A3's structural geometry as being
>very stiff, I didn't think it would be _that_ stiff. A deflection of
>40 mils, divided by a strut height of about 24" (my tape measure estimate)
>corresponds to a camber change of 5.7 minutes of arc. To my untrained
>eye, this looks negligible, although Randy might consider this the whole
>world (the stock suspension spec's call for -35 minutes + or - 20
>minutes).
>
>I performed this experiment solely to satisfy my own curiosity, and make
>no representations about any car but my own in this experiment: A1's and
>A2's might be totally different, and better tires and springs on my car
>(or a better driver) might produce a whole lot more deflection in
>cornering (but I doubt it, because skid pad tests don't improve quite
>that much with sport tuned cars).
>
>Comments, brickbats?
Wow, there's nothing like an engineer with way too much time
on their hands! <grin>
Seriously, a couple of things occured to me:
Are the plastic caps attached so solidly to the strut bearing
that they aren't moving side to side and thus not accurately
reflecting the movement of the strut? The strut bearings themselves
are mostly just a big chunk of rubber so there'd be some deflection
happening there.
A better way would be to measure the deflection between either
the top of the strut that pokes above the bearings or the metal "caps"
that sit above the bodywork. Don't know exactly how you'd do that
when all those parts rotate.
I'd love to test something like this in an autocross situation
with big sticky tires. I don't think the hardest onramp driving is
more than maybe 60-70% of the corenering capability of the car.
Also, the A3 is probably much stiffer than the A2 and A1 cars
so on those the deflection is probably much worse.
Todd
Seattle,WA
'86 GTI, Red of course. (exciting racey car) 221,000 miles
'87 Golf, Polar Silver. (boring work car) 253,000 miles
Wayne S. Hill
In article <358fe9e6...@nntp.ix.netcom.com>, race...@ix.netcom.com
says...
>
> Wow, there's nothing like an engineer with way too much time
>on their hands! <grin>
True enough! At least I'm keeping away from explosives (well, except
for gasoline, and spud guns, and...).
> Seriously, a couple of things occured to me:
> Are the plastic caps attached so solidly to the strut bearing
>that they aren't moving side to side and thus not accurately
>reflecting the movement of the strut? The strut bearings themselves
>are mostly just a big chunk of rubber so there'd be some deflection
>happening there.
The caps seem to really be on there. With the bracket rig installed, I
couldn't noticeably deflect any of this by grabbing the bracket and
yanking every which way. I agree that the strut bearings must deflect
a bit, but I think my rig should show a bit more deflection than the
strut towers themselves, which is what a stressbar addresses.
> A better way would be to measure the deflection between either
>the top of the strut that pokes above the bearings or the metal "caps"
>that sit above the bodywork. Don't know exactly how you'd do that
>when all those parts rotate.
This would take a bit more work, but could be done. It would be
interesting to compare the standard and sport strut bearings this way.
> I'd love to test something like this in an autocross situation
>with big sticky tires.
Great! When can we expect your report?
> I don't think the hardest onramp driving is
>more than maybe 60-70% of the corenering capability of the car.
No doubt. But if I generated only half the potential side forces, then
I'd expect the max deflection to account for a total camber change of 12
minutes or less.
> Also, the A3 is probably much stiffer than the A2 and A1 cars
>so on those the deflection is probably much worse.
I presume that's where the apparent consensus on the need for stressbars
came from. Does anyone want to measure those cars? I also wouldn't
mind if someone would reproduce my experiment on an A3 for confirmation.
Keith Watson
In article <EuxJ7...@world.std.com>, wshills...@world.std.com (Wayne S. Hill) writes:
|> Y'all may remember a thread here about a month back about the need for
|> stressbars in VW's of various vintages. I've got a 95 Jetta and, as a
|> compulsive engineer, just had to quantify what was going on here.
|> Oh, BTW, my other equipment: the car has stock springs and shocks (yech),
|> Autotech front and rear sways, and broken-in but very good BFG Comp T/A
|> HR4 185-60 R14's (love these tires, but wouldn't mind wider ones).
|> The results: although the carbon paper produced lines that were a bit
|> wide, making it a little hard to interpret accurately, the most that I
|> could reasonably say that the strut towers could have deflected relative
|> to each other is all of 1 mm, 0.040". I have to say that this was a bit
|> surprising: although I looked at the A3's structural geometry as being
|> very stiff, I didn't think it would be _that_ stiff. A deflection of
|> 40 mils, divided by a strut height of about 24" (my tape measure estimate)
|> corresponds to a camber change of 5.7 minutes of arc. To my untrained
|> eye, this looks negligible, although Randy might consider this the whole
|> world (the stock suspension spec's call for -35 minutes + or - 20
|> minutes).
|> Comments, brickbats?
|>
|> Cheers,
|> -Wayne
Cool posting. Thanks for sharing it with us.
I wonder how the deflection would change if you were using tires with more
grip. When I use an H4 type tire it has very little grip so there is only
some lateral force. When I use a high performance tire there is more grip
and thus more lateral force generated. Then there is also using a full on
race tire which of course allows for even more lateral force to be
generated. Also changing from a 185mm width tire to 195mm or 205mm might
have some impact.
It looks like marketing strikes again! :-)
Keith Watson
87 VW GTI 16V