Bump Steer - Theoretical Question

In determining where to mount a steering rack, and the distance between the tie rods, I think I understand this as it relates to Bump Steer, in the 2 dimensional sense- that is, looking at the front of the car head on.  But now, what about when we throw in the dimension of the placement of the rack fore-aft in the chassis?  Presumably, the neutral position -that is, when the wheels are facing straight ahead, the rack and tie rods all line up in a straight line as viewed from the top of the car straight down to the ground- is ideal.  If you move the rack fore and aft I can see it having an affect on Ackerman for sure.  But, what about the affect, if any, on Bump Steer?  

Susprog3d

I know when Ransom and I did it I was shocked at just how much change was made via very small movements in rack placement.  I am sure he will have more input later

Seems there has to be some effect.  Picture the difference between looking straight at a plane containing a circle, holding your viewpoint the same, and turning that plane about a vertical axis such that the circle now appears as a very mild ellipse, almost but not quite round.  The ellipse arc won't quite match up with the circle, and that "error" will be your change in bumpsteer.

Norm

Trent said:

Susprog3d

 

I know when Ransom and I did it I was shocked at just how much change was made via very small movements in rack placement.  I am sure he will have more input later

Was that fore - aft in the chassis, or up - down?  

Up - down, no question, big change.  Fore - aft is a bit murkier to me. 

Fore-aft is a tricky subject, I expect you won't get nearly as much bump steer change as moving the rack up-down. Fore-aft movement will have a big effect on your Ackerman angle and overall steering travel though, that's one of the tricks used to get huge lock angles on dedicated drift cars.

Edit: I should add that most of the travel change on drift cars comes from also moving the tie rod attachment point on the upright closer to the axle line.

In reply to GameboyRMH :

Right, makes sense.

The follow-up to this question is, let's say you have a rack that's too long to get the proper ideal bump - steer setup.  Could you move the rack fore - aft (assuming mounting location and steering shaft length isn't a concern) to correct for too long of a rack?  

Fore-aft will not affect bump steer that much unless you have some really screwy control arm pivoting angles.  You'll notice the Ackerman getting stuffed up long before the bumpsteer effects.

volvoclearinghouse said:

In reply to GameboyRMH :

Right, makes sense.

The follow-up to this question is, let's say you have a rack that's too long to get the proper ideal bump - steer setup.  Could you move the rack fore - aft (assuming mounting location and steering shaft length isn't a concern) to correct for too long of a rack?  

Not really, because the inner tie rod pivot is still going to be way too far outboard.  Depending on how much suspension travel you expect to have, you might be able to get away with just making do and making sure that the "toe error curve" (for lack of a better term) is centered at normal ride height instead of being on the slope.  If you have to fudge it  it's best if the toe stays stable or goes slightly negative in bump, because in droop the tire will naturally be under less load.

In reply to Knurled. :

Makes sense.  I've also read that a little toe out under bump is OK, but toe-in is not.  

EDIT: A bunch of theoretical stuff here that turns out to be not aimed at the practical question that turned up. I agree with Knurled; fore-aft rack motion isn't going to be a useful fix.

Right... As I understand it, which should be understood as a serious caveat...

The painful bit is that while looking at it from the front tells you something, and looking at it from above tells you something, but you really have to work through it in three dimensions. Most control arms aren't parallel to a major axis, and many suspensions' control arms are moving on axes that are not parallel. So the path taken by the outer tie rod end for zero bump steer is an arc matching neither control arm when viewed from the front, and moving more (or less) vertically as the suspension travels because the differing axes rotate the upright as it travels (e.g. anti-dive geometry on double A arms viewed from the side has the upper arm's pivot axis slanting downward at the rear relative to the axis of the lower arm. That means that during bump travel, the upper ball joint is traveling rearward slightly relative to the lower ball joint, which rotates the upright rearward, meaning that on a front-steer car the outer tie rod end moves more in bump than the lower ball joint)

When we were working with this I was really, really happy that Susprog3D has the ability to do some solving based on other constraints. The upshot as I recall it is:

1. It's not entirely intuitive (to me, anyway).
2. While I believe fore/aft affects Ackermann more than it does bump steer, finding the correct rack placement for minimizing bump steer involved a lot of both movement of the rack and adjustment of rack and tie rod lengths. For our project, it was handy that we were working with a Spitfire rack which (for Trent) wasn't a big deal to shorten once we had the numbers. Put another way, I came away with the impression that given the rest of the suspension design fixed, the two primary variables to minimizing bump steer are rack placement (EDIT: meaning fore-aft), and rack length; and that it's not something as obvious as locating the rack end on a line between the upper and lower control arm inner axes.

While we were able to keep bump steer tiny, even given the ability to do a custom rack length and that we were doing custom steering arms for the (two piece) uprights, given the packaging constraints in the wheels at the outside of the tie rods and given the area available for mounting the rack at the inside, we ended up with less positive Ackermann than we were shooting for.

I also wholeheartedly believe that for an amalgamation of parts that can be twiddled to minimize bump steer, it can be worked out in real life with a hinged board rested against the sidewall that allows you to see what the wheel's doing steering-wise as you run it through its travel. And if you're not starting with a blank slate, that last approach may be easier than getting all the measurements into a computer. But that's sort of for the state that you're just twiddling the location of the rack and some tie rod adjustment and leaving everything else be.

Shorten your rack? 

wheelsmithy said:

Shorten your rack? 

I'm leaning that way.  I googled "shorten Miata rack" yesterday and apparently this is a well- traveled road.  

In reply to volvoclearinghouse :

I believe this is in my  future, too. Mr tanner shortened a Miata rack in his excellent Lotus 7 replica book, He cut and welded the actual rack. I would be a might uncomfortable trying that with flux core. Trent has shortened I'm sure, more than one. The one I remember, he used a lathe, and the rack was shortened using mechanical means only. The inner ball joints had male threads, and, I believe, only one side of thr rack needed cutting, a new hole, and re threading. Shortening the housing seems relatively easy as long as it is steel.

In reply to wheelsmithy :

Generally speaking you can only shorten one side. The side with the pinion gear is generally as short as it can be.

I have a sneaking suspicion that for a lot of cars, maybe most or all, control arm pivot placement is dictated by how far inboard the pinion gear can be placed and therefore where the inner tie rod pivot can be placed.  Control arms used to be a lot longer when everything was recirc-ball steering and you could put the inner tie rod anywhere that you wanted.

In reply to wheelsmithy :

Yes, I believe I've seen both of those methods used.  I don't know that I'd worry about using flux core welding- in my job we use it a lot where gas is impractical, for structural welds.  I've flux core welded brake caliper mounting brackets successfully *knock on wood* and I think the loads, vibration, and heat in the steering rack are significantly lower.  

In reply to Knurled. :

Have seen that you can ony shorten one side of the rack.  Luckily for my particular application there should be plenty of room for mounting and dealing with positioning of the pinion.  It's a conversion of a recirc ball steering to a R&P, and yeah, the stock steering just had linkages and levers wherever.  There are some racks that are shorter, and some that they sell pre-shortened.  The Pinto rack is apparently popular for hot rodders.  I think the stock Pinto rack is an inch or 2 shorter than the stock Miata rack; still too long for my purposes but there's options.  

In reply to volvoclearinghouse :

Well, you're doing awesome work with your flux core. I may have a slight prejudice from years of only using them in the field in awkward situations. Carry on/Party on.

And now you know....the rest of the story.