818 bump steer

[Larry J W]

Here is a quick discussion on what I found on bump steer. The 818 started to get a little more twitchy so I checked the toe in. The original alignment was 1/16 inch toe in. When checked, the toe in was1/4 inch. Evidently the suspension had settled a bit over the years. Suspecting a bump steer problem I checked the toe in with the car jacked up to the point the wheels were just touching the ground. At that point the toe OUT was 1/4 inch! There was a 1/2 inch change!

This weekend the bump steer parts arrived and I made a gauge to measure it. The following graph is the results.

150796

Spacing is the distance from the bottom of the steering knuckle to the center of the heim joint.

From the fully extended suspension system (0) to almost fully collapsed 5 1/2 inches there was a 3/4 inch change in toe in per wheel. The width of the tire is about 24 inches. The change in toe in was measured on a 20 inch board. Needless to say it was shocking how bad it was! Yet the car handled great in the curves.

A second issue that made the problem worse, was to get ride height I was 1/2 inch from the limit of travel. The absolute worst place to be on the curve. The plan is to drop the ride height 1 inch and take the occasional scrape. This would put me in the flat part of the curve. Each data point is 1/2 inch of travel.

With 2.24 inch spacing the car is more stable. Before the change the car seemed to want to dive into curve. Probably what was happening is loading the outside tire with excessive toe in it grabbed the pavement more pulling it deeper into the turn. I have done some dumb things with it that scared me and should have spun out, but it has always behaved perfectly!

I now feel in hard corners it wanting to slightly over steer. I know there have been problems with bump steer from the rear resulting in people spinning out mid curve

Moving to the future, first is lower the ride height, realign the suspension, and correct any problems with the rear bump steer. The learning never ends!

[J R Jones]

Larry, This is confusing to me. I am still assembling my 818 and have not gotten to bump steer, but I have done other cars. I have seen other 818s with spacers between the steering arm and the tie rod, I assume that is part of their bump steer solution.
That is consistent with my experience. In many cases the bump steer solution makes the tie rod align with the lower control arm (LCA). Therefore the steering rack is (about) in line with the LCA pivots and the tie rod end is (about) in line with the LCA ball joint.

Your tire diameter is 24 inches? Your target board is 20 inches wide?
Your steering knuckle is the upright? The bottom is the flat surface above the ball joint? You replaced the tie rod end with a Heim joint?
My 2002 Impreza tie rod end is about 0.91 in below the steering arm. Is that where you are referencing "spacing"?

Your 2.24 inch spacing is close to what I saw in 818 pictures, and it is your best graph.
I use ride height as my "0" point and measure bump steer up and down from there. That puts you at -0.1 to +0.1 which is not bad. Once I get to my bump steer, I will find out how close to "0" it can be.

Unless you have significant chassis roll in turns I would not expect roll-steer.

Suggested ride height is 4.5 inches, I would not run lower than that on the street.

Bump steer in back is minimized by aligning the outer lateral rod pivots with the inner rod pivots.
Measure the inner pivot pitch angle, (or heights to floor) and adjust the upper track bar to make the outer pivots on the upright exactly the same pitch angle.
jim

[STiPWRD]

Is there another spacer that would put you on a more flat part of the curve at your current ride height? Thanks for posting the data btw.

[J R Jones]

Larry, It would be helpful to know where your ride height is on the graph.
The 0-5 inch is wheel travel defined by the shock absorber?

How are you measuring the bump steer? Dial indicators?
I use a laser fixed to the hub, projecting on the target field to the side.
Ride height position is traced on the board, then displacement is measured as the beam goes off the ride height line.
Practical distance of the target field is based on your tire diameter, but fine tuning can be done by moving the field further away.
jim

[Larry J W]

The zero on the graph was defined by the shock removed and the suspension hanging lose with no tire. This point was consistent. Bump steer was measured on a 12X20 melamine board with a dial indicator on one end and a stop on the other, similar to a commercial bump stop gauges. The board was not quite as wide as the tire but close enough for what I wanted to do. I looked at using a laser but the other system was quick and easy for the moment. Current ride height on the graph is 1/2 inch which is part of the problem.

I used the pinto pins per the recommendations in other post. These are a little large and do not seat completely in the steering knuckle. There is about and 1/8 gap between the seat and knuckle. Therefore spacer length is meaningless. Instead I measured from the bottom of the knuckle to the center of the joint. The actual spacer is 1 inch.

Another issue that may be adding to the bump steer is the caster is 6 1/2 degrees. This may be increasing the bump steer also.

I hope this clears up some of the questions.

[J R Jones]

Larry, I am not familiar with pinto pins, but you are right to measure total spacing.
6 1/2 degrees caster is a lot, "S" spec is 3-4 degrees. That promotes self-centering and requires more effort to steer.
Big caster has an effect on the toe dynamics by raising the steering arm as the upright tips back on top. In this case it raises the location of the steering rod end. Not a big number.
jim

[Larry J W]

The pinto pins have a similar taper as the tie rod ends on the Subaru, just a slightly larger diameter. Rather than risking screwing up a steering arm by drilling them out to 5/8 inch, I went with the pinto pins. I believe they are the same as the mustang pins so they were easily available. It is good to not burn too many bridges when making changes!

When I built the 818 in 2014 the ride height was set at 4.5 inches but I scraped at the bottom of the driveway if not careful as it went from 20% grade to level. Then there are the speed bumps! By cranking the springs down as tight as possible that problem went away and caused another.

When professionally aligned the castor was set at 4 degrees and the ride height at 5 1/2 inches but it did not feel right. To compensate, the castor was increased to 6 1/2 degrees. In hind sight it was a a band aid to cover the root cause. Steering effort is not too bad. Unless I learn something different I'll stay with the existing setting.

Larry

[J R Jones]

Larry, careful about matching tapers. I did a street rod project, building upper control arms with QA1 ball joints and housing. They had to fit in Lexus uprights but the Lexus (metric) taper did not match the SAE taper on the QA1 ball joint shaft.
I bought a 7 degree taper reamer to make the QA1 shaft fit the Lexus hole. You are OK if the Pinto taper is the same as Subaru. jim

What is the application for Pinto tapered pins? Steering or ball joint? Ball joint shafts are typically larger.

[Larry J W]

The pins are for correcting bump steer. As close as I can measure they are 7 degree taper (act 7.15 degrees) or 1.5" per foot or 8 to 1. Per the specifications the root of the pinto pins is .630". I measured the root of the Subaru at .61". Working through the the math they should bottom out at .16 inches short of full insertion. While I did not measure it, it is real close to what was observed.

What I do not like is very little thread shows above the nut. When the dust settles I may have them turned down to correct any potential problems.

[J R Jones]

Speedway has a 7 degree reamer, but using it to make the hole larger and deeper is not easy. The reamer likes to feed itself into the hole.

[DMC7492]

Look at this thread about bump steer
https://thefactoryfiveforum.com/showthread.php?22469-bumpsteer-kit

[RPGs818SNA]

DMC7492, I’m surprised that so few builders seem to be correcting the basic cause of bumpsteer the way you did, especially after Hindsight’s extensive posts (1164 to 1194) on his build thread. See https://thefactoryfiveforum.com/showthread.php?15217-Hindsight-s-build-thread

The way I look at it, the bottom ball joint of the spindle moves in a circular arc based on the lower control arm’s axis and length. The upper ball joint also moves in a circular arc based on the upper control arm’s axis and length. As a result, the tie rod ball joint also moves in an arc, although not strictly circular due to the different arm lengths. To minimize bump steer, one needs to match the length and center of rotation of the tie rod to best approximate the natural arc of the spindle’s steering ball joint. Thus properly positioning the joint at the end of the steering rack (in all 3 dimensions) and adjusting the tie rod length to match should minimize bump steer.

DMC7492, how did you determine how much to extend your rack and how much to move it up?

Thanks for any insight.
RPG

[Bicyclops]

DMC7492, I’m surprised that so few builders seem to be correcting the basic cause of bumpsteer the way you did, especially after Hindsight’s extensive posts (1164 to 1194) on his build thread. See https://thefactoryfiveforum.com/showthread.php?15217-Hindsight-s-build-thread

The way I look at it, the bottom ball joint of the spindle moves in a circular arc based on the lower control arm’s axis and length. The upper ball joint also moves in a circular arc based on the upper control arm’s axis and length. As a result, the tie rod ball joint also moves in an arc, although not strictly circular due to the different arm lengths. To minimize bump steer, one needs to match the length and center of rotation of the tie rod to best approximate the natural arc of the spindle’s steering ball joint. Thus properly positioning the joint at the end of the steering rack (in all 3 dimensions) and adjusting the tie rod length to match should minimize bump steer.

DMC7492, how did you determine how much to extend your rack and how much to move it up?

Thanks for any insight.
RPG

Very interesting. Not sure how to measure that and where else to start but spacer length.

I wonder if anybody has a good starting length for bump steer spacers for the S ride height?

Ed Holyoke

[Ajzride]

I wonder if anybody has a good starting length for bump steer spacers for the S ride height?

Ed Holyoke

46mm

https://thefactoryfiveforum.com/showthread.php?31639-Aj-s-818S-in-Houston&p=419290&viewfull=1#post419290

[Bicyclops]

Thanks for that.

Ed

[DMC7492]

RPG. The process is time consuming but not difficult once the correct measuring tools are used and the suspension is aligned. I followed these instructions.https://www.longacreracing.com/technical-articles.aspx?item=8162
I then used set collars welded to plates to hold the rack and clamped the plates to the frame. Once the measurements are documented, the chart showing the corrections needed are implemented, until the bump steer is minimized. The rack mounts were modified or replaced and the rack spacers were machined and ne tie rod ends were cut shorter,along with shortening the tie rod adjuster.
This car-is special. The corrections were very close at ride height but as the suspension is compressed or extended the bump steer changes, it was a give and take.
The original numbers were severe and would snap the steering wheel into your hand. Along with the steering column loose in the bottom grommet. Driving on a Wisconsin pot hole road was terrible. On a concrete highway was perfect.
151679151680151681151682151683
I believe there is not a set dimension as ride hieght effects the position of the rack

[RPGs818SNA]

Thanks so much for your post, DMC! The link to the instructions is most helpful, although it will take me a while to absorb it. Your work looks awesome. Thanks for sharing the photos. Your process makes a lot of sense, although I imagine it took quite a bit of work to implement. How much wheel travel up and down did you consider as you were minimizing bump steer? Did you have to check with the wheels turned left or right, or only straight ahead? What’s your ride height? We don’t have so many pot holes in North Carolina, but I want to get the suspension fixed such that most anyone can drive the car comfortably.
RPG

[J R Jones]

RPG, DMC,
I have yet to tackle bump steer on my 818, but obviously it is advantageous. I am perplexed that there is this packaging compromise in a custom design.
I have a replica of a 1929 Mercedes SSK built on a Studebaker chassis. I have integrated Lexus IS300 power and suspension/steering/brakes into it.
The bump steer was accomplished with a laser beamed from the wheel hub to a target field on the side. As the hub "bump steers" the beam goes off the ride height reference line.

The mechanical solution was complicated by replacing the lower control arm (LCA) trunnion pin with a ball joint, the shortened Lexus upright and a fabricated upper control arm (UCA).
Furthermore the Studebaker control arms sweep aft and the Lexus rack/pinion (R&P) was too wide.
I modified the R&P, replacing the tie rods with an "intermediate rod" or "drag link". Mounted ahead of the R&P, it gave me a greenfield to mount my tie rod inner pivots.
I benefited from the basic geometry of the IS300, the steering arms were perfect.
In the end I tested inner tie rod mount locations and tie rod length until bump steer was gone.
I will explore this technique on the 818.
jim

151686
151687
151685

[Dave 53]

I'm enjoying the discussion and respect the effort! But, for those that might be freaking out on the complexity of this like I was...

I did the Baer solution and called it good. My car drives and steers great.

[RPGs818SNA]

OK, Dave, Jim is freaking me out too with the complexity of his solution. It sure sounds a lot harder than the Baer solution, but I'm intrigued by the rack joint relocation solution.

Just for fun, in order to put some numbers on the concept, I measured the location of my control arm axes and ball joint centers and both tie rod ball centers as best I could. I only considered vertical and horizontal as viewed from the front, as 3 dimensions makes drawing it so much harder. After drawing lines representing the control arms, spindle, and tie rod at ride height, I moved the LCA up 3” and down 2” in 1” increments, drew the new spindle lines, and adjusted them to join the UCA, keeping all the lengths the same. Then I located the spindle tie rod joint center on each of the spindles, shown in orange. Next I drew the tie rods from the rack to the spindle and noted how much too long they were and the toe out they would cause. The numbers seem in line with what’s been reported, that is, not so bad for good roads but a problem with big bumps and holes.

Then I took the coordinates of each of the natural spindle to tie rod joint centers and computed a least squares best fit for the rack end location and tie rod length. The best fit for the rack joint was 2.75” to the right and 0.05” down. Using the new tie rod length, I redrew the tie rod lines in green and noted the length error and new bump steer results. The numbers are all smaller, most right much smaller, but there is some toe in/toe out which I hear is undesirable. Moving the rack joint a little less to the right and making the tie rod longer could make all the results toe out, with some additional toe out at the extremes.

151722

This was entirely a drawing and mathematics exercise, and did not take into account the third dimension, but I think it shows that it may be possible to substantially improve bump steer by simply moving the rack end joints to a better place, like DMC7492 did. Experience has shown that the Baer solution also works to reduce bump steer, but I don’t know why. Maybe that’s my next drawing project.

RPG

[DMC7492]

Ok I’m running high ride height 5.0-5.125” which is close to max and a stronger spring. SO. The compression and extension is probably less. There fore keeping the arc swing angle of the tie rod minimal.
This explains why every car will have different dimensions based on the ride height. I think the total travel of the suspension in real life is probably less than 1” extend and 3” compression. So the Baer spacers probably work due to minimal travel. In my set up the outer part of the front tires burn on the fender flares, so a limiter may be needed for compression. Or make the fenders wider like on the new style body

[J R Jones]

RPG, I guess you had to be there. (looks like the administrator did not like my photos)
I am OC about structure, and relocating the rack is a PIA.
Relocating the tie rod end on a spacer, below the steering arm gives me pause. Strength, deflection and deformation concern me. An OEM would not do that.
I am surprised that Subaru did not locate the steering arm (on the upright) on the same plane as the lower ball joint. That is (dynamically) a stable location.

The reference points I use are shock absorber stroke.
Measure shock mount eye to eye length at ride height.
Measure extended shock length eye to eye.
(without the spring) measure compressed shock length eye to eye.
Measure bump steer to these dimension extremes with ride height as the "0" point.

Being a skeptic, whatever package solution you choose, I would measure the results.
jim

[RPGs818SNA]

Jim, your solution was probably harder to understand without the pictures, but throwing in alternate parts and fabricating arms and having things work well is way out of my league. I am thoroughly impressed by (and totally envious of) your capabilities. Thanks for sharing your experiences.

I share your concern about extending the steering arm so far downward and generating additional stress on the spindle. I'm even more concerned that doing so doesn't reduce bump steer as much as moving the rack ends, although it may make the bump steer more acceptable. I'm still trying to figure that out.

I imagine the Subaru steering arm location above the lower ball joint was OK with a McPherson strut since that doesn't deform the steering arm arc as much as the short upper control arm does.

I can envision that moving the rack vertically or front to back any substantial amount would be a PIA. It appears to me that bump steer is much more sensitive to moving the rack than it is to lengthening the rack ends. Wouldn't that be as simple as using a threaded couple and threaded rod to lengthen the rack, since the tie rod joint relieves any bending or torsional stress on the rack end?

It sounds like you consider where the wheel CAN travel rather than where it is likely to travel. I might be willing to compromise with more bump steer at the extremes for less at the normal travel range. Maybe that's a mistake if big bumps and holes is where you need minimal bump steer the most.

I agree that any solution needs to be verified with measurement and road testing. I look forward to reading about how you tackle bump steer on your 818.

RPG

[Larry J W]

An update on correcting bump steer. The ride height has been lowered to 4.5 in the front and 4 in the rear. This gets me up and down the driveway without scrapping if I’m careful. I checked the bump steer in the rear and it was pretty good. When attempting to reduce it even further, the coil over did not fit in the lower bracket. A good reference point for the rear is if the lower bolt on the knuckle is level.

The result is the 818 is quite stable. It also appears to be neutral in handling. This is based on some spirited driving in the canyon where I live. There are lots of 15 and 20 MPH curves to verify the low speed handling and smell the heated tires. As for finding out which end brake loss it was not the place to try it with rock walls on one side and nasty drop offs on the other!

Jim, you had me second guessing the 6.5 degree caster I am running. A little research says run as much as you can steer. Even with a smaller steering wheel the effort is no problem because the front is so lite. Even for a 76 year old man. The new C8 vette runs 7.4 degrees castor. 6.5 degrees is the maximum you could get without modifying the upper arm.

I am not worried about the spacers to correct bump steer. The weak point is shear as the hiem joint on the original tie rod end. The extended pin does add bending to the stress. Using spacers greatly reduces the bending stress. I would be a little more concerned using a double nut like some of the systems.

Moving the rack and pinon or extending the pivot points is a much better solution. On the other hand “Done is better than perfect”. As bad as it was it still handled better than anything I have driven. As Jim says measure! Ride height and wheel alignment will affect how you solve the problem.

This is a good discussion to point out problems that many people don’t realize. Gee the 818 is twitchy, but why? The 818 is a diamond in the rough and through discussions like this we can refine it into a great car.

Larry

[RPGs818SNA]

Hmm. Doing the same graphical analysis after dropping the spindle steering joint 1.69” per the FFR recommended Baer solution, I get the following toe changes with wheel travel:

1.54” Toe out @ +3 compression
0.99” Toe out @ +2 compression
0.48” Toe out @ +1 compression
0.00” Set to 0 @ 4.5" ride height
0.46” Toe in @ -1 droop
0.85” Toe in @ -2 droop

That’s what I expected, given the geometry, but not what I expected given the actual experience of those that implement the Baer solution. Does anyone understand what’s going on here?
RPG

[Ajzride]

I dropped mine 62mm or 2.5”. I’m suspecting that will give you numbers that are closer to “stable”

[J R Jones]

RPG, Larry, Yikes, your graphed numbers are extraordinary, an empirical test is in order before reaching conclusions.
Your logic is spot on. I had not connected the steering arm position to the original strut configuration, I believe that you are right.
Yes, bump steer reduction is realized more often with moderate suspension travel, and maybe a little bump steer at the extremes is acceptable. Big bumps are always problematic.
I can't say what adjustment is more significant, tie rod or rack. In the end, if you find a solution, you have to decide if the ends justify the means. Read on:
On rack/tie rod dimensions. In many successful steering systems the ends of the rack (inner tie rod joints) are the same width as the distance between the lower control arm pivots. Then the tie rod length is the same as the lower control arm. Then if the lower control arm and tie rod are at the same angle, there should be minimal bump steer. Raising and lowering for angle.

On alignment and caster, Cobra guy CraigS is very bright and made this observation, paraphrase:
The Cobra is not straight line stable without a lot of castor (6+ degrees) but steering effort becomes so great that autocross requires power assist.
I found one day, jockeying into a parking space, steering lock to steering lock, that big caster numbers lift the inside (front) fender and lower the outside (front) fender. The increased effort may be from lifting the front end.
Experimentally I reduced the caster and the lift/lower went away.
This was in response to a theory that you should run 2/3 of kingpin angle or SAI, (steering access inclination). In that case caster would have been 8-9 degrees!

I have never run over 4.5 degrees caster, so I do not have a learned opinion on extreme caster.
jim

[RPGs818SNA]

OK, I think I found my explanation in Appendix H 818R Supplement of the manual. I hadn’t read that part as I’m building an 818S.

To run at track height, the control arms are raised an inch or so, raising the spindle relative to the frame. However, the rack is not raised, so the spindle steering joint has to be lowered about an inch with the Baer bump steer kit (see Rev Z page 611 Parts and page 616 showing the kit installed). In addition, per notes on page 634, the castor should be increased, which raises the steering joint as well, needing more lowering with the kit. Finally, on page 634, FFR recommends toe out bump steer on compression which is achieved by lowering the steering joint even more. So that’s why the Baer solution works at track height.

I'll rerun the analysis taking the arm raising and extra castor into account and see if the numbers are more reasonable.

For an 818S at 4.5” ride height and factory castor, I suspect lengthening the rack around 2.75” may be the better alternative. I don’t think the Baer solution would reduce bump steer in that instance, although it may cause more desirable bump steer resulting in improved handling and feel.

Azride, I assumed a 2 7/16th spacer, or 2 13/16 from the bottom of the steering arm to the ball center. The stock ball is about 1 1/8 below the arm, so i used 1 11/16 as the additional lowering of the ball. I think our numbers are pretty close. What's your ride height? Did you raise the control arm axes per the 818R instructions? How does your steering feel during bumps?

Jim, I agree that aligning the tie rod with the lower control arm would make bump steer go away. I wonder if FFR did that on any of their other models.

Larry, I can't wait to get to the point you are with a stable ride and tight canyon curves to drive through. What fun awaits.

RPG

[Ajzride]

I'm at 4.5" on all corners with no change to control arms. There is no bump steer that you feel when driving, and Ive tested on some roads that have pretty bad surfaces.

[J R Jones]

RPG. The process is time consuming but not difficult once the correct measuring tools are used and the suspension is aligned. I followed these instructions.https://www.longacreracing.com/technical-articles.aspx?item=8162
I then used set collars welded to plates to hold the rack and clamped the plates to the frame. Once the measurements are documented, the chart showing the corrections needed are implemented, until the bump steer is minimized. The rack mounts were modified or replaced and the rack spacers were machined and ne tie rod ends were cut shorter,along with shortening the tie rod adjuster.
This car-is special. The corrections were very close at ride height but as the suspension is compressed or extended the bump steer changes, it was a give and take.
The original numbers were severe and would snap the steering wheel into your hand. Along with the steering column loose in the bottom grommet. Driving on a Wisconsin pot hole road was terrible. On a concrete highway was perfect.
151679151680151681151682151683
I believe there is not a set dimension as ride hieght effects the position of the rack

DMC, I have noticed in your photos that our steering racks are different, maybe chassis too. I see you have the alloy (2006) control arms.
My 818 is a 2014 #291. My donor is a 2002.
My rack is shorter, and the inner tie rod joints align with the lower control arm pivots at 27 inches. My sense is my rack will make correcting bump steer easier.
I have purchased the alloy control arms, as far as I know they should be a direct replacement.

Where in WI are you?
jim

[lance corsi]

For anyone interested, there is a great book on suspension geometry that answers all the above questions and offers solutions. The book is called CHASSIS ENGINEERING by Herb Adams. It’s a paperback and reasonably priced. It explains bump under steer, bump oversteer, roll center, instant center, and much much more.

[RPGs818SNA]

Thanks for the book tip, Lance.

The diagram below shows a recomputed best location for the rack end joint for a 3.5” ride height achieved by raising both control arms per the FFR Appendix H and using the Baer solution to lower the spindle steering joint 1”. (That requires a 1.75” spacer.) The bump steer is virtually eliminated by also moving the rack end joint over 2” and down 0.2”. That places the rack end joint very close to the lower control arm axis, which Jim pointed out makes fixing bump steer easy. Since moving the rack end joints down is hard, the better solution is to only lower the spindle steering joint 0.8”.

151883

I have assumed +3 degrees caster. Additional caster will raise the spindle steering joint 0.09” per degree. That would need to be added to the Baer spacer. I also did not add desirable toe out on compression. That can be added by moving the steering joint over less than 2” (maybe even 0”).

Please note that these numbers are all mathematically derived from my approximate measurements and should be considered only as a help in understanding various bump steer solutions.

RPG

[RPGs818SNA]

If you are interested in trying out the rack extension method for reducing bumpsteer at a 4.5” ride height, here are several ways to make them. Mechie3 machined a short rack extender to fix the early rack misplacement problem. It’s beautiful.
152158

DMC7492 made his own rack extender for his bump steer solution. I wish I could make one like that.
152159
Here’s an inexpensive rack extender that doesn’t require machining.

152165
It consists of 1 and a half 1.9” threaded couples and a 2.75” fully threaded bolt with the head cut off. It also requires a longer bellows which DMC7492 found for us.

Here are the parts and sources assuming M16 or M18 threads on the rack end. I really hope it’s M16.
152166
M16 90009A225 $4.00 McMaster-Carr M16 Class 8 1.89” long Need 3.
M18 90009A233 $33.60 McMaster-Carr M16 Class 8 2.12” long Need 3.

152167
M16 91280A273 $8.11 for 5, McMaster-Carr M16 Class 8, 2.75” long. Need 2 of the 5.
M18 91280A997 $13.76 for 5, McMaster-Carr M16 Class 8, 3.15” long. Need 2 of the 5.

152168
910-34347 $11.99 Speedway Motors Need 2.

It seems like this would be strong enough since there are no bending or torsional loads on the extender. Now we just need a way to easily shorten the tie rods by around 2.75” or find shorter replacements.
RPG

[DMC7492]

Hi guys, Jim I’m up in Chase Wisconsin north of Green Bay!

Tie rods are easy too but need a little work. The originals re threaded does not work so well as can be seen the threads are into the Hex. Not good.
New tie rods will be shortened along with cutting the tie rod end. The maximum allowable threads into the tie-rod. I think I drilled the new tie rod ends with the proper size bit and threaded deeper with bottoming tap. These were the test parts and the new tie-rod ends will be cut to match the test part.152202152203152204152205

[J R Jones]

Two years later I got around to measuring and correcting the bumpsteer on my 818 chassis build. I reviewed this 7/21 thread by Larry J W, 10/16 by Lumpyguy, and 5/14 by C Plavin plus suggested links.
There were various solutions by spacing the tie rod end and substituting a spherical rod end for the tie rod. Jim Schenk weighed in stating that is a good solution and spacing the rack is likely no better.
Interpretation of builder geometry plots and modification results are obfuscated by no consistent metrics. I used my favored method to document four configurations.
I attach a laser to a wheel stud and project 25 inches to the side resulting in toe change X 2 on a target page with a vertical reference line. The suspension is stroked from bottom to top shock limits without the spring.

818 as designed:
193739
193740

2 7/16 spacer on tie rod (bottom of steer arm to pivot ball centerline)
193741
193742

1.8 spacer on tie rod (bottom of steer arm to pivot ball centerline)
193747
193743

Rack spaced on mounts 0.65 in and shifted 0.75 forward
193744
193745
193746

The last two traces are similar. I did not have to change the tie rod so the joint structure is not compromised by nut/bolt/spacer.
The rack mount is shaped to conform to the rack isolators and attachment structure is not compromised by nut/bolt/spacer.
jim

[Dave 53]

Excellent work!

[Larry J W]

Great work Jim. You did it the right way. I did it the quick and easy way. I may put a little bump steer back in the system this spring. It felt good with a little over steer where it wanted to dive into the curves.

A heads up, bump steer in the rear is real bad!

Larry

[J R Jones]

L J W, Thanks.
Bump steer in the rear may be what Sgt Gator warns us about and it causing snap oversteer. The rear (lower) lateral links keep the knuckle from steering but if they are not the same exact length, they can steer the rear in turns.

If you want more "turn-in" you might experiment with "0" toe or a little tow-out in front. Before that you might explore less castor.
Oversteer in a mid engine car can have consequences.
jim

[Dave 53]

Clarification needed...

It seems if the rear has zero toe, that lateral arms are by definition, equal length.

My understanding was that rear bump steer was caused by the inboard lateral link mounting points and the outboard lateral arm mounting points not being parallel to each other. Which is adjusted by the upper trailing arm. In other words, the lateral links have a bit of a twist.

I found without enough castor, my car was a dancing bucking bronco under acceleration. Dailing in as much castor as possible calmed it right down. If you intentionally reduce castor, I'd suggest you hold on and do some hard accelerations on the test drive.

I'm going to be doing some spring swapping soon and when I have the coil overs out, I'm going to do your laser on the stud bump steer test front and rear.

[lance corsi]

The front wheels should toe in, not out. The rear wheels have no provision for steering unless you are talking about toe-in under compression, in which case, the outside wheel receiving the compression should toe in slightly. Toe out in the rear under compression or otherwise is bad. Under compression, the wheel most heavily loaded should increase toe-in under load. This helps steer the rear of the car in the same direction as the front wheels. Without this, snap oversteer is likely.

[Dave 53]

A bit of front tow out makes for a great feeling "turn-in" which is desirable on a track, autocross or if one simply likes the feel of it (which I do). The downside of too much toe out is straight line instability. I have 0.11 total front toe out. Turn in feels great and the compromise in straight line instability is minimal on hundreds of freeway miles to the track and at 125 mph on the track straights. If I trailered my car to the track, I probably would have a bit more front toe out.

I assume the factory alignment on my 2023 Toyota Corolla is toe in which is appropriate for a mass produced daily driver. If I ever enter the Silver State Classic, I think I'd go for a bit of toe in.

When you get your alignment (and on corner weight scales), make sure to be sitting in the car and maybe half a tank of gas. I make up almost 10% of the weight on the suspension, so it makes a difference.

[J R Jones]

Clarification needed...

It seems if the rear has zero toe, that lateral arms are by definition, equal length.

My understanding was that rear bump steer was caused by the inboard lateral link mounting points and the outboard lateral arm mounting points not being parallel to each other. Which is adjusted by the upper trailing arm. In other words, the lateral links have a bit of a twist.

I found without enough castor, my car was a dancing bucking bronco under acceleration. Dailing in as much castor as possible calmed it right down. If you intentionally reduce castor, I'd suggest you hold on and do some hard accelerations on the test drive.

I'm going to be doing some spring swapping soon and when I have the coil overs out, I'm going to do your laser on the stud bump steer test front and rear.

David, absolutely correct, I once knew better and checked my suspension then forgot the detail you mentioned. So the lateral links should be parallel in yaw (azimuth) and pitch.
I am not a fan of that design so I am changing it, ultimately to control arms.

BTW when I replaced the 2002 steel stamped front lower control arms with 2006 alloy LCA I found the front bushing mounts did not align with the rear Z bracket mounts. I had to elongate the front bracket holes to swing the Z brackets out about 0.6 inch. PIA.
jim

[RPGs818SNA]

What an elegant solution to bump steer Jim! I especially like your laser on a stud solution to measuring steering deflection. It’s simple, understandable, repeatable, and sufficiently precise. I also appreciate your dispelling the idea that lowering the steering arm end is substantially different from raising the rack. I think your solution results in a stronger mechanism too. Your rack supports are a work of art. It doesn’t appear that they would cause you any radiator mounting issues.

I have a question about your process. I understand using various block sizes between the tie rod end and the spindle to determine what size minimizes bump steer as evidenced by the laser plot. It sounds like you ended up with a 1.8” optimum block. You then reported “Rack spaced on mounts 0.65 in and shifted 0.75 forward” and appear to have about the same tiny amount of steering variation through the wheel travel range. How did you get from a 1.8” block size to your rack adjustments? Are they vertical and horizontal rack displacements?

Thanks for sharing this excellent bump steer solution.
RPG

[Dave 53]

I just checked my bump steer front and rear. My method isn't as sexy as using a laser, but it provided some rough data. I measured the angle of the chassis with a phone app and shimmed the ruler so it was perfectly perpendicular to the chassis.

Rear has an extremely small amount of compression toe in. Perfect.

I have the Baer kit with all of the spacers. Front still has toe in under compression. Nothing that I notice while driving. I didn't take any measurements or do any math to quantify. If I add any more spacers to the Baer kit, the nylock part of the bolt won't be grabbing any threads. Is it safe to add spacers and just count on thread lock? (I'm thinking, yes).

Or if Jim goes into production...

Under compression 194187

Relaxed 194188

Measuring setup 194189

[J R Jones]

What an elegant solution to bump steer Jim! I especially like your laser on a stud solution to measuring steering deflection. ItÂ’s simple, understandable, repeatable, and sufficiently precise. I also appreciate your dispelling the idea that lowering the steering arm end is substantially different from raising the rack. I think your solution results in a stronger mechanism too. Your rack supports are a work of art. It doesnÂ’t appear that they would cause you any radiator mounting issues.

I have a question about your process. I understand using various block sizes between the tie rod end and the spindle to determine what size minimizes bump steer as evidenced by the laser plot. It sounds like you ended up with a 1.8” optimum block. You then reported “Rack spaced on mounts 0.65 in and shifted 0.75 forward” and appear to have about the same tiny amount of steering variation through the wheel travel range. How did you get from a 1.8” block size to your rack adjustments? Are they vertical and horizontal rack displacements?

Thanks for sharing this excellent bump steer solution.
RPG

RPG, Thank you. First a note on measurement resolution.
The axis of toe rotation is a line through the ball joint centerlines (duh). My laser target was a convenient 25in plus distance. IMO toe is a function of tire radius, or 25% of circumference. That means my graph represents twice the actual toe change.
I speculate that the 2002 donor WRX did not have much bump steer but I never had it with the (2014) unfinished kit I bought. I could not measure the donor for reference.
I theorized that FFR copied the control arm mount geometry, but did not copy the rack mount geometry. The rack mount is the B/S root cause, not the steering arms.

My review of forum data showed two dimensions for the 818 tie rod B/S spacer, 1.8in (46mm) and 2.65in. I measured those spacers with my laser as reference, not a target. I wanted consistent method/metrics. That had no bearing on my rack reposition.
In my professional life I did advanced development including proof of concept vehicles and one-off running prototypes. I worked mostly without a net (support) but with an aggressive timeline so in most cases it was empirical design and validation.

In this case I removed the rack from the FFR mounts and put it on a motorcycle chassis lift. I jacked it up 2 inches and flipped the FFR B/S curve.
I dropped the rack to 1 inch and the curve was pretty straight.
I made shims and used clamps to hold the rack on the FFR mounts. The 0.650 wood shim at the FFR mount angle has a vertical hypotonus of about 1.2 in. That curve was pretty flat and 0.70 and 0.60 shims made the curve worse. I built the aluminum blocks at 0.675 thick plus appropriate stands for securing the Subaru isolator clamps. That dimension was not as good as the 0.65 wood and I fine tuned the blocks to that dimension. (measure twice, cut once)
Where the blocks attach to the FFR mounts was another opportunity tor tuning. I tested in 0.62 increments, three up and three down. In the end the Subaru clamp mount boss centerlines are about 0.75 forward of the FFR mount boss centerlines. The steering column joints have to compensate, but it fits.
jim

[RPGs818SNA]

Thanks so much for explaining your design approach, Jim. I suspect your empirical design, validation, and implementation approach is the fastest way around most problems.

I spent some time measuring my standard FFR rack mount, which is a chore at this stage of assembly with the body, radiator, battery, etc.in place. I drew that up and overlaid it with what I believe to be an approximation of your design.

194235

Given the dramatic reduction in measured bump steer, I suspect your design would reduce experienced bump steer to between acceptable and negligible. Getting out that last little bit of curve would require lengthening the rack and shortening the tie rods, no small task.

I imagine the rack would need to be raised another inch for race configurations using the upper wishbone mounts, but a revised bracket should be simple to implement. There are some variations in rack design verses donor year and model, but again easy to compensate for.

Overall, I think you have introduced a simple and effective alternative to the Baer solution. Well done.
RPG

[J R Jones]

RPG, your suggested rack position will interfere with the 818 frame structure.
Your through bolts contribute to that and ignore the stagger between frame fasteners and clamp fasteners.
My blocks have (4) threaded bosses, the upper clamp bosses are off-set forward ~0.75 inches.
Also the left (rectangular) clamp needs a bit of edge removed from the corner flange to clear the left diagonal frame tube.

I am not concerned about using the higher UCA mounts. That position may require a bit more vertical shimming; the UCA has less impact on bump steer than the LCA.
I do not know how repositioning the UCA impacts the steering arm B/S spacer length.
jim

[RPGs818SNA]

Sorry Jim, I knew something was wrong, but didn't understand the staggered holes. I corrected my drawing. I'm sure it's still imperfect, but hopefully explains the concept of your design for the benefit of others.
RPG

[lance corsi]

The uca does affect bump steer. As the lca rises, it describes an arc larger than the arc followed by the uca. If the uca causes camber increase, and the lca does not follow(larger radius) the distance from the steering knuckle to the steering rack changes, thus causing bump steer. The trick is to position all three elements such that the distance between the steering rack and the knuckle remains constant throughout the travel.

[J R Jones]

Qualifying the control arm influence on the knuckle camber and B/S, the (vertical) distance LCA ball joint to tie rod ball is 2.5 in. The distance from the UCA ball joint to the tie rod ball is 13.5 in. 5.4 to 1.
Rosanna Danna: "It just goes to show ya, It's always something".
Extreme castor will lift the steering arm, altering the relationship to the steering rack.
My geometry has been established prior to alignment, so I will measure it again, post alignment.
Lance, have you evaluated bump steer on your Corvette suspension/steering?
jim

[lance corsi]

Yes as well as I could, using my autocad. I must however correct my earlier post regarding bump steer. Since the uca and lca are different lengths, the resultant effect is a modified arc that the knuckle travels in. My main objective was to try and match that arc to the radius of the tie rod. It’s a compromise at best. I was primarily concerned with the numbers mainly through the usable suspension travel, not the full articulation, which is more than my suspension travel. It took a lot of fiddling to minimize b/s and was even more challenging to reproduce on the car since the “v” of the frame was what I had to accommodate for my attachment points. My biggest downside was the wheel offset used on the original corvette wheels. I couldn’t buy wheels with the required back spacing other than corvette wheels(inexpensive wheels) to eliminate scrub radius, so I will eventually need to buy wheels with more offset in the future. Probably some 3-piece wheel$. Oh my! Mo money! I tried the corvette wheels, but much too wide. Also GM mounts the uca at an angle to introduce anti-dive, which I felt unnecessary with such a light car, so I mounted my uca level to minimize the confusion of increasing caster, etc.

[lance corsi]

Fwiw, I settled on a 5 degree caster angle. And just for giggles, I measured my wheelbase and just by luck it measures within .100” of perfect square. Anyone else measure their wheelbase?

[J R Jones]

Fwiw, I settled on a 5 degree caster angle. And just for giggles, I measured my wheelbase and just by luck it measures within .100” of perfect square. Anyone else measure their wheelbase?

Symetry is a trickier evaluation with IFS and IRS, considering all things adjustible. My chassis suspension mounts are spot on, but by F to R wheel hubs differ 0.25, again without adjustment.

I bought a (repaired) C4 Corvette as a donor that had a RF accident history. As It came apart I also found the floor pan broken under the driver seat.
The RF frame horn was accordianed behind the front crossmember resulting in ~0.75 short wheel base on the right. A buddy was a professor at the local tech school body shop. I let him use the car for school demonstration purposes but their straightening left it ~0.375 short on the right. I ended up building a new frame anyway as my Avanti project required a 102 inch wheelbase.
jim