Rack n Pinion Steering Adventure - 1954 Pontiac
Well, what can I say? After 3 months of checking, testing and positioning the rack n pinion, I've come to the conclusion that the engineers in 1947 who designed the first IFS for GM didn't know or care about bumpsteer. I started off with the assumption that the original center position of the inner tie rods would be close to the optimum position - boy was I wrong!
I slowly moved the inner tie rod position outward and shortened the tie rod length while measuring the toe-in/out throughout the suspension travel. I graphed the results and through interpretation (see below) I found the correct steering geometry with the control arms for minimum bumpsteer. It is ABSOLUTELY no where close to the GM specs!
I've switched to a 1986 Buick Park Avenue/Cadillac end steer rack, shortened the tie rod length, used Volkswagon tie rod ends and 1973 Ford Courier steering arms to get the whole mess to work! As soon as I clean and paint the components, I'll post some pictures. I was going over this webpage and realized that I haven't mentioned how I selected the components (which is a pretty important step) so here's a webpage that discusses the criteria and methods that I used.
Here's info on what I found about using a center steer rack and pinion unit and an example of how to adapt it to the original kingpin front suspension.
Positioning the rack
I built a jig that simulates the rack in mid position. The jig is c-clamped to the bottom of the frame rails right behind the lower crossmember.
I started with some large angle iron and found out that I had to add some spacers at the end to drop the jig low enough to clear the oil pan. Then there was interference between the jig and the lower control arm. And finally, I got tired of trying to slip in 3 blocks of wood on each side to lower the jig so I welded in some more angle iron at 22 deg angle so that the tie rods would run straighter.
The jig that is under constant modification!
I originally used a laser to align the front wheels to the rear wheels as a first step. I've documented the procedure on the steering info page. I know use the simpler string method (see previous link). I checked for 0 toe-in using my pathetic toe-in checker made out of an old broom stick handle and two c-clamps. The right c-clamp has the t-handle fixed in place so it can't slide back and forth. The left hand side c-clamp allows the t-handle to slide. I measure the length of the t-handle for toe-in. I used some rubber hose and a wood screw to get around the tire and to the rim. The measurements are at the front of the wheel's rim. I've since made a more sturdy toe-in checker out of the original steering shaft and rebar.
Here's the low buck toe-in jig after many evolutions with many more yet to come...
To position the rack, you have to make a lot of measurements and then graph them. The resulting graph will indicate where the rack or tie rods should move to. You graph toe-in/toe-out versus body position. This means that you start as low as the car will go and raise it in 1" increments while measuring the toe-in. An important note, is that the springs are removed from the car so that the wheels are free to move up and down throughout their travel.
Here's a yardstick held in a drill press vice for measuring height and a jack for raising the body
Some terms that are confusing are bump and rebound. Bump is when a wheel goes over a bump like a speedbump. Under normal driving conditions, the wheel rises up in comparison to the frame. To simulate this, the opposite is done, the frame is lowered in relationship to the wheel.
The term rebound under normal driving conditions describes when a wheel drops into a pothole. The wheel lowers in comparision to the frame. To simulate this, the opposite is done, the frame is raised in relationship to the wheel. The gist of all this is that the lowest point measured on the frame is actually the highest point that the wheel will be. This is indicated on the following graphs with 8" at the top and 14" on the bottom.
The lowest point that I drop the frame was 8", this was as low as my jack could go. The ideal ride height is 11" so I figured that I would measure +/-3" from 11" which works out to 8" to 14". I found that starting at 14" and lowering the frame is much better than starting at 8" and raising the body. The suspension is so stiff without springs to push the control arms down, that I would raise the wheels of the ground at 12" and get false readings. Starting at 14", I made sure that the tires were flat on the ground by pushing them down.
I used Steve Smith's excellent book called "The Trans Am and Corvette Chassis Design/Theory/Construction" as a guide for measuring bumpsteer. I've graphed a few of the many tests (over a dozen) that I've taken and provided a quick explanation under each. Each test takes about an hour to do and I've spent at least 6 hours modifying the jig.
Test 1: This indicates that the rack jig (inner tie rods) is mounted too high. If it slanted to the right, the rack jig would be mounted too low.
When you see an "S" shaped graph, it indicates that there is play in the steering. In my case, one wheel bearing was loose and the right hand side tie rod needed tightening.
Here I've lowered the jig but not enough. Still have the left going slant. Notice the straight vertical section. The jig was interfering with the lower control arm and jamming the control arm in position. I had to knotch the jig to clear the control arms.
Here's the jig almost at the right vertical height, it could be a little lower. Now we have a nice curve appearing. This means that the tie rods are too long. Initially, I had the inner tie rods spaced 2" apart - the same as the rack spacing. The tie rods should be about 6" apart which will shorten them up about 2" each.
This test shows that the correct height has been found by the almost mirror image between bump and rebound. There still is the issue of tie rod length to deal with. If the curve opened to the right side, this would indicate that the tie rods are too short.
There's a couple of ways of "shortening" a tie rod without physically shortening it. You can move the rack farther forward. This may work, unfortunately, it didn't in my case. It looks like I'll have to widen the jig's inner tie rod mounts and shorten the tie rod lengths.
Before I do this, I'm going to do some measurements. The tie rods should run parallel to the lower control arm. This would be determined by a line drawn through the lower control arm's ball joint to the center of the lower control arm's frame mount.
The height of the tie rod should run parallel to the height of the lower control arm. I measured the difference in height from the ball joint and the outer tie rod end and it measured 1". The diffence between the inner tie rod end and the lower control arm inner arc axis should be the same. Now to rebuild the jig to these dimensions and test it in the next week.
Surprisingly, the correct steering geometry followed the upper control arm's arc and not the lower control arm's. The following picture shows the upper control arm and the rnp tie rod connecting to the steering arm.
RnP tie rod follows upper control arm's arc.
You can clearly see the 73 Ford Courier steering arm, the Volkswagon tie rod end and the shortened 86 Buick Park Avenue tie rod and rack end. The rack is mounted just under the frame and behind the lower control arm.
The Volkswagon tie rod end comes in a left and right side. The difference is there is a neat little offset which looks cool. I don't know what the unused bolt is used for on the steering arm. It's frozen in place and has a tiny head smaller than my smallest wrench (1/4") so I can't remove it. Just another conversation piece.
You can also see where the steering shaft connects to the rnp unit on the right side. There's about 1/2" of an inch of clearance between the shaft and the frame. The blue shaft is between the two universal joints. The lower universal joint just about covers the top of the rnp shaft. You can't see the narrow shaft in between in this picture (it's hidden by the universal joint).
I needed to use the 86 Buick rnp for its inner tie rod center to center distance of 25.4", the Courier's steering arm for correct placement of the outer tie rod and the Volkswagon tie rod end to mate the Buick rack to the Courier's arm. The critical dimensions for the rack are the inner center to center distance, the tie rod length, and the thread for the outer tie rod which is usually metric. The Buick uses 14 mm x 1.5 thread.
The steering arm critical dimensions are the length from the upright's vertical arc, the height in relationship to the rack's inner tie rod end, the not so critical Ackerman angle, the horizontal distance between the tie rod end and the rack's inner tie rod and the tie rod taper. The steering arm's placement is super critical in all 3 dimensions! The Courier's taper was 1:10 (1 inch for every 10 inches) and inner/outer hole dimensions.
The Volkswagon's tie rod end had a taper which closely matched the Courier's taper, had a short length of 2.910" and used 14mm x 1.5 thread. Plus it was a lot cheaper than the BMW 320i choice.
Steering arms starting at the bottom: 70 Nova, 63 Biscayne, 70 Ford van, 69 Ford 1/2 ton, 69 Impala, 63 GMC 1/2 ton, 67 Chevelle.
I tried all of the above steering arms and they all had problems with interference or positioning. The Ford arms were used upside down as their taper is the reverse of what I needed. Interestingly, I noticed that 67 Chevelle's upright/spindle could be used as a dropped spindle for 68-74 Novas and early Camaros. Rather than have the spindle at the bottom of the upright, the Chevelle's spindle is in the middle about 2" higher. May be worth investigating...
RnP installed on mounting plate and crossmember
The rack is mounted on 5/16" thick steel plate that has been plasma cut to shape by my friends at Duratec. The front is at the bottom of the picture, the driver's side is on the right. There is a small plate welded onto the front with two mounting holes. This forms a lip so that the steel plate is mounted flush with the bottom of the original front crossmember. I've used the original steering mounting holes in the front crossmember.
The rear of the plate is welded to a 1" thick steel tubing crossmember that is held in place by angle brackets bolted to the frame. There is very little clearance between the rnp and the oil pan. I will most likely end up knotching the oil pan when the engine mounts are tightened up. The plate rests between the two lower control arms and pretty well covers the bottom of the engine compartment from the front of the deepest part of the oil pan forward.
Knotched oil pan to clear rack n pinion unit (upside down on engine stand)
The rnp mount is bolted to the frame with 8 x 3/8" grade 8 bolts. The rack is bolted to the mounting plate with three 1/2" and two 3/8" grade 8 bolts.
Power Steering Pump
I hooked up the 76 Camaro's power steering pump as it had the correct pulleys and mounting brackets. I had a 1985 Park Avenue rack's pump that was identical except for the pulley and mounting bracket. I switched the pressure bleed off spring between the two.
The return line has little pressure and can be hose clamped. Normally, there is extra line looped in front of the rad for cooling. I will be using a small tranny cooler instead. I had to have the pressure line made for $60. The guy who made it was pretty helpful. He suggested that pressure fittings be used at the hose ends so that the pipe ends can be swiveled for a better fit. That little suggestion sure made a big difference when I hooked it up.
Mystery Electrical Connector
On the rack's steering shaft housing is an electrical transducer of some type that has 2 electrical connections. My first thought was that it was some sort of speed governed power assist control for changing the feel of the steering the faster that you drive. In fact, it is a hydraulic pressure sensor that detects high pressure when the engine is running at idle. It is used to shut off the air conditionner pump to reduce the load on the engine under high steering efforts (high pressure) at idle. Mystery solved!