PS Rack Seal Removal/Install

[OK-944]

So far so good. But, as usual…I’ve saved the best for last! Here it is…the dreaded inner tower seal:



After a bit of head scratching and probing around…I’ve decided to try using my pilot bearing puller:



Placed as follows, and adjusted for moderate pressure against the seal…I will give the vice grips a few upward taps, and if this does not work - I will then heat up the area of the tower which surrounds the seal, to soften up the seal rubber a bit prior to giving the vice grips a few more taps:



The above mentioned “seal heating” worked well for getting the seal out of the rack cylinder (long tube) - so I figure it should work here, although using the bearing puller will be less ideal than getting something to push evenly around the lower seal circumference from the other end. In fact…Bruce Arnn recommends punching out the lower tower bearing plug to get at the seal from that end, but this would still not allow a “full circumference” push - unless something wide enough could be snuck in through the spring cylinder hole.

But back to the rack cylinder (long tube). There is a plastic piece (see photo below), which I’ve seen referred to as a “seal guide,” which sits outside of both the cylinder seal and its plastic backing ring. Looks like this could be guided in from the outside end, with the larger circumference “flange” facing outwards…then squeezed (facilitated by the slots - see photo) to allow this flange to snap into what looks like an appropriately sized slot in the cylinder, prior to dropping in the plastic backing ring and driving the actual seal - which are done from the inner end of the cylinder. But…am I correct about the orientation of this plastic (“seal guide”) piece? I can find absolutely no reference to this - and my act of pounding out the old seal/backing plate/seal guide basically destroyed the guide itself, leaving no clue as to its orientation. At any rate, here is the “seal guide”:

[OK-944]

Success! Took some heat and a few whacks...but it came out! Thing about heat - you want just enough to soften the rubber a bit - but not enough to melt it...which could leave residue behind and make the installation of a new seal difficult:



I think the pilot bearing puller was a good choice here, especially as its jaw tabs are nowhere near long enough to make contact with the ps tower casting. Here is a photo of the inner tower after pulling the seal. The bit of roughness in the lower opening is the casting itself...while the inner circumference of the walls/bed which make actual contact with the seal are smooth. Yay!

[OK-944]

Later:
Yes…that plastic guide went in from the outside end of the rack cylinder, and when carefully driven with a 15mm socket (slightly smaller than the smallest I.D. of the slightly tapered cylinder wall), the flange snapped into place in its groove:



I then dropped the flat plastic spacer ring into the other end of the cylinder, followed by the seal…allowing it to rest above the spacer and orienting it with a pick from the other end - to make sure the “lip” side faced towards the inside. To drive the seal, I first began with an oversized driver…consisting of a 26mm socket (on a long socket extension) - reversed to present a broader “shoulder,” and with this dropped the seal to be flush with its metal locating flange. I then moved to a 22mm socket and, peering down the tube of the cylinder with a good headlamp to assure concentricity, drove the seal down further until it bottomed out on the plastic spacer.

When I use sockets as seal drivers, I like to use them upside down - for a greater contact area pushing the seal. Do note, from the photo…that there is still a very slight amount of taper on the end of the socket’s backside. Despite this, I still find a backwards socket orientation preferable to forwards for seal driving. But YMMV.

[OK-944]

Getting things back together...and I realize that I did not clock the ps tower splined shaft position prior to pulling things apart. Is there a specified position for how the flat part of this splined shaft should be oriented...assuming that the rack is "on center" (rack shaft center-dimple lines up with hole in housing)?

I'm guessing that if there is nothing specified, I can turn the steering wheel until it locks to center...then crawl under the car and map the orientation of the steering shaft splined end (where it joins to the rack's splined steering shaft), then transfer this to the rack tower.

I did remember to gouge witness marks to orient the tower shaft to the steering shaft - but right now I need to know the "centered" orientation of the steering shaft to tower spline set screw, so I can match the flat on the tower shaft to this. Make sense?

[timmay]

Great information! thanks for sharing

[OK-944]

And Later Still:
Very curious…having gone ahead and rotated the steering wheel to lock, then mapped out the angle of the lower steering shaft to rack spindle lock screw hole, and then using this to establish the assembled position of the upper ps tower spindle flat coinciding with rack centering via the inspection hole (see first photo, took a bit of fiddling), I come to find that the witness mark on the spindle cap ends up aligning perfectly with the spindle flat! (see second photo) I have never seen a reference to this, but for now - I’ll assume that I’m on to something!

[OK-944]

Update - after my seat reupholstery adventure, I got back to finishing up the steering rack.

I made sure to assemble this right on the crossmember, to make sure everything fit correctly for later installation under the car.

I painted steel pressure lines rust-proof black, and part of cylinder gray - each after removing corrosion. Also replaced feed and return lines with new ones (needed to bend some of the steel tubing to match the old line profiles), as well as the compensating line.

At any rate...here is one "before" photo, followed by two "after" photos, with the finished rack mounted on the cross member. (Will likely leave it mounted for final crossmember placement under suspended engined, prior to securing engine with mounts):