As of yesterday, I successfully, though laboriously, replaced all the front lower control arm bushings. I had previously replaced the front suspension's coil springs. About a 3/8-inch to 1/2-inch height difference still exists between driver's side and passenger's side, using my crude measurement techniques. The car does seem to handle better. It seems to take road bumps with smoother bounces rather than some clunkiness prior to the new bushings.
For the front passenger side lower control arm bushing outer sleeves ( = second side I did), the new air hammer-chisel I purchased wasn't enough. Nor were the new diamond cutting wheels used on the air die grinder. Instead, I mostly hand sawed and hand chiseled about one-third of each bushing outer sleeve, constantly applying PB Blaster. The PB Blaster was being sucked into the crevices between sleeve and arm; I think it helped. Had I more time, I would have let the PB Blaster sit over night. Hammering against a well-fitting socket (with outside diameter almost exactly that of the outer sleeve) seemed to finally shake the remains loose. All together it took hours to remove the two old, bushing outer sleeves. I think the smaller one was the tougher one, perhaps because its small size means it has less "springiness" to it.
I beat up the smaller control arm bushing hole pretty badly from my efforts. I filed a bit and cleaned both control arm holes with emery paper.
I heated the control arm for an hour in an oven set to 200 degrees F. I also froze (overnight) the new bushings. I started the larger bushing by manually hammering against a well-fitted socket, making sure the bushing was going in straight. Then I used the socket-bolt-nut-washer method to press it in. This worked well. While it did not just "slide in," progress was steady, noticeable, controlled, and much less labor intensive than whacking repeatedly with a hammer. I do think the heating and freezing helped. The larger new bushing was installed within an hour or so.
By the time I finished installing the larger new bushing, the control arm was about at ambient outdoor temperature again. For the smaller bushing, I did try heating the arm again but hesitated to leave it in the oven for too long with the new, larger bushing already installed. I do not think 200 degrees F is enough to destroy the bushing, but I was a bit worried about degrading the rubber somewhat.
I was on schedule for finishing the one control arm within a day (12 hours). But then installing the new, smaller bushing became a problem. I think the control arm hole just had too many nicks and burrs in it. I could not get the new, smaller bushing to line up straight for quite awhile. Eventually I got it started. I used the socket-bolt-nut-washer set up, and saw some progress. But pressing it in was taking more force than usual, perhaps because of the aforementioned nicks and burrs. I was using a roughly five-foot pipe extension throughout. After getting the bushing in about one-third of the way, I had stripped the 3/8-inch, fine thread, Grade 8 bolt/nut.
About the force required to push a bushing into place -- Not having access to a hardware store at 7 PM at night, I gave up on the socket-bolt-nut-washer method for the evening. I did try (1) a vise (which seemed to be working but also seemed to be breaking the vise); (2) a C-clamp (same); (3) rigging my car, with a jackstand for backup protection, to apply about a quarter of its weight (about
0.4 ton) to the bushing. This weight did not budge the bushing.
In the Usenet archives I read about a guy who used the weight of the "back end" of a Cadillac to press in a control arm bushing. I googled for Cadillac weights and estimate the typical Cadillac weighs a bit over twice as much as my Civic (3.6 tons vs. my Civic's 1.6 tons), so this guy applied maybe 1.8 tons or more. The guy wrote the Cadillac nearly lifted up before the bushing started moving.
The 3/8-inch bolt never yielded in tension (that is, pulled apart). This suggests the force the socket-bolt-nut-washer setup applied to the bushing was probably less than 7 tons, by my calculations using various, basic bolt formulae and theory. The bolt/nut threads are not supposed to strip until about 9 tons are applied. I figure I had tilted the nut somewhat when tightening, causing the stripping to occur. The washers and other doughnut shaped items beneath it were in fact yielding (= bending) somewhat, = causing things to get a little uneven. (The sockets by contrast held up well.) So with the nut under load and tilted relative to the bolt axis, at high enough loads, the threads would strip.
The next morning I went to the hardware store and bought two more 3/8-inch, fine thread, Grade 8 bolts. The second attempt with a 3/8-inch bolt advanced the bushing further, about 3/4 way altogether. Then this bolt and nut stripped. I finally got the bushing fully into place with the last
3/8-inch bolt.
I think if I had taken more care to keep the socket-bolt-nut-washer setup straight, the two bolts might not have stripped. New and tougher washers, or bolts of easier lengths with which to work with the sockets I had, might have prevented stripping.
To me the lessons here are (1) For a younger car driven in a non-corrosive environment, either a 12-ton shop press or the socket-bolt-nut-washer method might very well work to remove old bushings. This is not necessarily so for an older car driven in a corrosive environment such as the Midwest or Northern U.S.
(2) A 12-ton shop press would likely successfully and easily install new bushings. I do not think a two-ton press would always be enough, based on my experience and calculations and reading about others trying two-ton presses.
(3) Even if one goes with the shop press, one will have to buy the sockets or other bushing press accessories.
(4) Based on my calculations of the tons of force required, I have doubts about the wisdom of using any hammer to beat bushings into an older control arm already scuffed up from bushing removal. One has to possess a great deal of upper body strength and a lot of endurance, not be susceptible to heart attacks(!), etc. OTOH a newer control arm not beat up by the bushing removal effort might be a piece of cake.
In other words, perhaps if my 91 Civic were only around seven years old (vs. 15) and had not been driven in a corrosive environment for most of its life I would not have had so much difficulty removing the old bushings.
If I had to do this over for an older, rusty bushing sleeved-car, I would strongly consider buying used control arms with not more than a 100k miles on them from a nice low humidity, non-corrosive climate etc. part of the country. I'd replace the bushings in these arms at a casual pace. Remember that the old bushings in the bent junkyard control arm (the one I used to work out a methodology) came out much more easily. I think an air chisel would have had those out pronto. My hammer and old screwdriver worked fine.
Buying spare control arms would keep my car available. Then I'd just swap the arms.
Buying new arms with bushings already installed IIRC is also an option. The rear arms for a 91 Civic are less expensive than the front ones. Considering all the trouble to which I will go to cut rear control arm bolts out, I think I might go this direction for at least one of the rear control arms. The other rear control arm is already a junkyard one whose bushings may press out easily. From previous work on it a few years ago, I think the arm's bolts will come free pretty easily.