'Thought I'd pass something along for the OP. I just finished replacing the Cave drives with Byers drives on my Astrola mount with 1.5" axis. I want to discuss my findings related to the pier-mount attach point in some detail as I suspect this may be your problem.
Originally, the mount was mated to a casting that fit on the pier, let's call this casting "the cap." The top portion of the cap consists of two parallel vertical plates drilled through for a 3/4" bolt. The interior spacing between the two vertical plates is 7/8". The RA casting has a tab on the bottom that fits between the two vertical plates that is also drilled for a 3/4" bolt. The thickness of this tab is coincidentally 3/4". That means, in the absence of spacers, that there is a 1/8" void when the tab is mounted to the cap with the through bolt.
Cave supplied two spacers to take up this void, one on either side of the cap. One of these appears to be a relatively thin washer of teflon. The other, long destroyed on my mount, was if memory served a thin washer of bakelite. Together, they filled the void. Note, however, that both of these materials are relatively soft. This has two important ramifications. First, they have a certain springiness when the bolt is tightened. This will allow vibration of a certain small but definitely detectable amplitude. Second, they both can be damaged (and mine were) by the two retraining machine screws just above the 3/4" bolt on each of the two vertical plates.
OK, what not to do. First, DO NOT torque that 3/4" bolt tight without some form of spacers to fill the gap. You may stress the casting and crack one or both of the vertical plates. What if you skip the spacers but don't tighten the bolt much? In that case, you will experience a large-amplitude, relatively slow (~1 s) vibration that will take 8-10s to settle. Is this what you saw? Consider the frequency f of a pendulum of length x with gravitational acceleration g: f = 1/(2*pi) * SQRT(g/x). We have g = 980 cm/s^2 and x (the vertical height from the bolt center to the mount center of mass, which is about where the "tee" of the mount is) ~ 10 cm. We then have f ~ 1.6 s. Is this about what you saw?
I fixed this problem in combination with the latitude adjust issue. I made a 2" wide by 3/32" thick by 3/4" hole washer using a large, old washer buried in my spares collection but drilled out on a drill press (a drill press is mandatory to do this) using a spade bit. Yeah, I know you machinists out there will object but it worked. Incidentally, the standard 1/8" thick washer available in stores was too thick for me to slip in, hence the need for a 3/32' washer (it needn't be round on the outside). Then I attached an eye bolt to the RA housing plate and another to the pier, and connected these two with a small, stainless turnbuckle. I arranged the geometry of the turnbuckle to be parallel to the RA plate so as not to put any "push/pull" stress on the relatively small 8-32" machine screws that Cave used to bolt the RA plate to the RA axis casting (the one with the tab). I ca now both adjust the latitude easily and smoothly while eliminating the side-to-side play of the RA axis between the vertical plates. The two set screws on each vertical plate are now superfluous. Incidentally, my time-to-settle dropped from 3 s to 2.5 s (at high power) from the original "soft" washers Cave supplied.
I sense your frustration, and I hope this solves your problem. Really, properly tuned, the old Cave 1.5" mounts are quite wonderful and very stable.
Happy observing always,