Vintage DD turntables. Are we living dangerously?

Aigenga.
re unipivot arms.
You raise an interesting point, different arm bearings may well show a difference in the graphs. (Arm geometry may also show difference, as BT argues.)
That said, have any of the rigs recently listed by Halcro had unipivot arms?
The W&F visible on the raw data for the TT-101 seems to be quite similar from arm to arm. This would imply that the speed changes are localised in the drive unit itself.

I believe that Halcro posted three gimballed arms - the only unipivots that I noticed was the VPI on the Classic Direct. It showed a rather ragged trace.

I lean towards the idea that the regular speed changes on the Feickert traces are the results of record eccentricity; the fact that they occur once per rotation (16 - 17 per 30 second measured period) tells me that they reflect an out-of-round condition and are not due to speed correction feedback which would probably occur far more frequently.

Continuing my last post, I believe that the out-of-round condition must be in the record not the table because records are imprecise and therefore very likely to have that problem and if both the table and the record were at fault the trace would show them either reinforcing or canceling each other depending on where the record was on the platter - so every time you measured the same record on the same platter but rotated the record relative to the platter you would get significantly different results. This is easily tested.