Remember to store the tape tails out.
This advice probably sounds odd without explaining the reason, so I'll elaborate. The reason for doing this is that all the recorded tapes develop print-through, which creates a faint copy of the recorded signal from one turn of the spooled tape to the former one and to the next one. This is audible as an echo and a pre-echo. For some reason the print-through is stronger on the inner neighbouring turn than on the outer one, so you mostly hear a pre-echo, which is unnatural. One way to solve the problem is to store the tape tail out in order to promote the echo instead of the pre-echo, another way is to wind and rewind the tape from time to time, which smears the print-through as the new turns will not be exactly at the same places as the former ones, so both echo and pre-echo are reduced. Note that on B-wind machines - the old European broadcast recorders that played the tape oxide-out instead of oxide-in - the tape should be stored tails-in instead of tails-out.
Revox roller bearing guides which I used on both feed and take up sides aren't that accurate and at high speed the tape can shift back and forth almost nudging or nudging the reel flanges which can cause tape drop outs on channels 1 and 4 with 1/4 trk tapes.
The Revox in & out tape guides are not the best on the market, no doubt about that. However I definitely don't blame them for the tape going up and down at high speed. I blame the tape itself, and more specifically the slitting. Evidence of this is that the same tapes that don't pack evenly on my Revoxes, don't pack evenly on my broadcast recorders either. Conversely, the tapes that pack evenly on my broadcast recorders, also pack very evenly on my Revoxes. Experience has taught me and many others that the European tapes (Agfa, Basf, Emtec, RMG and RTM) pack well at high speed, while 3M and Ampex/Quantegy are prone to winding problems leading to almost systematic tape spillage when fast winding. BASF produced a video showing the difference on a Studer, I will find the link for you if you want. Common wisdom is that the different slitting quality is due to the widespread use of flangeless AEG hubs in Europe, which are unforgiving to packing problems.
My tests on both 1/2 track and 1/4 track have shown that the Revox guides are only related to a phase fluctuation higher than on a broadcast recorder, but this minor issue can be solved rather easily.
Plus I got better signal to noise, better dynamic range by being able to place more signal on the tape using 2 trk.
More specifically, consumer 1/2-track (2 mm wide) is 3 dB less noisy than 1/4 track (1 mm wide). Wide tracks (2.75 mm wide) add an extra 1.4 dB at the cost of slightly reduced channel separation.
I never used 10 1/2 inch plastic reels as they had a tendency to have warped flanges.
My experience is exactly the opposite: metal reels were almost all warped due to shocks after being used for a while, whereas the plastic reels remained flat. I don't use cheapies, only good quality reels made by Agfa, which probably explains our different experiences with plastic reels.