Previous posts on this matter indicate thet this is just the way it is.
Some have calculated the difference and use that as a correction. (eg 99.8% speed change)
As Craig points out, this is unfortunately all too common. Going back to the cine-days, we used the crystal clock from either the recorder to control the camera, or the other way around, i.e. the Nagra taking the sync from the crystal in the camera.
There are some older posts on similar, with comments by Jim Simon, on his adjustment of the remote audio recorder's files. IIRC, it seems that for his equipment, a Time Stretch of 104% for the audio did the trick for him. Each equipment setup might require a different amount of compensation, but once established, should work universally, so long as the gear does not change.
Strange thing is that thirty + years ago we could synchronise independent hardware ( Nagras , Arris etc) without any issue yet modern hi tech hardware is exhibiting problems of this basic nature.
Showing my age I guess.
Yep, we were always able to get the crystals to lock in those days too. All it took was a bit of setup and a cable. Now, it seems to be a bit more difficult, and each mfgr. has their own idea of sync.
Even when we could not cable and lock, once one found sync, it seemed to stay. Had a 2 hour open-heart surgery, where the Nagra was in the OR, but the camera had to be in the theater above the OR, shooting through glass. We did wild sound, and it could not be slated. I found a scene in the beginning with some surgery tools being unwrapped, and used that to sync. Two hours later, we still had good sync, except when we changed mags, or the audio-guy switched tapes. A few jogs got us back in sync, with zero drift. Maybe we just got lucky? We'd budgeted about 20 hours for syncing that footage, and had it done in two.