I also came across this pheonemon some time before. However, it didn't affect much on the analyses. Besides, there is no so-called acceptable % of 219 and 502 relative to 69 suggested in the 5973MSD hardware manual. Wondering if you have tried using this AUTOTUNE value on your routine analyses.
As you may know that vacuum (i.e. column flow rate) affects very much on the tune values. Did you record the conditions, say column flow rate, oven temp, ion source/quad temp., etc when tuning, vacuum pressure? Were they the same every time?
Also did you change the heat filament, the filament position in the ion source? Did you get the similar results?