I did a lot of sputter and i noticed that Vbias wasn't stable occuring the deposit (decreasing) at differente pressure 10-20-30 x 10-3 mbar and RF power 60 W
Check your target is not cracked and that clamp and shield are properly mounted, with no debris/flackes in between. If not, secondary plasmas and sparks can occur and make the voltage unstable.
If you vary the process parameters, in this case the pressure, it's completely normal that different selfbias voltages will result from that.
Also, the selfbias not being constant throughout the process is normal, I also observe it to start at a higher value and then converging to a lower one.
However, at identical nominal process parameters the value towards which it converges should always be the same, if that's not the case, Filippo Annoni 's suggestions are the hottest candidates for what's wrong.