I don’t think PC would have a different effect between AMR or legacy codecs. The step sizes and factors should be the same… The thresholds are different.
Now, you could take into acocunt that AMR is more resistant, so you don’t need to have the PC tuned as reactive as when used with Legacy codecs. This is where you have to do trials… though I doubt you’d find much improvements.
For the moment we are using fix power change step size, but in which cases is recommended to use variable step size? Is it true that the power control can happen only in intervals of 480 ms, in both cases?
yes, the power commands are sent at the SACCH interval, every 480ms.
the variable or fix step size refer to the amount of power that cn be increased or reduced.
If UL RXLEV is very high, then the BSC will tell the MS to reduce its TX Power:
– if fixed step : power is reduced at a rate of “n” dB every 480ms. Usually, n = 2dB. so it takes 5 seconds to decrease power by 10 dB.
– if variable step : power is reduced by “k” * (rxlev_target – rxlev_measured)
k is usually between 0.6 and 1.
So it takes usually 1 or 2 SACCH to reach the desired power.