SNAPIN reads the magnetics waveform MV-VS-SL to obtain the surface voltage Vsur. The surface voltage is comprised of two components: a resistive part, required to sustain the toroidal current against plasma resistivity, and an inductive part, which produces a time rate of change of magnetic stored energy. Because high electron temperatures are achieved in TFTR, the resistive equilibration time is rather long--of order seconds. As a consequence, the current profile, the inductance parameter l i, and the magnetic stored energy can have significant time derivatives long after the plasma current has reached a flattop.
SNAP uses the voltage for two calculations: to calculate the ohmic power input, and to calculate the plasma resistance. The appropriate input for both calculations is Vres, the resistive component of Vsur.
If no special action is taken, SNAPIN will simply pass along to SNAP the voltage it reads from MV-VS-SL as the resistive voltage, Vsur. This will cause SNAP to miscalculate the ohmic input power and the plasma resistance. In beam heated plasmas this is not a significant problem since typically , but it can be important for ohmically-heated plasmas.
If you manually enter zero as the surface voltage in the magnetics menu of SNAPIN, a special subroutine VSURLAM will be activated to correct the Vsur value to obtain a better estimate of the resistive component. VSURLAM reads in the time history of Vsur, the Shafranov Lambda (), Ip, R, a, the toroidal field current, and . It calculates li from and . Then it corrects Vsur to get Vres by subtracting the component associated with d/dt of the magnetic stored energy. Also, an attempt is made to correct for the fact that the surface voltage as reported by waveform MV-VS-SL is D/Dt (poloidal flux between magnetic axis and a), while the true surface voltage should be evaluated at fixed toroidal flux.
The correction is:
where is the total toroidal flux in the plasma,
. The
VSURLAM routine allows you to use a smoothing time for the
waveforms which enter this calculation different from the averaging
time used for all other waveform reads in SNAPIN. This can be useful
because the surface voltage is quite noisy, and may require a longer
averaging time than most other diagnostic inputs (be careful to avoid
including the beam turn-on transient and periods when the VC
current is ramped. In large plasmas [ cm] which might
deposit heat on the outer RF limiter, the VC current is
often ramped up a few hundred msec before the start of auxiliary
heating).