Peeling modes are observed at the plasma edge in the Pegasus Toroidal Experiment under conditions of high edge current density (J(edge) similar to 0.1 MA m(-2)) and low magnetic field (B similar to 0.1 T) present at near-unity aspect ratio. Their macroscopic properties are measured using external Mirnov coil arrays, Langmuir probes and high-speed visible imaging. The modest edge parameters and short pulse lengths of Pegasus discharges permit direct measurement of the internal magnetic field structure with an insertable array of Hall-effect sensors, providing the current profile and its temporal evolution. Peeling modes generate coherent, edge-localized electromagnetic activity with low toroidal mode numbers n <= 3 and high poloidal mode numbers, in agreement with theoretical expectations of a low-n external kink structure. Coherent MHD fluctuation amplitudes are found to be strongly dependent on the experimentally measured J(edge)/B peeling instability drive, consistent with theory. Peeling modes nonlinearly generate ELM-like, field-aligned filamentary structures that detach from the edge and propagate radially outward. The KFIT equilibrium code is extended with an Akima spline profile parameterization and an improved model for induced toroidal wall current estimation to obtain a reconstruction during peeling activity with its current profile constrained by internal Hall measurements. It is used to test the analytic peeling stability criterion and numerically evaluate ideal MHD stability. Both approaches predict instability, in agreement with experiment, with the latter identifying an unstable external kink.

Peeling modes, an instability mechanism underlying deleterious edge localized mode (ELM) activity in fusion-grade plasmas, are observed at the edge of limited plasmas in a low aspect ratio tokamak under conditions of high edge current density (J(edge) ∼ 0.1  MA/m2) and low magnetic field (B ∼ 0.1  T). They generate edge-localized, electromagnetic activity with low toroidal mode numbers n≤3 and amplitudes that scale strongly with measured J(edge)/B instability drive, consistent with theory. ELM-like field-aligned, current-carrying filaments form from an initial current-hole J(edge) perturbation that detach and propagate outward.

Localized current injection near the outboard midplane is used to form 0.1MA plasma discharges with no induction supplied from a central solenoid in the ultra-low aspect ratio Pegasus Toroidal Experiment. The discharges are initiated by driving open-field-line currents that perturb the vacuum magnetic field such that the magnetic topology transitions to a tokamak-like configuration. The plasma is subsequently driven via helicity injection from the edge current sources and poloidal field induction. Intermittent n = 1 MHD activity is observed during periods of strong edge current drive and each event leads to a rapid inward expansion of the plasma volume and a drop in the plasma inductance. The plasmas are sufficiently turbulent such that the equilibrium approaches the lowest energy state described by Taylor relaxation theory. In agreement with that theory, the maximum I-p scales with (ITFIinj/w)(1/2), where I-TF is the toroidal field rod current, I-inj is the injected edge current and w is the radial width of the average poloidal magnetic flux in the driven open flux region.

Coaxial helicity injection in the HIT-II [A. J. Redd et al., Phys. Plasmas 9, 2006 (2002)] spherical torus is modeled with time-dependent resistive magnetohydrodynamics computations run to steady state for conditions without strong relaxation. Laboratory and computed results on injector current and plasma current agree reasonably well as toroidal magnetic field and injector flux are scaled. The scalings are consistent with a dimensional estimate from the Grad-Shafranov equation that provides a new perspective on a previously published model based on a current-sheet equilibrium and the magnetic pressure required for the “bubble-burst” criterion. Numerical solutions of the Grad-Shafranov equation with an assumed current profile also indicate large qualitative changes as the predicted criterion is crossed.