An earlier tether experiment ended prematurely when problems arose with the deploying mechanism, but the one on February 25, 1996, began as planned, unrolling mile after mile of tether while the observed dynamo current grew at the predicted rate. The deployment was almost complete when the unexpected happened: the tether suddenly broke and its end whipped away into space in great wavy wiggles. The satellite payload at the far end of the tether remained linked by radio and was tracked for a while, but the tether experiment itself was over.

It took a considerable amount of detective work to figure out what had happened. Back on Earth the frayed end of the tether aboard the space shuttle was examined, and pieces of the cable were tested in a vacuum chamber. The nature of the break suggested it was not caused by excessive tension, but rather that an electric current had melted the tether.

The electric conductor of the tether was a copper braid wound around a nylon string. It was encased in teflon-like insulation, with an outer cover of kevlar, a tough plastic also used in bullet-proof vests, all this inside a nylon sheath. The culprit turned out to be the innermost core, made of a porous material which, during its manufacture, trapped many bubbles of air, at atmospheric pressure.

Later vacuum-chamber experiments suggested that the unwinding of the reel uncovered pinholes in the insulation. That in itself would not have caused a major problem, because the ionosphere around the tether, under normal circumstance, was too rarefied to divert much of the current. However, the air trapped in the insulation changed that. As it bubbled out of the pinholes, the high voltage ("electric pressure") of the nearby tether, about 3500 volts, converted it into a plasma (in a way similar to the ignition of a fluorescent tube), a relatively dense one and therefore a much better conductor of electricity.

The instruments aboard the tether satelite showed that this plasma diverted through the pinhole about 1 ampere, a current comparable to that of a 100-watt bulb (but at 3500 volts!), to the metal of the shuttle and from there to the ionospheric return circuit. That current was enough to melt the cable.

As the broken end whipped away from the shuttle, the plasma established electric contact with the ionosphere directly. The satellite on the distant end monitored the current: after about half a minute it stopped, then it reignited and flowed again for about another half minute, stopping for good when (presumably) all the trapped air was gone.

Because of the unexpected break, the tether experiment at the time was widely viewed by the press as an expensive failure. True, the planned operation at full deployment, for several hours, could not take place, nor could the tether and its satellite be retrieved, which was to have demonstrated the feasibility of deployable tethers.

However, many of the scientific experiments had already begun during deployment and yielded good data. And the break itself, though unfortunate, added an unscheduled experiment to the mission, one which highlighted the risks and complexities of operating scientific equipment in space.