Early drawing stages may be performed by draw-benches with the filled tube remaining straight, but if a finished tape hundreds of metres long is to be obtained it is inevitable that at least the later drawing stages will be performed using a drawing machine in which the drawn intermediate is taken up by winding it in some kind of roll.

It is our experience, whether the fully drawn intermediate is annealed before flattening or not, that the flatted tape so produced is far from perfectly straight. If it is not annealed immediately before flattening, it acquires stresses from the coiled form it had on the roll and these result in it wandering across the width of the flattening rolls producing edgewise bending that survives all subsequent rolling stages, while if it is annealed at that stage it is so soft that handling produces local deformations and defects that also survive subsequent rolling.

The present invention permits the production of superconducting tapes with much improved edgewise straightness, facilitating assembly with additional tapes and other components.

The invention comprises filling a tube with particulate superconducting material (as defined) to form an intermediate, drawing the intermediate through a plurality of

dies to reduce its cross-section, annealing the reduced intermediate, and rolling it though a series of roll passes to flatten and further reduce it and is characterised in that after the intermediate has been annealed it is drawn through a further die by tension applied to it by a first, flattening roll pass in which at least one of the rolls is driven.

In this way, imperfections arising in the handling of the intermediate in the annealed state may be substantially removed by the said further die and it reaches the flattening roll pass in a straight condition under tension and free from deleterious stresses. If required, it can be reeled after flattening (before and/or after further rolling stages), since the plane of the tape is now clearly defined and it can be confidently rolled flatwise without substantial edgewise bending.

We prefer, except in special circumstances, that all the drawing dies, including the further die, are round, but the use of an oval or otherwise flattened die for the further die is not absolutely excluded.

If the number of drawing stages makes it desirable, one or more additional annealing step may be interposed between drawing stages. Annealing may be a batch process, using a suitably heat-resisting reel if necessary, but we prefer that at least the annealing step preceding drawing through the said further die is an in-line step using a tubular oven or other appropriate heating arrangement. For silver-clad ceramic superconductor intermediates, annealing can be effected by heating to about 500°C for a few minutes.

The invention is applicable to the manufacture of powder-in-tube superconductors of any composition (even, if desired, in which the superconducting particles are of Nb3Sn or other superconducting metal alloy powder), but our present preference is for use of ceramic superconductors of the BiSCCO family, which approximate more or less closely to the formula (Bi, Pb) 2Sr2Ca2Cu3Ox, where x is a variable quantity in

the range 9-11 (depending on the heat-treatment history of the material, and not usually determined).

The filled tube intermediate may be a simple tube initially of round cross-section with a single bore and uniform cross-section, or it may be a composite tube with multiple filled bores, most easily produced by assembling several simple filled tubes together and compacting them.

Example A number of pure silver tubes having internal and external diameters of 8 and 10 mm respectively were filled with superconducting ceramic powder of nominal composition (Bi18PbO. 33) Srl. s7Ca2Cu30x and mean particle size about 2Tm.

These were compacted and drawn to a diameter of 2.82 mm using a total of 24 dies each effecting a reduction in area of about 10%, with three anneals (all anneals were effected by passage through a tube furnace at 500°-C, with a dwell time of a few minutes). Initial drawing was on a linear draw bench, but after reaching a diameter of 5.31 mm a first bull-block drawing machine with a block diameter of 600 mm was used and after reaching 5.31 a second bull-block machine with a drum diameter of 300 mm. They were then drawn through a hexagonal die to an across-flats dimension of 2.52 mm and then sets of seven were assembled in an untwisted hexagonal group and compacted together and further drawn through 35 successive dies, again each effecting a reduction in area of about 10%, to a diameter of 1.58mm. The draw bench and each of the two bull block machines were used with the same change-over sizes as before, with anneals at 5.31 and 2.98 mm and at the finished size of 1.58 mm.

The annealed intermediate was rewound (by hand) onto a plastics reel 300mm in diameter and 80 mm wide, in a traversed multilayer winding. The exposed end was pointed and pulled by hand through a further circular die with a diameter of 1.50mm until it could be inserted into the nip of a pair of rolls driven at a peripheral speed of 16.7mm/s. Each roll

was 110 mm in diameter and the pair were spaced to flatten the intermediate to a thickness of 1.21mm (nominal reduction in this roll pass 19%). Subsequently, in a sequence of twelve rolling steps each effecting a reduction of about 10%, the tape thickness was reduced to 0.33mm. The tape was then 3mm wide, and a 40m length was so made.

At this stage (still requiring heat-treatment and some further rolling to produce a finished tape), the a 1m sample of the tape was laid out on a flat surface and its deviation from a straight line measured an found to be lmm on the whole length of the sample and 0.5mm in any O. lm length. For a conventionally prepared tape at this stage of production, a deviation of around 10 mm would be expected.