Quantum Design Oxford is reconfiguring superconducting magnet production at Tubney Woods, adding reaction furnaces, winding capacity, automated testing and separate manufacturing streams as it works towards a further 15% increase in output. The programme is already well beyond its halfway point and follows additional investment after Quantum Design acquired the Oxford operation in January 2026, with further winding, potting and test equipment planned over the next 18 months.
The production problem is unusual because the factory has to accommodate repeat magnets alongside high-value systems built in much smaller numbers. Quantum Design Oxford supplies dry Cryofree magnets that routinely reach 14 Tesla and liquid-helium-cooled products reaching 20 Tesla, so a standard repeat product and an extreme-field build can require very different amounts of specialist equipment, engineering input and time even when they share parts of the same manufacturing route.
Running those workloads through one sequence can make a one-off design disproportionately disruptive. The company is therefore working towards a factory-within-a-factory structure, separating repeat-volume products, lower-volume magnets that can be manufactured to forecast and specialist builds requiring their own end-to-end route. Common products are already being made ahead of confirmed orders where demand is predictable enough, allowing completed or partly completed units to reduce customer lead time.
Reaction furnaces are one area where that separation is becoming physical. Three smaller ovens have been installed, replacing a previous two-stage air-burn and reaction sequence with a single automated programme that controls gas, temperature and the reaction environment. More frequent smaller batches can be processed without occupying the large ovens, leaving those larger assets available for the highest-field magnets that genuinely require them.
A dedicated niobium-tin termination room has also been completed, while the winding area has been reorganised to create space for additional machinery. High-field superconducting magnets depend on tightly controlled coil manufacture and subsequent processing, so increasing capacity at one stage without matching winding, termination and testing would simply move the constraint further down the line. Quantum Design Oxford’s programme is consequently spread across several parts of the manufacturing sequence rather than concentrated on one headline machine.
People are being added alongside equipment because much of the work remains specialist. More than 20 manufacturing technicians, manufacturing engineers and technical engineers support the magnet line, with recruitment continuing and cross-training intended to broaden the range of assemblies individual employees can handle. That gives the operation more flexibility when production mix changes, reducing dependence on a narrow group of specialists without assuming that complex magnet manufacture can be automated end to end.
Testing is one area where automation can remove elapsed time. Engineers have developed scripts that allow measurements to continue once a magnet has reached the required cold condition, enabling test sequences to run beyond normal working hours while upgraded instruments and data loggers collect a fuller record of performance. The test requirement itself remains unchanged — the magnet still has to demonstrate the specified behaviour — but less time is lost waiting for an operator to initiate each step manually.
Quantum Design Oxford points to a recently completed 20 Tesla magnet as an early indication of what the revised process can achieve. The system operates at 4.2 Kelvin, weighs around 650kg, stands approximately 2.4 metres high and has a diameter of about 64cm; after passing qualification and system testing first time, it was delivered seven months ahead of a quoted two-year lead time.
One successful build does not establish a new lead-time standard, particularly when high-field magnets are produced in low numbers and individual programmes can differ considerably. The more meaningful evidence will be whether the same manufacturing structure improves delivery across a broader mix of repeat and specialist products without increasing rework, because an output target is useful only when qualification performance is maintained.
The remaining 18 months of investment will include winding-machine upgrades, potting equipment and a new magnet test laboratory, completing more of the separation between production streams. If the line can sustain its planned 15% output increase while preventing large one-off systems from repeatedly interrupting routine work, the reconfiguration will have addressed the difficult part of specialist manufacturing growth: adding capacity without allowing product complexity to consume it as quickly as it is created.
