The small value of the permeability of vacuum implies that an extremely high current density (∼ 1 MAcm-2) is needed to generate a magnetic field of one tesla. Since the maximum field strength of superconducting magnets is presently limited to about 20 tesla, one has to rely on resistive conductors to generate higher fields. The power consumption of these high field resistive magnets runs into megawatts and these are therefore only feasible in the few large-scale multi-million dollar infrastructures like the ones in Grenoble, Nijmegen, Tallahasse and Tsukuba.

An elegant way around this problem (and to also achieve even higher magnetic fields), is to generate very large power for only a very short time as a transient pulse. Consequently the total energy requirements can be limited to about a hundred kilojoules (about the energy contained in a cup of coffee) and one can still achieve fields of about 50 tesla. Depending on the size of the installation, anywhere between a few kilojoules to a few megajoules of energy is stored in a bank of capacitors which are discharged through a solenoid wound with a resistive conductor (e.g. copper). The constraint of heating is kept under control by the time duration for which the field is generated (10–1000 milliseconds). During this time the magnet, which is immersed in liquid nitrogen, adiabatically heats up to around room temperature. The experiment can be repeated once the magnet has cooled down again.

Schematic circuit of a pulsed magnet. It is essentially an LCR circuit. A bank of capacitors (600 mF in our case) stores energy which is rapidly discharged through a suitably designed solenoid with resistance R2 and inductance L. The thyristor (Ty) is used to close the circuit and the diode (D) prevents oscillation of current.

• Capacitor Bank—The capacitor bank (75 kJ) comprises of 60 electrolytic capacitors, 10 mF each with a peak voltage of 500V, connected in parallel. The capacitors are made by a Pune-based company (Alcon) and the bank was wired by the students here.

• High Power electronics and switching—The magnet is energized with a transient current pulse of about 20 ms duration and a peak current of about 25 kA. This requires a high current electronic switch (thyristor) that completes the circuit after receiving a trigger pulse and a crowbar (power diode) to prevent the LCR circuit from oscillating. Both these were purchased from the market in Kolkata.

• Magnet Coil— For pulsed magnets the fundamental constraint on the peak field is only the strength of the conductor wire which has to withstand the massive Lorentz force. For copper the limit turns out to be around 30 tesla. By a careful design that involves reinforcing the conducting wire with a high strength fibre composite and distributing the stress within the volume of a coil, the peak field can be increased by about a factor of two.Our coil has a 16 mm bore with 7 layers of winding, each appropriately reinforced to withstand the stress at 35 tesla. Since winding such a coil is an a very specialized task, Dr Tao Peng (of Wuhan University of Science and Technology, China) improved our design and wound the coil for us.

• Low temperature cryostat—Since we do not have access to liquid helium, we have developed a magnet cryostat around an old 4 K helium closed cycle refrigerator having a sapphire (thermal conductor and electrical insulator) cold finger. Currently, the temperature range accessible is about 6 K to 300 K.

• Electronics for synchronous data acquisition—The trigger pulse that fires the magnet also triggers a high speed data logger (20MHz, 16 bit, 4 channel). A low noise voltage preamplifier and ac or dc current source is used to excite the sample. We can do Hall effect and magnetoresistance measurements.

• Digital Signal Processing—Measurements in a pulse magnetic field, by their very nature, are noisy. Thousands of amperes of current suddenly flows through the magnet coil, giving the whole set up a large mechanical impulse. Moreover, the changing magnetic field superimposes a large inductive pick up on the sample signal. To get around this, following by now a routine practice, we have developed a high frequency digital lockin amplifier where the sample is excited at around 1MHz and the sample signal and the reference waveform during the pulse are stored. The lockin procedure (multiplication and low pass filter) is later digitally implemented using Matlab. Due to the short time of the pulse, commercial lockins are not suitable for transport measurements in pulsed fields. While the software part is implemented and working well, we are still trying to get around the large cable capacitance and inductance that interfere with the high frequency measurement.

• Magneto-photoluminescence spectroscopy (Under development)—We have recently acquired very sensitive electron-multiplying charged coupled device camera-based spectrograph that would allow us to record luminescence spectra synchronized with the pulse. The associated fibre-optics assembly and cryostat are not ready yet.

You may write to Bhavtosh Bansal [bhavtosh[]iiserkol.ac.in] for more information about the set-up and its capabilities and/or if you are interested in trying some magnetotransport measurements. If you have an interest in setting up such a facility in your lab, we would be delighted share our expertise and resources.