… or what used to be called rocket science.

Sometimes, we also do stuff that is not so fancy, but still a lot of fun. See these projects:

Yoctoscope
Gyroscope
Lectures on Identity

Project: Yoctoscope

We contributed a stage prop for photographer and physicist Andrzej Dragan‘s first short film Hierarchy Lost. It also features a small Tesla coil that we made, briefly seen at 2:12 min. Our working title Yoctoscope was meant to signify the extreme magnification power of the device. In the end, it became totally EhVIL, the Event horizon Virtual Imaging Lens.

We made the device from discarded ultra-high vacuum parts, fancy, alien spider feet that we milled in our postgraduate workshop, lots of useless electronic bits, hard drive parts, hi-fi cable (most expensive), plastic tubes, vintage electronic valves and other bits and bobs. The device actually has a real function. A stepper motor opens and closes the sample chamber. True PhD skills were at play here. And did you notice the shiny CPU-cooler? But hey, (spoiler alert) even Darth Vader’s mask was made with a hard drive head arm. Yes! See Episode III, when Anakin Skywalker becomes Darth Vader. But we have to warn you. You will never be able to unsee it.

Project: Gyroscope

One of our ongoing students’ projects is to build a gimballed, magnetic gyroscope. It is a nice mechanical model for nuclear magnetic resonance, which has naturally many similarities with quantum (pseudo-)spin systems. We can observe magnetic resonance, Rabi cycles, decoherence, and can hopefully run Ramsey sequences to demonstrate the principle of an atomic clock.

Mechanical basics

The angular momentum vector (usually pointing along the axis of rotation) of a spinning top will maintain its direction, if no external torque is applied. This can be seen on a space station, and it forms the basis of spin-stabilized magnetic levitation, where a spinning magnet can hover above a repelling magnet because it doesn’t flip over. Unfortunately, we can neither switch off gravity nor friction and we don’t want to be restricted to a particular direction. So we need to mount our gyroscope in a set of gimbals and continuously drive its rotation with a motor. But there are two issues: gimbal lock and the torque that acts back on the motor. This torque could spin the gimbals instead of the gyroscope. Luckily, there is a solution that tackles both: the feedback-controlled extra gimbal for Christmas.

Our magnetic gyroscope is made from an old brushless hard drive motor with a neodymium ring magnet mounted on its one-inch diameter shaft. The magnet provides sufficient moment of inertia as well as a strong magnetic moment. The motor/magnet assembly is mounted at the centre of two freely rotating gimbals with orthogonal axes. Together with the axis of the motor shaft (that takes the role of an inner gimbal), this allows the magnet to adopt any orientation in 3-dimensional space. (That’s not the same as “performing any rotation”, see gimbal lock.) It can point in any direction and rotate about its axis, but it may get stuck, in a way.

Gimbal lock occurs when the middle gimbal rotates such that the motor axis (or inner gimbal axis) coincides with the axis of the outer gimbal. In that case a degree of freedom is lost and the gyroscope cannot reorient anymore in both directions. This situation can occur even for arbitrarily many nested gimbals. But it can be avoided by adding a redundant, actively controlled gimbal. Whenever the middle (passive) gimbal starts to rotate about its axis, the active gimbal is oriented such that the outer (passive) gimbal takes the necessary rotation instead. As a result, the motor axis always remains nearly orthogonal to both axes of the passive gimbals. Whichever way the gyroscope would like to reorient, at least one of the two passive gimbals will allow for that motion. The additional benefit is that the torque from the motor cannot spin the passive gimbals, because it never has a component parallel to those axes.

Electrical basics

The hard drive motor is run with a little brushless motor driver board that is mounted directly under the motor. We power it via the metal gimbal bearings, without the need for any further wiring. (You may have guessed, the metal frames all consist of two isolated parts, which is why we use plastic screws everywhere.)

In order to provide the required feedback to our active gimbal, we need to measure the rotation of the middle gimbal about its axis. Since any friction is an issue, we chose a contactless method, namely magnetic induction. We wound radio-frequency (RF) driven, paired saddle coils on the gimbal axis as well as on an outer sheath, fixed to the bearing. The saddle coils are oriented orthogonally, such that no cross-induction occurs at the preferred, relative angle. As soon as the middle gimbal rotates, amplitude and phase of the induction can be used to determine the angle, which we do with a simple, home-built lock-in amplifier. The RF-signal is also passed via the bearings of the outer gimbal to the main frame, this time angle independent, with axially oriented transformer coils. At the servo motor driven, active gimbal axis, we do not care about friction. There, we simply use slip rings to make galvanic contacts.

When the gimbals rotate, the electrical resistance of our ball bearings changes slightly and alters the power that is delivered to the motor. We can overcome this by monitoring the motor speed and adjusting the drive voltage accordingly. Interestingly, the back-EMF from the motor propagates back through the gimbals and their bearings and can be picked up at the power supply. This gives us a fairly easy way to determine the current motor speed.

The whole gyroscope is controlled by an Arduino board, connected to a computer. A simple program evaluates the lock-in signal and feeds it back to a servo-motor that controls the angle of our active gimbal.

What we would like to do next is to include optical angle encoders and be able to record the gyroscope’s orientation and thus its behaviour. Then the real fun will start. Wanna join?