Abstract
Integrator drift is a problem strongly felt in different measurement fields, often detrimental even for short-term applications. An analytical method for modelling and feed-forward correcting drift in magnetic flux measurements was developed analytically and tested experimentally. A case study is reported on the proof of principle as a novel kind of quasi-DC field marker of the 5-ppm Nuclear Magnetic Resonance (NMR) transducer Metrolab PT2026, applied to the Extra Low ENergy Antiproton (ELENA) ring and the Proton Synchrotron Booster (PSB) at CERN. In some particle accelerators, such as in ELENA, the resulting feed-forward correction guarantees 1 T field stability over 120-s long magnetic cycle on a plateau of 50 mT, reducing by three orders of magnitude the field error caused by the integrator drift with respect to the state of the art.
Highlights
The problem of voltage integrator drift is strongly felt in different domains where real-time measurements are needed, from inertial to magnetic sensors
These measurements are needed in real time, for example to feed back magnetic field information to various machine control systems
Based on the mathematical modeling of the drift and of the results obtained experimentally in Extra Low ENergy Antiproton (ELENA) and the Proton Synchrotron Booster (PSB), combined with the results of the metrological characterization for the Nuclear Magnetic Resonance (NMR), we have attempted to predict the value of the uncertainty of the field at the end of the interval, through Equation (23), using the new feed-forward correction algorithm and the experimental results
Summary
The problem of voltage integrator drift is strongly felt in different domains where real-time measurements are needed, from inertial to magnetic sensors. Often the magnets are operated with fast magnetic cycles, so that measurements with a fixed induction coil are useful to capture dynamic effects In certain cases, these measurements are needed in real time, for example to feed back magnetic field information to various machine control systems. These measurements are needed in real time, for example to feed back magnetic field information to various machine control systems The requirements for these applications may be very demanding, up to 100 ppm or better in terms of relative accuracy, and hundreds of kHz of bandwidth, which makes the drift correction techniques described above hardly suitable. Drift correction is similar in both cases: the DC offset voltage is first estimated by averaging the voltage at the input of the integrator over a time interval when no excitation is applied to the magnet, and is subtracted from the input signal. The method proposed below is implemented in a new generation of B-train systems which is being developed and tested at CERN, in the context of a general consolidation project, with the goal to improve their performance and guarantee their long-term maintainability
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
