298 IEEE TRANSACTIONS ON MAGNETICS, VOL. 46, NO. 2, FEBRUARY 2010 Measurement of Electrical Steels With Direct Field Determination Oleksandr Stupakov , Richard Wood , Yevgen Melikhov ,and David Jiles Institute of Physics of the AS CR, v.v.i., 18221 Prague, Czech Republic Wolfson Centre for Magnetics, Cardiff University, Cardiff CF24 3AA, U.K. The objective of the presented work was to introduce the principles of direct field control into magnetic measurements. Different grades of nonoriented and grain-oriented steels were tested in dc magnetization regime. A vertical array of three Hall sensors was used to control the sample field. With small gradients of the sample surface field our results showed comparable measurement errors with the single sheet tester. Measurement of the hysteresis loss and the coercive force was mostly independent of the field determination method. However, for accurate estimation of the magnetic induction parameters the sample field should be determined directly. No unique dependence was found between the measured dc and 50 Hz Epstein/single sheet tester parameters except the remanent induction of the nonoriented steels. Index Terms—Magnetic field measurement, magnetic hysteresis, magnetic variables measurement, silicon steel. I. INTRODUCTION T HE ultimate goal of our investigation is to find a simple method for fast localized repeatable testing of magnetic properties of the electrical steels, which can be applied on the production line for quality control. This problem is of interest to our industrial partner Cogent Power Ltd, which is a large-scale producer of electrical steels [1]. Magnetization of the moving steel sheet should be realized with a substantial air gap between the magnet and the sample. However, under this condition the magnetization current cannot provide a reliable base for evalu- ation of the sample field [2]. Therefore, the hysteresis measure- ments were performed with direct field control to check if the proposed approach can solve this problem [3]. The current work presents the first results of the project—reliability of the method was studied in the simplest dc magnetization regime. The con- cept of further development of a unique measurement system, capable to solve this practical task, is described in Section IV. II. EXPERIMENT Six different grades of nonoriented steels and six grades of grain-oriented steels, produced by Cogent Power Ltd, were used [1]. The nonoriented steels were tested parallel and perpendicu- larly to the rolling directions. Two strips of each steel of standard sizes mm mm mm were locally magnetized at three different positions by a single yoke with the inner pole distance of 40 mm (see Fig. 1). The 3 mm air gap between the yoke and the strips was chosen as a compromise between the de- viation range of the moving sheet and the distance, needed for its homogeneous magnetization. With the air gap increase the direct field method provides the same results, whereas the error of the current field method rises considerably [2], [4]. Manuscript received June 19, 2009; revised September 03, 2009; accepted September 14, 2009. Current version published January 20, 2010. Corre- sponding author: O. Stupakov (e-mail: stupak@fzu.cz). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TMAG.2009.2032861 Fig. 1. Scheme of measurement setup (shown sizes are in mm). The magnetization frequency was 0.5 Hz to stabilize the re- sponse at dc regime. The induction coil was wrapped around the strips. The measurements were done with the fixed maximum induction , 1.5 T for the nonoriented and , 1.7 T for the grain-oriented steels, respectively. The extra-sen- sitive Hall elements HW-108C from Asahi Kasei (1.6 mV/G) were integrated in the array to simultaneously measure the tan- gential surface fields at 1.5, 4, and 6.5 mm above the sample [5]. It was placed in the middle of the yoke-free side of the strips, where the surface field gradients are small. All measured signals were sampled at 150 kHz, averaged on-the-fly over 300 adjacent points and 10–25 consequent magnetization cycles. This proce- dure was necessary for smoothing of the raw data, especially of the Hall sensor outputs [6]. The measured magnetic induction was referred to the current field , evaluated to be pro- portional to the magnetization current, to the surface field , measured by the Hall sensor at 1.5 mm distance, and to the field , linearly extrapolated to the strip surface using three-point readings of the Hall sensor array [2]. Standard 50 Hz measure- ments [7], [8] were also done for the reference: Epstein data was provided by the producer, single sheet tester (SST) experiments 0018-9464/$26.00 © 2010 IEEE