Abstract
Most accelerator magnets contain iron or other ferromagnetic materials to increase the magnetic field of a coil. Unfortunately these materials have a strong dependence on their history (hysteresis). To obtain a desired magnetic field, a particular current must be supplied with respect to that history. Usually a history map is chosen in such a way that one of the main branches, up or down, of the hysteresis curve is selected. The choices are arbitrary and not natural. The disadvantages of these schemes are, for instance, long standardizing times going up and down the hysteresis, different slopes for increasing or decreasing the field and unstable, but quite reproducible values along the hysteresis loop. These problems can be overcome by choosing the curve showing the physical dependence of a magnet upon the current. This curve, in the middle of the hysteresis, shows the following advantages: reproducibility by cycling (up and down, going closer) to a certain current; stability to shock, small current changes, heat, even a temperature rise over the Curie temperature and back is possible; the same behavior (slope) going up or down in current for a small range of adjustments, therefore easier to correct by hand or computer (feedback); faster to adjust, especially for small changes. Different theoretical ideas are discussed and some experimental tests are described