Journal of the Society of Motion Picture and Television Engineers (1950-1954)

cur at about 0.4-mil wavelength or wavelengths across one side of the p and 40.5 wavelengths across the her side of the gap. This represents angular error of 6 min and a linear ror of 1.2% for the gap of the experiental head. Since this head is not itable for measurement at ^-mil wavength, it is evident that a degree of curacy is required which may not be actical to obtain. The parallelism problem can be reced in long-gap heads in various ways. T one thing, it might be advisable to ve a series of heads of successively orter gaps, placing the maxima farther art at the high end but improving the curacy. For another thing, the longp head could be made with a narrow lack width so that for a given angular ror in gap edges the linear error would ; reduced. Neither method has been vestigated prior to this paper. A comparison of the results of the two .ethods of measuring surface induction shown in Fig. 13. The upper set of irves shows the 15-in./sec surface inaction, and the lower set shows the 5-in./sec surface induction. The solid irves are the short-gap method and the Dt-dash curves are the long-gap method. The most striking difference between ie curves is a difference of slope of about 5 db/octave. A 2-db/octave differice was found by the BBC and the Dani participants in the CGIR. Various cplanations have been offered for the ifference in slope, and the matter is ill under investigation.3 It is apparent lat there is a wavelength dependent ariable that does not appear in the ssumed equation. 3 Since presentation of this paper, the illowing publication has appeared which ay contain the explanation of the difxence in slope between the two methods : K. Westmijze, Philips Research Laboitories, Eindhoven, Netherlands, "Gapngth formula in magnetic recording," custica, 2, No. 6: 292, 1952. The short-dashed curve shows a correction of 1.5 db/octave in the longgap curves. With this correction, the curves fit closely to about 1.5-mil wavelength with the short-gap curves. After this, the long-gap curves drop below the short-gap curves. It may be stated that measurements reported by the BBC and Denmark were only carried out to 1.5-mil wavelength. Up to this point they found the two methods to agree after applying the slope correction. In conclusion, it is felt that it is desirable to be able to measure surface induction. With the information gained, it is easier to evaluate recording and playback losses of a system as well as to evaluate tapes. Furthermore, it is desirable to be able to make the measurements in the field and be assured of substantial agreement of these measurements with others made on other machines. Methods are being devised for making the measurements. Of the two methods described in this discussion, the shortgap method is easier to perform and does not require a special head construction. Its accuracy is probably good if wavelength losses are held within 5 db. The long-gap method may be promising in the laboratory, but its accuracy must be improved at short wavelengths. Also we need to determine the nature of the 1.5 db/octave correction. Until this is explained fully, the validity of the long-gap method is questionable. Other methods may be devised, and with them we may have further checks on our results. The writer acknowledges the guidance and suggestions of W. E. Stewart, Chairman of the Subcommittee on Magnetic Tape Recording of the NARTB, and W. H. Erikson of Advanced Development Engineering, RCA Victor Div., as well as assistance in the project through design of special heads provided by L. W. Ferber of Commercial Sound Engineering, RCA Victor Div. J. D. Bick: Measuring Surface Induction 525