First progress report on a multi-channel magnetic drum inner ...

A-2
Using this formula (i.So, assuming each pole to be a line pole)
let us see to vmat extent vje can approximate the shape of an experimentally
deterndned voltage patterno The pattern for a single-wire head spaced 2 m.ils
from the mgnetic medium is shown in Drawing C-5-2021 along with the corres-
ponding computed paUern. This pattern was computed as follows „ Measuring
the value of y to the center of the wire we have y = 2o5 mils and hence
e = k
t
2,$ _ 2o5
(^x)^ > 6„25 W>^)^ > 6. 25 J
The value of x for which e is a maximum, (minimum) on the experimental curve
is X = + 1,8 mils. Setting the derivative of the above expression to zero
at X = + 1.8 we obtain Z = 1.25 mils « The voltage pattern can then be
computed. Kote that the overall length of the dipole is 2.5 rrdls, or lass
than 1/4 of the extrapolated length of 10.4 milso This shows that if we
space dipoles according to the extrapolated width (.see Section 3ol) there
is a consi derable separation of the poles of adjacent dipoles and hence
little interference between them^ This is consistent with the fact that
our studies of dynan.ic recording showed that pulses could be safely spaced
according to the extrapolated length. It should be noted in this connec-
tion that a dynarrically recorded dipole is longer than th-. corresponding
statically recorded one because of the motion of the head during the
recording. However, the record nulses used in the investigation had a
duration of only 0.2 microsecond which corresponds to a motion of only
0.2 mils at a peripheral velocity of 1000 iriches per second.
It will be noted in Drawing G~5-2021 that the calculated pattern
is broader than the m.easured one. This may seem surprising since the cal-
culations were based on a line-dipole, whereas we would expect some spread-
ing (especially at the surface — ivhcre we liave ass'on.ed the dipole to be;,
" I