It sounds like you heard right. There's a reason for using pulsed
fields. You want the field as big as possible in order to flip the
magnetic domains into lining up with the applied field. For small
applied fields, the domains stay stuck. However, running lots of
current through your coil to make a big field will make it heat up and
ultimately melt or start a fire. So you run a big current for a short
time. The domains that don't respond to a pulse of a few milliseconds
mostly wouldn't have responded on longer times anyway. In other words,
it's more effective to have a big field a small fraction of the time
than a small field all the time.
You might experiment with having the steel at different
temperatures. The hotter it gets, the easier it is to get the domains
to line up during the pulse, at least until they get so hot that they
start to lose their magnetism altogether. However, the hotter it gets
the easier it is for the domains to relax back to the un-aligned state
after the pulse ends. There may be a best temperature a bit warmer than
Unfortunately, I don't have a circuit diagram of an automated pulse
circuit available. Maybe you could just make pulses with a manual
switch, say by discharging a capacitor through the coil.
I just tried magnetizing a paperclip by very brief contact with a
permanent magnet. It worked ok, and didn't seem to fade over the
weekend. So maybe you won't need many pulses.
(republished on 07/13/06)