9 Dec
2003
9 Dec
'03
12:03 p.m.
Tony wrote:
Then there's the issue of writing out the last
cylinder before
power-down. If there's no seek from said cylinder, then how will the
interfacee know it has time to write it?
I don't know if the recent disk drives still do it the same way, but
at one time some 5.25-inch winchester drives used the spindle motor as
a generator to provide power to write any dirty buffers to the media
when power was lost. This won't work if much seeking is involved, so
they have a dedicated area of the media to store the buffers near the
head landing zone. On power-up, they read the buffer area, write any
dirty buffers back to the correct locations, and write a "clean" flag
to the buffer area.
Obviously this won't solve the power fail problem for the proposed
drive emulator, but it should be reasonable to have battery backup
for the buffer memory (two cylinders' worth).
Assuming the use of modern circuitry, everything in the disk emulator
other than the buffers for the actual ST-506 interface signals should
run on 3.3V or less. So another possibility is to detect 5V fail, and
write the dirty buffers to flash before the 3.3V drops too low. It
shouldn't take an unreasonably sized electrolytic capacitor on the
3.3V rail to provide enough time for this.