Most floppy
drives have a tachogenerator on the motor anyway. On the old
full-height belt driven units, the spindle motor, a permanent magnet DC
motor -- has 4 wires. 2 go to the motor, the other 2 to an AC-output
tachogenerator inside the motor. I can look up the colour coding, IIRC
most mangufacturers used the same motor.
The problem being getting enough resolution to figure out how much the speed
varies over a single revolution. A single pulse per revolution just isn't
enough to say that with any degree of accuracy. It's enough to say that
It's a lot more than 1 pulse per revolution (In other words, I know of no
drive where the spindle speed is cotnrolled by the index sensor, which
normally does give one pulse per revolution).
However, I don;t know the sort of frequencies involved, the best way to
find out would be to make some measurements I guess.
revolution #1 took 0.2% too long and revolution #2 was
0.34% too fast, but not
enough to say where in the cycle that error was.
Incidentally, I do have that 'Microtest' unit I mentioned in an earlier
message. For those who don't know about it, it's a way of aligning floppy
drives without a 'scope. You link up the drive as drive B: on a PC (it
shows the age of the unit that the spec of the PC is '256K or RAM and 1
serial port'). There's a little box that contains a microcontroller (8035
IIRC) and an ADC, and which links to the serial port of the PC. You then
run the supplied program under MS-DOS on that PC and select the drive
yupe you ant to align from a menu.
Yo uthen get a picture of the drive PCB shown on the PC monitor using
the IBM line-drawing characters (so it'll even work on an MDA display).
It shows you where to connect probe clips that come from the ADC box --
the test points, IIRC, are ground, the differential outputs of the read
amplifier, the index pulse and the track 0 sensor.
Yo uthen put an alignment disk -- a standard CE disk, not a special one
-- in the drive and select various options on the PC. It'll measrue the
spindle speed by timing the index pulse. It'll move to the CE track and
display the CE pattern (again using block graphics) so you can do a head
alignment. And so on.
Yes, I do use it, running on an old Amstrad PPC640 laptop that I've
modified to have a DC37 connector in place of drive B:.
Anyway, the point of me mentioning it is that if you're working on an
older drive (3.5", 5.25", 8"), I would be happy to see if this system
knows about it, and if so, what the testpoints used are (and yes, I do
know what the various clips are supposed to be connected to).
-tony