RL02 head gain adjustment: Limits?
cz at alembic.crystel.com
Wed Jan 8 19:42:08 CST 2020
So in the continuing quest of fixing my second RL02 I replaced the heads
and found the drive was slow and would generate about 80 block errors on
a pack where my reference RL02 drive would never generate more than 10.
After doing a positional radial check (fine) and a head alignment check
(head 0 and 1 were way off from each other and head 0 was on a crest of
the stepper motor (the part where you are between steps, highly
unstable). Adjusting that so head 0 and 1 were in a valley on the
stepper motor (inherently stable) the drive was much quicker but still
would generate about 80 errors. Odd.
So on a lark I swapped the head amplifier module from the good drive to
the suspect drive. Bingo, 10 errors. So the issue is in the head
amplifier, maybe the voltage was too low.
Next step was to check the read signal amplitude: Because I have a
reference drive I decided to check the signal voltage at track 0 on the
reference drive, then move the same pack to the problem drive and
verify. On the reference drive track 0 has a signal value of a smidge
below 1.5 volts on head 1. On my suspect drive the signal value was 1v.
Adjusted it up to 1.5 volts and errors went down to 9. Not bad. Went up
to 1.6 volts and errors went down to 6.
So a question: Is it a problem to adjust the read amplitude up, and if
so how far can you go. The manual says 2.25, is there a benefit of
running up that high? More important, does adjusting the read amplitude
also change the write current through the heads? That could result in a
pack that may be readable, but will be a mess when written to.
What would be nice would be to have a blue amplitude reference pack, but
I don't have one of those. Also I haven't tried any other packs in the
suspect drive, but maybe I should adjust the amplitude in my reference
drive as well.
(All readings done with a tektronix scope. I haven't checked the
calibration in awhile, but it should be reasonably in sync)
More information about the cctalk