9 Mar
2004
9 Mar
'04
2:23 a.m.
On Tue, 9 Mar 2004 Gooijen H <GOOI(a)oce.nl> wrote:
I am not sure that removing the brake will help.
More likely not. In fact, the break is there for a reason. To stop the
drive when spinning down. Otherwise, it will spin for a long time. I also
suspect that the break is engaged until you try to spin the disk up, so
that the disk don't swing back and forth just because you move the drive
around.
I once heard the following rumour:
the brake in combination with the 'heavy' motor
makes sure that the disk platters always rotate
in the same direction, never for a short instance
in the other direction (vibration, power flutter,
whatever reason). Rotation in the opposite direction,
for any short moment, will cause the air on which the
r/w head fly to disappear and result in an immediate
head crash ...
Are you suggesting that a platter, weighting several kilograms, spinning
at something like 3600 rpm, suddenly can change spin direction at a
millisecond notice?
As I said, it is a rumour I once heard a few years
ago.
Has anybody heard of this too?
That's one of the more outrageous rumours I've ever heard. :-)
The amount of energy required to do that trick would blow every fuse in
your house, and then some. We're talking about living mass here. Physics,
you know... ;-)
However, vibration can cause a head crash...
Johnny
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt(a)update.uu.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
9 Mar
9 Mar
2:50 a.m.
On Tue, 2004-03-09 at 13:23, Johnny Billquist wrote:
Maybe the motor's prone to kick back in the other direction for a
fraction of a turn when it comes to rest, and that would be enough to
trash the heads...
I am quite surprised that they bothered with a brake without a good
reason - these drives weren't exactly intended to shifted around a lot,
so who cares if it took 5 minutes to spin down naturally? Mind you,
different era I suppose - the same thing could be argued about a serial
diagnostic interface to the drive, but I'm glad it's there :-)
cheers
Jules
I also
suspect that the break is engaged until you try to spin the disk up,
I believe you can hear the solenoid brake make a loud clunk noise as it
disengages just prior to spin-up.
Aside: I noticed on both of our units that have spin-up problems, that
the pulley on the bottom of the HDA was worn quite well. Although belt
tension is fine on both units and the belt appears to be gripping the
pulley very well, I did wonder if under start-up conditions it was
slipping.
I once heard
the following rumour:
the brake in combination with the 'heavy' motor
makes sure that the disk platters always rotate
in the same direction, never for a short instance
in the other direction (vibration, power flutter,
whatever reason). Rotation in the opposite direction,
for any short moment, will cause the air on which the
r/w head fly to disappear and result in an immediate
head crash ...
Are you suggesting that a platter, weighting several kilograms, spinning
at something like 3600 rpm, suddenly can change spin direction at a
millisecond notice? 
4:06 a.m.
Jules Richardson wrote:
I believe you can hear the solenoid brake make a loud clunk noise as it
disengages just prior to spin-up.
humm. interesting. And makes sense that it would be "engaged" until
power comes to it. I hear no clunk and it's definitely not disengaging.
another clue :-)
I need schematics, but have found none...
-brad
6:15 a.m.
I am quite surprised that they bothered with a brake
without
a good reason - these drives weren't exactly intended to
shifted around a lot, so who cares if it took 5 minutes to
spin down naturally? Mind you, different era I suppose - the
same thing could be argued about a serial diagnostic
interface to the drive, but I'm glad it's there :-)
I don't have anything to hand with prices for an RA8x drive,
but if we skip along one era to the days of the RA90/RA92,
the 1991 DECdirect catalogue shows ?14270 for an RA90 (and
an RA92!) and ?8028 for an RA70 (280MB in case you'd
forgotten). (Obviously you'd go for 4 RA70 drives and
get your 1.1GB discounted to ?28550 :-)).
Having spent a large fortune on your storage you would
almost certainly have it on contract. So the last thing
FS want to have to do is swap it out (??? in time and effort)
only to find back at base that that some lead was loose or
a simple adjustment or fix could have been made on-site.
So I expect that the RS232 connector was there initially
to help debug the thing while it was being designed and
then left in as a useful FS tool.
Antonio
--
---------------
Antonio Carlini arcarlini(a)iee.org
2:32 p.m.
> I once heard the following rumour:
> the brake in combination with the 'heavy' motor
> makes sure that the disk platters always rotate
> in the same direction, never for a short instance
> in the other direction (vibration, power flutter,
Depending on how the motor is wound, etc, it might be possible
*WHEN STARTING THE MOTOR FROM A DEAD STOP* that if initially spun
the wrong direction when power is applied, that it could spin
up reverse direction. The rotor is pushed around the stator by
discrete angular magnetic "bumps" and unless the magnetic field
is bent with shade poles or whatever the motor design uses,
it could start 'backwards'. Most motors provide for such a
thing one way or another. There are a lot of motor designs.
I have no specific information on the motor used in this drive.
Are you suggesting that a platter, weighting several
kilograms, spinning
at something like 3600 rpm, suddenly can change spin direction at a
millisecond notice?
This would be ummm interesting if you could make it happen... :-)
3:29 p.m.
Are you
suggesting that a platter, weighting several kilograms, spinning
at something like 3600 rpm, suddenly can change spin direction at a
millisecond notice?
This would be ummm interesting if you could make it happen... :-) 
3:38 p.m.
Quantum Mechanics tells us that the platter is
simultaneously
spinning both forward AND backward, and motionless, until YOU had to
go and open up the cover and make it assume a given state ;)
Well...until _something_ causes wavefunction collapse. It's not clear
when wavefunction collapse happens, though there are some good rules of
thumb; the closest I've seen to a theory including a mathematical
specification of when to collapse is Evan Walker's _The Physics Of
Consciousness_ (a book I highly recommend, by the by - his theory, for
all that it has weak spots (and what theory hasn't?) is also one of the
best cases I've seen of explaining more than it sets out to, including
such trivia as why so many creatures sleep).
/~\ The ASCII der Mouse
\ / Ribbon Campaign
X Against HTML mouse(a)rodents.montreal.qc.ca
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
13 Mar
13 Mar
11 a.m.
Breakthrough! (for no apparent reason)
We revisited one of our sick RA81 drives again today, having aquired
some proper schematics. Several of you had suggested the HDA spindle
transducer, and it does indeed look like that's the problem, at least on
one of our drives.
In glorious ASCII-vision, the connection to the spindle sensor on the
read/write board bolted to the top of the HDA is as follows:
Spindle Read / write
Sensor board
conn.
82ohm
red o----------XXX----------+------< +5V
|
black o-----<GND |
|
white o-----------------------+
|
black o-----<GND X
X 470 ohm
o (KEY) X
|
green o----->TEMP SENS. OUT |
|
yellow o-----------------------+------> TACH. OUT
Putting a scope on the tacho output from the sensor showed it at 5V
(tied to +5V via the 470 ohm resistor) with the drive at rest. When we
tried spinning the drive up, there was some fluctuation on the output,
but it dropped to only 4.5V max or so and obviously didn't look healthy.
Voltage at the pin where the red wire connects - i.e. supply to the
sensor LED(s) was around 1.3V and so seemed healthy enough.
We pulled the HDA and took the sensor off to try it on the bench in
isolation by replicating the circuit above and repeatedly breaking the
beam with a bit of card. Note that it showed no signs of being dirty.
This is where things got strange - at first it behaved just as in the
RA81, the output which was tied high by a 470ohm resistor dropping by
maybe 0.5V. But then it quickly began to pick up, the voltage ducking
further and further below 5V until it suddenly seemed to be operating
perfectly, the output swinging between +5V and close to 0.
We left it to cool down, thinking it might be a heat related thing. Came
back - nope, still working just fine.
We let it cool off for an hour or so again, then put the sensor back in
the drive and put everything back together. Guess what - one fully
operational RA81, spun up without any trouble!
I'm totally confused by that - the sensor showed no signs of dirt. It
hadn't worked in place in the drive, and nor did it work in isolation on
the bench at first. Yet it showed signs of *some* life when originally
in place within the drive, so there obviously wasn't a cable break.
I'm at a loss to explain that one, and it also means we're just waiting
for the drive to fail yet again, as we haven't actually fixed anything
(at least we'll know where to look this time!). Next test will be an
early morning start when the room that the drive's in is particularly
cold I suppose, but today seemed to prove that it would at least work at
normal room temperature and above....
cheers
Jules
1:47 p.m.
>>>> "Jules" == Jules Richardson
<julesrichardsonuk(a)yahoo.co.uk> writes:
Jules> ...I'm totally confused by that - the sensor showed no signs of
Jules> dirt. It hadn't worked in place in the drive, and nor did it
Jules> work in isolation on the bench at first. Yet it showed signs
Jules> of *some* life when originally in place within the drive, so
Jules> there obviously wasn't a cable break.
One guess would be a dirty connector. Unplugging and replugging can
rub off oxide that may have accumulated over the years.
paul
14 Mar
14 Mar
12:11 a.m.
On Sun, 2004-03-14 at 00:47, Paul Koning wrote:
Absolutely - I'd already unplugged and plugged the connector back into
the read/write board a few times just to try and rule that one out
though. I suppose we'll just see how it goes from very cold at next
startup.
Nothing else in the world seems to use those sensors - yes, floppy
drives, tape drives and printers tend to have them, but always in a
vertical orientation whilst the ones on the RA81's are horizontal. I
suspect we can make up a bracket for a vertical type to somehow mount it
horizontally if needs be though...
cheers
Jules
>>>> "Jules" == Jules Richardson
<julesrichardsonuk(a)yahoo.co.uk> writes:
Jules> ...I'm totally confused by that - the sensor showed no signs of
Jules> dirt. It hadn't worked in place in the drive, and nor did it
Jules> work in isolation on the bench at first. Yet it showed signs
Jules> of *some* life when originally in place within the drive, so
Jules> there obviously wasn't a cable break.
One guess would be a dirty connector. Unplugging and replugging can
rub off oxide that may have accumulated over the years. 
15 Mar
15 Mar
3:59 a.m.
Paul Koning wrote:
Or banging on the sensor dislodged the dust that it had collected. The card
you were using to test with may have scraped the lens clean.
>>>>"Jules" == Jules Richardson
<julesrichardsonuk(a)yahoo.co.uk> writes:
>>>>
>>>>
Jules> ...I'm totally confused by that - the sensor showed no signs of
Jules> dirt. It hadn't worked in place in the drive, and nor did it
Jules> work in isolation on the bench at first. Yet it showed signs
Jules> of *some* life when originally in place within the drive, so
Jules> there obviously wasn't a cable break.
One guess would be a dirty connector. Unplugging and replugging can
rub off oxide that may have accumulated over the years.
7435
days inactive
7441
days old
11 comments
9 participants
participants (9)
-
arcarlini＠iee.org -
bqt＠update.uu.se -
brad＠heeltoe.com -
julesrichardsonuk＠yahoo.co.uk -
jwest＠classiccmp.org -
kth＠srv.net -
mouse＠Rodents.Montreal.QC.CA -
pkoning＠equallogic.com -
tomj＠wps.com