13 Jun
2010
13 Jun
'10
6:47 p.m.
Hi folks,
I'm just reviving a completely misaligned RK05 drive. I want to get it fully up and
running and
tested before I use an alignment pack. So I started debugging and got it to the point that
I can
play with the formatter.
Here comes the question: The formatter (RKLFMT under OS/8) threw MANY CRC errors when
reading back
the freshly formatted disk. Changing the head cables at the drive proved that the problem
is only
with one head i.e. the failures change "side" when you change the head plugs.
Now I ran the test several times. With random failing disk addresses. And they decreased
and
decreased... Now I've run two error-free formatter rounds.
Could it be possible that new data on new tracks (keep in mind that the drive is actually
misaligned
by means of mechanical alteration!) suffers from interference from old tracks? That's
the only
reason I can see for this kind of behavior.
On the other hand that would mean that I'd have to degauss the pack before I ever
could use it
reliable again.
Any ideas?
Kind regards,
Philipp :-)
--
http://www.hachti.de
14 Jun
14 Jun
2:55 p.m.
Hi folks,
I'm just reviving a completely misaligned RK05 drive. I want to get it fully up and
running and
tested before I use an alignment pack. So I started debugging and got it to the point
that I can
play with the formatter.
Here comes the question: The formatter (RKLFMT under OS/8) threw MANY CRC errors when
reading back
the freshly formatted disk. Changing the head cables at the drive proved that the problem
is only
with one head i.e. the failures change "side" when you change the head plugs.
Now I ran the test several times. With random failing disk addresses. And they decreased
and
decreased... Now I've run two error-free formatter rounds.
Could it be possible that new data on new tracks (keep in mind that the drive is actually
misaligned
by means of mechanical alteration!) suffers from interference from old tracks? That's
the only
reason I can see for this kind of behavior.
I think that's unlikely. There is a tunnel erase head which essentially
wipes out any old data that might be picked up by the read head.
Do you know if the heads are good? I've had at least one RK05 head that
tested OK electrically, looked clean, but which procuded no signal
On the other hand that would mean that I'd have to degauss the pack before I ever
could use it
reliable again.
Since there's no servo infromation or anything like that on an RK05 pack,
there is no reason why you can't degauss it.
Do yuo have any more packs? If so, try formatting another one. If that
passes with no CRC errors, it would appear to be a head problem which has
got better as the drive was used. If you get CRC errors at first which go
away if you format it a few times, it would apperar to be your suggestion
of old data causing problems.
-tony
6:01 p.m.
Salut Tony,
Of course, I *can* degauss the
pack. The question is if I *must* degauss a pack before using it with
new track positions.
I think that's unlikely. There is a tunnel erase
head which essentially
wipes out any old data that might be picked up by the read head.
You mean that the
erased trace is guaranteed to be broader than the area that can be sensed by the
read head?
Do you know if the heads are good? I've had at
least one RK05 head that
tested OK electrically, looked clean, but which procuded no signal
Oh, I cleaned
them. And they look white. And I actually CAN read and write with both heads. But
there are errors from time to time. And the error frequency seemed to drop with every time
I format
the pack... Perhaps it's something at the write side... No idea.
I'll grab some heads and throw them at the drive. Then I'll try again. Currently I
have to find a
new blower for the unit 0 RK05 of my test system. It shorted and burned yesterday while
playing with
the rotten drive.
On the other hand that would mean that I'd have to degauss the pack before I ever
could use it
reliable again.
Since there's no servo infromation or anything like that on an
RK05 pack,
there is no reason why you can't degauss it. Do yuo have any more packs?
Of course.
If you get CRC errors at first which go
away if you format it a few times, it would apperar to be your suggestion
of old data causing problems.
If I had a degausser I'd immediately try it out.
I already thougt about a program that writes random data to the disk. Again, again, and
again. That
could level out the area around the tracks.
Another possibility: My idea is complete bullshit :-)
Best wishes,
Philipp :-)
--
http://www.hachti.de
15 Jun
15 Jun
5:27 a.m.
On Tue, 15 Jun 2010, Philipp Hachtmann wrote:
Well, erasing takes place *after* writing. The R/W head writes the data
and the erase head then erases the borders of the just written data.
The error just sounds like the one I had in my RK05 drive. I could read
all my packs without any errors, but writing the lower side was very
unreliable. Tests showed that I could write (and read back, of course)
data from track 0 up to 127. Starting at track 128 I just had lots of read
errors. The maintenance manual revealed that the write current is reduced
starting with track 128, so that gave me a clue. Swapping the cables from
the upper and lower head moved the errors from the lower to the upper side
(logically, i.e. from the software point of view), so the fault was not in
the R/W amplifiers. It could only be a defective R/W head. I then changed
only the lower head, and the drive is working fine since then.
I suspect that the head had shorted windings or something like that. It
otherwise looked fine. It is possible that it had crashed sometimes in
the past before I got the machine.
Christian
I think
that's unlikely. There is a tunnel erase head which essentially
wipes out any old data that might be picked up by the read head.
You mean that the
erased trace is guaranteed to be broader than the area that
can be sensed by the read head? Do you know if the heads are good? I've had
at least one RK05 head that
tested OK electrically, looked clean, but which procuded no signal
Oh, I cleaned
them. And they look white. And I actually CAN read and write
with both heads. But there are errors from time to time. And the error
frequency seemed to drop with every time I format the pack... Perhaps it's
something at the write side... No idea.
I'll grab some heads and throw them at the drive. Then I'll try again.
Currently I have to find a new blower for the unit 0 RK05 of my test system.
It shorted and burned yesterday while playing with the rotten drive. 
1:36 p.m.
I think
that's unlikely. There is a tunnel erase head which essentially
wipes out any old data that might be picked up by the read head.
You mean that the
erased trace is guaranteed to be broader than the area that
can be sensed by the read head?
1:45 p.m.
Salut Tony,
Yes, I beleive so.
I think that's unlikely. There is a tunnel
erase head which essentially
wipes out any old data that might be picked up by the read head.
You mean that the
erased trace is guaranteed to be broader than the area that can be sensed by the
read head? I'll grab some heads and throw them at the drive.
Then I'll try again.
Currently I have to find a
new blower for the unit 0 RK05 of my test system. It shorted and
burned yesterday while playing with
the rotten drive.
They don;t normally short, what happens is the insulation on the windings
breaks down, and you get an unwanted earth. And then te magic smoke (and
flames) come out.
I will have to check the scheamtics, but since this is a 115V motor run
off the primary of the PSU transformer, is it wired so that it has to
stand 230V between one end of the windings and chassis ground? If so, it
might explain why they seem tofail all too often over there.
I am sure you know the blower comes apart. After removing it (4 Allen
screws hidden by the foam seal), you can remove the ring under that seal
(a few self-tapping screws), then the fan blades (Use a long hex key
between the blades to undo the setscrew), and then seprate the motor from
the fan housing (2 nuts). The motor itself comes apart by flicking off
the clims and removing the end plate and rotor. But it is almost
imposible to rewind it.
I mention this because when one ofmine failed, a friend gave me a faulty
blower with good winidings and bad bearings. And of course I made one
good one from the 2.
If you can find a suitable motro, you could probably fit it to the
housing and fit the fan blades. A motoe may be easier to find than an
RK05 blower.
Of course all the screws are UNC sizes, you need imperial (inch-size)
Allen keys and nutdrivers. But if you work on DEC stuff, you must have
thsoe anyway.
On the other hand that would mean that I'd have to degauss the pack before I ever
could use it
reliable again.
Since there's no servo infromation or anything like that on an
RK05 pack,
there is no reason why you can't degauss it. Do yuo have any more packs?
Of course.
If you get CRC errors at first which go
away if you format it a few times, it would apperar to be your suggestion
of old data causing problems.
If I had a degausser I'd immediately try it out.
14 Jun
14 Jun
4:58 p.m.
On 06/13/2010 03:47 PM, Philipp Hachtmann wrote:
I'm just reviving a completely misaligned RK05
drive. I want to get it
fully up and running and tested before I use an alignment pack. So I
started debugging and got it to the point that I can play with the
formatter.
Here comes the question: The formatter (RKLFMT under OS/8) threw MANY
CRC errors when reading back the freshly formatted disk.
If the drive can't
read back a pack it has just formatted, you've got
more problems than the positioner alignment. I'm not an expert on RK05
maintenance, though.
5:54 p.m.
On 14.06.2010 22:58, Eric Smith wrote:
If the drive can't read back a pack it has just
formatted, you've got
more problems than the positioner alignment. I'm not an expert on RK05
maintenance, though.
The alignment *should* be no problem. My question was if there
*could* be interference of old data
with differently aligned new data. That could happen if the read side is able to sense
more (in a
way of geometry) than the erase/write part treat when writing. I already like the drive.
It had many
more problems. It's one of the worst I've got so far :-) The first I saw of this
drive was a fallen
apart transducer and smoke from the PSU...
--
http://www.hachti.de
5128
days inactive
5130
days old
7 comments
5 participants
participants (5)
-
ard＠p850ug1.demon.co.uk -
cc＠informatik.uni-stuttgart.de -
cisin＠xenosoft.com -
eric＠brouhaha.com -
hachti＠hachti.de