Is this hard disk dying? :-(

[sk8terg1rl]

Hi everyone,

Does it look like my hard disk is failing here, or is there any chance
it could be the IDE controller?

The login screen is responsive but ssh into this PC doesn't work.
After typing in the password the system locks up, and Ctrl-Alt-Del
gives the message:
college login: /etc/initdscript: line 106: /sbin/shutdown: Input/
Output error

I've had to put both the HD & DVD drive onto the secondary IDE channel
because the primary one is dodgy on this motherboard. The motherboard
and hard disk have just been replaced and are fairly new too (under a
year) so I'm puzzled why this PC keeps failing.

Thanks for any advice
skate

Contents of /var/log/warn follows:

Jun 23 23:02:48 college kernel: hdd: dma_timer_expiry: dma status ==
0x61
Jun 23 23:02:58 college kernel: hdd: DMA timeout error
Jun 23 23:02:58 college kernel: hdd: dma timeout error: status=0x58
{ DriveReady SeekComplete DataRequest }
Jun 23 23:02:58 college kernel: ide: failed opcode was: unknown
Jun 23 23:02:58 college kernel: hdc: status error: status=0x50
{ DriveReady SeekComplete }
Jun 23 23:02:58 college kernel: ide: failed opcode was: unknown
Jun 23 23:02:58 college kernel: hdc: status error: status=0x58
{ DriveReady SeekComplete DataRequest }
Jun 23 23:02:58 college kernel: ide: failed opcode was: unknown
Jun 23 23:02:58 college kernel: hdc: drive not ready for command
Jun 23 23:02:59 college kernel: hdd: status timeout: status=0xd0
{ Busy }
Jun 23 23:02:59 college kernel: ide: failed opcode was: unknown
Jun 23 23:02:59 college kernel: hdd: no DRQ after issuing
MULTWRITE_EXT
Jun 23 23:04:30 college kernel: ide1: reset: success
Jun 23 23:04:30 college kernel: hdd: dma_timer_expiry: dma status ==
0x41
Jun 23 23:04:30 college kernel: hdd: DMA timeout error
Jun 23 23:04:30 college kernel: hdd: dma timeout error: status=0x58
{ DriveReady SeekComplete DataRequest }
Jun 23 23:04:30 college kernel: ide: failed opcode was: unknown
Jun 23 23:04:30 college kernel: hdd: status error: status=0x50
{ DriveReady SeekComplete }
Jun 23 23:04:30 college kernel: ide: failed opcode was: unknown
Jun 23 23:04:30 college kernel: hdd: no DRQ after issuing
MULTWRITE_EXT

Jun 23 23:04:30 college kernel: ide1: reset: success
Jun 23 23:04:30 college kernel: hdd: dma_timer_expiry: dma status ==
0x41
Jun 23 23:04:30 college kernel: hdd: DMA timeout error
Jun 23 23:04:30 college kernel: hdd: dma timeout error: status=0x58
{ DriveReady SeekComplete DataRequest }
Jun 23 23:04:30 college kernel: ide: failed opcode was: unknown
Jun 23 23:04:30 college kernel: hdd: status timeout: status=0xd0
{ Busy }
Jun 23 23:04:30 college kernel: ide: failed opcode was: unknown
Jun 23 23:04:30 college kernel: hdd: no DRQ after issuing
MULTWRITE_EXT
Jun 23 23:04:30 college kernel: ide1: reset: success
Jun 23 23:04:30 college kernel: hdd: dma_timer_expiry: dma status ==
0x41
Jun 23 23:04:30 college kernel: hdd: DMA timeout error
Jun 23 23:04:30 college kernel: hdd: dma timeout error: status=0x58
{ DriveReady SeekComplete DataRequest }
Jun 23 23:04:30 college kernel: ide: failed opcode was: unknown
Jun 23 23:04:30 college kernel: hdd: status timeout: status=0xd0
{ Busy }
Jun 23 23:04:30 college kernel: ide: failed opcode was: unknown
Jun 23 23:04:30 college kernel: hdd: no DRQ after issuing
MULTWRITE_EXT
Jun 23 23:04:30 college kernel: ide1: reset: success
Jun 27 01:30:02 college kernel: hdd: status error: status=0x7f
{ DriveReady DeviceFault SeekComplete DataRequest CorrectedError Index
Error }
Jun 27 01:30:02 college kernel: hdd: status error: error=0x7f
{ DriveStatusError UncorrectableError SectorIdNotFound
TrackZeroNotFound AddrMarkNotFound }, LBAsect=149570080178047,
high=8915071, low=8355711, sector=40984406
Jun 27 01:30:02 college kernel: ide: failed opcode was: unknown
Jun 27 01:30:02 college kernel: hdd: drive not ready for command
Jun 27 01:30:09 college kernel: ide1: reset: success
Jun 27 01:57:51 college kernel: hdd: status error: status=0x7f
{ DriveReady DeviceFault SeekComplete DataRequest CorrectedError Index
Error }
Jun 27 01:57:51 college kernel: hdd: status error: error=0x7f
{ DriveStatusError UncorrectableError SectorIdNotFound
TrackZeroNotFound AddrMarkNotFound }, LBAsect=149568184352639,
high=8914958, low=8355711, sector=40986358
Jun 27 01:57:51 college kernel: ide: failed opcode was: unknown
Jun 27 01:57:51 college kernel: hdd: drive not ready for command
Jun 27 01:58:13 college kernel: ide1: reset: success

 

[Roby]

Hi everyone,

Does it look like my hard disk is failing here, or is there any chance
it could be the IDE controller?

The login screen is responsive but ssh into this PC doesn't work.
After typing in the password the system locks up, and Ctrl-Alt-Del
gives the message:
college login: /etc/initdscript: line 106: /sbin/shutdown: Input/
Output error

I've had to put both the HD & DVD drive onto the secondary IDE channel
because the primary one is dodgy on this motherboard. The motherboard
and hard disk have just been replaced and are fairly new too (under a
year) so I'm puzzled why this PC keeps failing.

Thanks for any advice
skate

Contents of /var/log/warn follows:

Jun 23 23:02:48 college kernel: hdd: dma_timer_expiry: dma status ==
0x61
Jun 23 23:02:58 college kernel: hdd: DMA timeout error
Jun 23 23:02:58 college kernel: hdd: dma timeout error: status=0x58
{ DriveReady SeekComplete DataRequest }
Jun 23 23:02:58 college kernel: ide: failed opcode was: unknown
Jun 23 23:02:58 college kernel: hdc: status error: status=0x50
{ DriveReady SeekComplete }
Jun 23 23:02:58 college kernel: ide: failed opcode was: unknown
Jun 23 23:02:58 college kernel: hdc: status error: status=0x58
{ DriveReady SeekComplete DataRequest }
Jun 23 23:02:58 college kernel: ide: failed opcode was: unknown
Jun 23 23:02:58 college kernel: hdc: drive not ready for command
Jun 23 23:02:59 college kernel: hdd: status timeout: status=0xd0
{ Busy }
Jun 23 23:02:59 college kernel: ide: failed opcode was: unknown
Jun 23 23:02:59 college kernel: hdd: no DRQ after issuing
MULTWRITE_EXT
Jun 23 23:04:30 college kernel: ide1: reset: success
Jun 23 23:04:30 college kernel: hdd: dma_timer_expiry: dma status ==
0x41
Jun 23 23:04:30 college kernel: hdd: DMA timeout error
Jun 23 23:04:30 college kernel: hdd: dma timeout error: status=0x58
{ DriveReady SeekComplete DataRequest }
Jun 23 23:04:30 college kernel: ide: failed opcode was: unknown
Jun 23 23:04:30 college kernel: hdd: status error: status=0x50
{ DriveReady SeekComplete }
Jun 23 23:04:30 college kernel: ide: failed opcode was: unknown
Jun 23 23:04:30 college kernel: hdd: no DRQ after issuing
MULTWRITE_EXT

Jun 23 23:04:30 college kernel: ide1: reset: success
Jun 23 23:04:30 college kernel: hdd: dma_timer_expiry: dma status ==
0x41
Jun 23 23:04:30 college kernel: hdd: DMA timeout error
Jun 23 23:04:30 college kernel: hdd: dma timeout error: status=0x58
{ DriveReady SeekComplete DataRequest }
Jun 23 23:04:30 college kernel: ide: failed opcode was: unknown
Jun 23 23:04:30 college kernel: hdd: status timeout: status=0xd0
{ Busy }
Jun 23 23:04:30 college kernel: ide: failed opcode was: unknown
Jun 23 23:04:30 college kernel: hdd: no DRQ after issuing
MULTWRITE_EXT
Jun 23 23:04:30 college kernel: ide1: reset: success
Jun 23 23:04:30 college kernel: hdd: dma_timer_expiry: dma status ==
0x41
Jun 23 23:04:30 college kernel: hdd: DMA timeout error
Jun 23 23:04:30 college kernel: hdd: dma timeout error: status=0x58
{ DriveReady SeekComplete DataRequest }
Jun 23 23:04:30 college kernel: ide: failed opcode was: unknown
Jun 23 23:04:30 college kernel: hdd: status timeout: status=0xd0
{ Busy }
Jun 23 23:04:30 college kernel: ide: failed opcode was: unknown
Jun 23 23:04:30 college kernel: hdd: no DRQ after issuing
MULTWRITE_EXT
Jun 23 23:04:30 college kernel: ide1: reset: success
Jun 27 01:30:02 college kernel: hdd: status error: status=0x7f
{ DriveReady DeviceFault SeekComplete DataRequest CorrectedError Index
Error }
Jun 27 01:30:02 college kernel: hdd: status error: error=0x7f
{ DriveStatusError UncorrectableError SectorIdNotFound
TrackZeroNotFound AddrMarkNotFound }, LBAsect=149570080178047,
high=8915071, low=8355711, sector=40984406
Jun 27 01:30:02 college kernel: ide: failed opcode was: unknown
Jun 27 01:30:02 college kernel: hdd: drive not ready for command
Jun 27 01:30:09 college kernel: ide1: reset: success
Jun 27 01:57:51 college kernel: hdd: status error: status=0x7f
{ DriveReady DeviceFault SeekComplete DataRequest CorrectedError Index
Error }
Jun 27 01:57:51 college kernel: hdd: status error: error=0x7f
{ DriveStatusError UncorrectableError SectorIdNotFound
TrackZeroNotFound AddrMarkNotFound }, LBAsect=149568184352639,
high=8914958, low=8355711, sector=40986358
Jun 27 01:57:51 college kernel: ide: failed opcode was: unknown
Jun 27 01:57:51 college kernel: hdd: drive not ready for command
Jun 27 01:58:13 college kernel: ide1: reset: success

The controller and hard drive are certainly having a hard time
communicating, but I can't say who is unwell.

Get the manufacturer's maintenance floppy for that drive and run it.
I've seen error messages like that when the low-level format on a
hard drive fades to the point that the controller can't find tracks.
If that's the problem, you can "fix" it for a while by doing one or
a few low-level formats followed by writing zeroes to the drive to
force remapping of bad sectors (dd if=/dev/zero of=/dev/hda bs=1024).
Better to replace it though.

 

[Erik Max Francis]

Does it look like my hard disk is failing here, or is there any chance
it could be the IDE controller?

The login screen is responsive but ssh into this PC doesn't work.
After typing in the password the system locks up, and Ctrl-Alt-Del
gives the message:
college login: /etc/initdscript: line 106: /sbin/shutdown: Input/
Output error

Something's either wrong with your hard drive controller or the hard
drive. Have you tried to check the SMART status of your drive? You can
ask the drive itself whether or not it thinks it's sick.

man smartctl

 

[sk8terg1rl]

Something's either wrong with your hard drive controller or the hard
drive. Have you tried to check the SMART status of your drive? You can
ask the drive itself whether or not it thinks it's sick.

Thanks everyone. I'll bring in a CD-RW into college on Monday to do a
Seatools boot test. I was just hoping someone could say for certain
just by looking at the logs what the problem was. I wonder if this is
a quality control issue thanks to Seagate acquiring Maxtor.

skate xx

 

[~misfit~]

Thanks everyone. I'll bring in a CD-RW into college on Monday to do a
Seatools boot test. I was just hoping someone could say for certain
just by looking at the logs what the problem was. I wonder if this is
a quality control issue thanks to Seagate acquiring Maxtor.

skate xx

Hey skate,

Maybe not applicable in this case but I had a (Windows) machine that refused
to boot 3 out of 4 times, had to keep hard-restarting it. Tried the
secondary IDE and it helped for a while until that too gave the same
problem. Ran Seatools and Spinrite with no problems. I decided that it must
be the controller so bought a PCI/IDE controller and stil had the same
problem!!

Ya know what it was? Capacitors on the mobo failing so that the CPU vcore
was unstable, especially at startup when everything was spinning up! I
raised vcore a couple notches in BIOS and all was well. Frustrating, weeks
of messing around wth the HDD/controllers and the problem was actually the
CPU/mobo. (Once it started it ran Prime95 just fine, it was just sagging at
startup).

Just a thought, sometimes it pays to think outside the box.....

Luck,

 

[Plato]

Does it look like my hard disk is failing here, or is there any chance
it could be the IDE controller?

As an aside. Always replace your IDE cables with brand new ones whenever
trying to fix a hardware drive problem.

 

[Jean-David Beyer]

Maybe not applicable in this case but I had a (Windows) machine that refused
to boot 3 out of 4 times, had to keep hard-restarting it. Tried the
secondary IDE and it helped for a while until that too gave the same
problem. Ran Seatools and Spinrite with no problems. I decided that it must
be the controller so bought a PCI/IDE controller and stil had the same
problem!!

Ya know what it was? Capacitors on the mobo failing so that the CPU vcore
was unstable, especially at startup when everything was spinning up! I
raised vcore a couple notches in BIOS and all was well. Frustrating, weeks
of messing around wth the HDD/controllers and the problem was actually the
CPU/mobo. (Once it started it ran Prime95 just fine, it was just sagging at
startup).

Just a thought, sometimes it pays to think outside the box.....

Thinking outside another box, could it have been a marginal power supply
that had enough power to run the computer, but not start it? If the power
supply output is too low, the vcore regulator might not be able to produce
the required voltage for the processor(s).

On my main system, the hard drives are SCSI, so I have them set up to start
one at a time. Actually, two at a time, since they are not all on the same
controller.

 

[~misfit~]

Thinking outside another box, could it have been a marginal power
supply that had enough power to run the computer, but not start it?
If the power supply output is too low, the vcore regulator might not
be able to produce the required voltage for the processor(s).

Hehee! Good thinking. Actually it /was/ the mobo though. I replaced most of
the capacitors around the CPU slot and am still using it now. With the same
PSU. (I tried swapping out the PSU before replacing the caps/while waiting
for the caps) It's a Barton XP2500+ overclocked to 3200+ speed (200Mhz FSB)
and with the vcore actually *dropped* from 1.65V to 1.60V. Before I replaced
the caps the high/low page in MBM5 showed that the vcore (once in Windows)
was fluctuating quite a bit. New caps stopped that.

On my main system, the hard drives are SCSI, so I have them set up to
start one at a time. Actually, two at a time, since they are not all
on the same controller.

Always a good idea if you have the option and multiple discs. I have three
SATA HDDs in this machine (250 x 1 and 320 x 2)but they're on a fairly basic
PCI-SATA controller without that option. Some SATA mobos give you the option
of HDD spin-up delay.

Cheers,

 

[Jean-David Beyer]

Jean-David Beyer wrote:

Hehee! Good thinking. Actually it /was/ the mobo though. I replaced most of
the capacitors around the CPU slot and am still using it now.

Wow! I used to be a hardware techie when I graduated from college around
1960. I would have no trouble soldering 2-sided PC boards. But the MoBo in
my present computer is 12 layers, and I would be scared to death to approach
it with a soldering iron for fear of messing up the inner wiring. Also,
unless soldering irons are designed differently from the early 1960s, the
electrical leakage would probably cause problems... . And they are too big
for the super-tiny conductors.

With the same
PSU. (I tried swapping out the PSU before replacing the caps/while waiting
for the caps) It's a Barton XP2500+ overclocked to 3200+ speed (200Mhz FSB)
and with the vcore actually *dropped* from 1.65V to 1.60V. Before I replaced
the caps the high/low page in MBM5 showed that the vcore (once in Windows)
was fluctuating quite a bit. New caps stopped that.


Always a good idea if you have the option and multiple discs. I have three
SATA HDDs in this machine (250 x 1 and 320 x 2)but they're on a fairly basic
PCI-SATA controller without that option. Some SATA mobos give you the option
of HDD spin-up delay.

My MoBo and power supply have one 24-pin cable for the Mobo, and an
additional 8-pin cable for the processors. Each processor gets 4 conductors:
two ground and two 12V ones. Somehow the regulator gets this down to 1.44
volts for my processors, and it could not be a series or shunt regulator,
because each processor takes about 70 Amp, and the power supply cannot
provide that. It must be some kind of reverse voltage doubler type circuit,
but I have never tried to reverse-engineer that to see how to do it. Some
kind of capacitive switching trick, I suppose.

 

[Paul]

Wow! I used to be a hardware techie when I graduated from college around
1960. I would have no trouble soldering 2-sided PC boards. But the MoBo in
my present computer is 12 layers, and I would be scared to death to approach
it with a soldering iron for fear of messing up the inner wiring. Also,
unless soldering irons are designed differently from the early 1960s, the
electrical leakage would probably cause problems... . And they are too big
for the super-tiny conductors.

My MoBo and power supply have one 24-pin cable for the Mobo, and an
additional 8-pin cable for the processors. Each processor gets 4 conductors:
two ground and two 12V ones. Somehow the regulator gets this down to 1.44
volts for my processors, and it could not be a series or shunt regulator,
because each processor takes about 70 Amp, and the power supply cannot
provide that. It must be some kind of reverse voltage doubler type circuit,
but I have never tried to reverse-engineer that to see how to do it. Some
kind of capacitive switching trick, I suppose.

Some of the motherboards are four layer, and some are six layer. Layers cost
money (10% added cost per pair of added layers, for the blank PCB), so there
is good incentive to stick with four layer boards. And current twelve layer
boards can be just as reliable as a four layer board. You just aren't likely
to see them in a factory making millions of boards a year.

You can get soldering stations, with "controlled tip" irons. Some even come
with a dial, so you can set the tip temperature, and there is a temp display
on the soldering station. That is a much more precise way of soldering than
an uncontrolled $25 iron from Radio Shack.

In terms of power level, I have a 15W, a 25W, and an 80W cheap uncontrolled
set of irons for home use. Using the 15W and the 25W, it is possible to heat
the ends of SMT resistors and caps, and remove or replace them. The 80W
is for heating up solder joints that have no isothermal relief on the
hole pattern. Like if an electrolytic was hard to get out, I'd point the
80W at it. Computer motherboards use thin foil, and aren't that tolerant
of abuse, so you cannot leave the iron touching them for long periods of
time.

You can get datasheets for the voltage regulators used on motherboards.
This is the one on my motherboard - a "synchronous buck controller".
These devices are so sophisticated now, they have separate eight pin
chips to drive the MOSFET gates. That helps to keep the main chip
cooler when it is running. (At one time, the main chip did everything
and got hotter as a result.) I love the complexity of the design process
for these - in the real world, you have to prototype the circuit first,
and tune it up, before connecting the design to a processor. But that
hasn't stopped hobbyists from doing "droop" mods to these circuits.

http://www.analog.com/UploadedFiles/Data_Sheets/ADP3180.pdf

A buck regulator does not isolate input from output. In this case,
it takes 12V on one side, and has 1.5V on the other side. Inductors
(those powdered iron toroids with three pieces of wire wound in
parallel) store the switched energy. The multiple capacitors
on input and output, smooth the ripple. Since the MOSFETs are
fully on or are open circuit, there is little power dissipated
in them. At least, compared to the alternative, which would be the
more wasteful, linear regulator. A motherboard switching regulator
could be 90% efficient, and switching regulation is preferred for
high current loads. Simple linear designs, tend to stop at about
5 amps (some of those are used for DIMM termination voltages and
the like). The multiphase motherboard processor regulator, can in
some cases produce 100 amps of current (at 1.5V or less), on demand.
The motherboards used to overclock Pentium D 805 processors, were
being called on to provide even more than that. So that regulator
concept is quite an amazing achievement.

http://www.overclockers.com/articles1386/index03.asp

Paul

 

[~misfit~]

Wow! I used to be a hardware techie when I graduated from college
around 1960. I would have no trouble soldering 2-sided PC boards. But
the MoBo in my present computer is 12 layers, and I would be scared
to death to approach it with a soldering iron for fear of messing up
the inner wiring. Also, unless soldering irons are designed
differently from the early 1960s, the electrical leakage would
probably cause problems... . And they are too big for the super-tiny
conductors.

Heh! I'm a bit of a cowboy. I just used a $10 soldering iron from the NZ
equivalent of the Radio Shack and took my time. My board is 6 layers, I have
three of that model here and I've swapped out the capacitors on all of them
without problems. They're Soltek SL75FRN2-L Nforce2 Ultra 400 Socket A
boards. Great chipset, nice board layout but suffered from cheap components.
Soltek went out of business not long after they produced these boards. For a
while they'd send out packs of replacement caps free-of-charge to anyone who
had a board that gave problems. Unfortunately I didn't find that out until
*after* they'd stopped doing that and had closed the doors.

I've done the same on several other boards too, I have three mATX Gigabyte
boards here (with on-board graphics, audio and ethernet) that are running
Intel Tualatin Celerons (I have a 1.2, a 1.3 and a 1.4Ghz). They're handy
machines for low-power consumption applications, leaving running 24/7 for
torrents or, with a PCI/SATA card, as a NAS box.

I guess that I'm just lucky with a soldering iron. Then again, I am/was
legendary (heh!) in alt.comp.hardware.overclocking for having been doing a
pin-mod on a coppermine CPU and removing the wrong pin! Not to worry, I did
what I was told was impossible and soldered a replacement pin onto the CPU
using the same $10 tool.

I've only had one failure when replacing capacitors and I'm convinced that
it was due to me using second-hand caps from 'donor' 486/P1 boards to get
the 16-odd caps I needed to re-cap a Gigabyte 440BX 133 RAID board. It was
my first, <sigh> before I managed to find sources for new low ESR caps. I've
had quite a bit of trouble with BX boards as I used to run Tualatins in them
using Upgradeware adapters. It tends to stress the CPU power components as
they weren't designed to produce low voltage/high current. I still run one
BX/Tui combo daily. It has a 60 x 25mm fan blowing directly onto the CPU
power MOSFETs as, without that, they very quickly become too hot to touch.
In fact I had three indentical boards and with one of them the solder pad
that one of the MOSFETs sits on liquified and the FET slipped out of place!
That was before I realised I had to have some pretty aggressive cooling in
that area. It helps the caps too.

My MoBo and power supply have one 24-pin cable for the Mobo, and an
additional 8-pin cable for the processors. Each processor gets 4
conductors: two ground and two 12V ones.

Sounds like a good set-up. Sadly, my finances gave away about 5 years ago
when I had an accident, injured my back and susequently lost my business and
life-savings. Consequently the newest tech I have consists of nForce2/Barton
systems which will have to serve me well into the future. Hence learning how
to re-cap mobos with a $10 soldering iron. You don't have many options when
you're on welfare.

I've done a couple of 64-bit/dual-core builds for friends but really my
hands-on knowledge is a bit out-of-date now. It's no problem when everything
works properly, a 12 year-old could build a PC. It's having the knowledge to
troubleshoot that makes the difference and that mainly comes from
experience. You can only be /so/ good just by reading web-pages/news-groups.

Somehow the regulator gets
this down to 1.44 volts for my processors, and it could not be a
series or shunt regulator, because each processor takes about 70 Amp,
and the power supply cannot provide that. It must be some kind of
reverse voltage doubler type circuit, but I have never tried to
reverse-engineer that to see how to do it. Some kind of capacitive
switching trick, I suppose.

Yeah, Paul posted a good explaination of that. CPU Vregs are a lot more
efficient that they were a few years ago.

Cheers,

 

[CBFalconer]

.... big snip ...

Sounds like a good set-up. Sadly, my finances gave away about 5
years ago when I had an accident, injured my back and susequently
lost my business and life-savings. Consequently the newest tech I
have consists of nForce2/Barton systems which will have to serve
me well into the future. Hence learning how to re-cap mobos with
a $10 soldering iron. You don't have many options when you're on
welfare.

Don't know how bad your back is, but if possible watch the
roads/dumps/etc. for discarded hardware, within about 8 years of
age. You can get some nice systems that way, for free.

 

[kony]

I guess that I'm just lucky with a soldering iron. Then again, I am/was
legendary (heh!) in alt.comp.hardware.overclocking for having been doing a
pin-mod on a coppermine CPU and removing the wrong pin! Not to worry, I did
what I was told was impossible and soldered a replacement pin onto the CPU
using the same $10 tool.

I remember that... your success motivate me to try a few
different methods of reattaching pins. I tried tinning the
pin then using flux, tried that liquid automotive defogger
copper paint (had to scrap away excess later), but neither
of these worked as well for me as plain old plumbers tinning
flux, the two former attempts resulting in the pin coming
off when I tried to fasten it in the socket while the last
is still running today. The key to using it was to make
sure the iron tip was very clean, de-tinned, so while
heating the pin it didn't leave but a trivial amount of
solder residue from the tip, only relying on the solder dust
in the tinning flux to make the bond.

 

[~misfit~]

I remember that... your success motivate me to try a few
different methods of reattaching pins. I tried tinning the
pin then using flux, tried that liquid automotive defogger
copper paint (had to scrap away excess later), but neither
of these worked as well for me as plain old plumbers tinning
flux, the two former attempts resulting in the pin coming
off when I tried to fasten it in the socket while the last
is still running today. The key to using it was to make
sure the iron tip was very clean, de-tinned, so while
heating the pin it didn't leave but a trivial amount of
solder residue from the tip, only relying on the solder dust
in the tinning flux to make the bond.

Hehee! I just tinned the end off of a pin off a 486 (the actual pin off the
Coppermine was bent). The 486 pin had a wee 'mushroom' at the end. then I
dipped the soldering iron into the crater where the pin was, leaving the
tiniest bit of solder in the 'hole'. I then wiped the soldering iron clean
with a damp cloth, let it reach it's temp again, held the pin in place with
tweezers and help the iron against the side of the pin for a couple seconds.

I still have the CPU, a Celeron 600 that does 900 easy. I've had it in and
out of sockets a few times since, the pin holds on fine. I can recognise it
by the little bit of solder rsidue on the side of the pin. I guess I didn't
get the iron completely clean. <g> It's in a plastic CPU package now, in a
drawer. <sigh> What once was a great achievement is now a keepsake. I'm too
sentimental to throw it away.

Cheers,

 

[~misfit~]

Don't know how bad your back is, but if possible watch the
roads/dumps/etc. for discarded hardware, within about 8 years of
age. You can get some nice systems that way, for free.

My backs bad, but if I eat my whole day's PKs at once I can get some things
done. I'm fine when my spine is straight. leaning over PC cases is painful
though.

I have plenty of older systems here; Couple Bartons, an Applebred Duron 1.6,
three tuis, three Coppermines, even an old IBM system that's running a
Katmai 500 (I got it when it was running a Deschutes 350). All in working
condition, all on my LAN.

I spoke to the local refuse transfer place and they're only interested if I
take *all* computer-related stuff. I worked out that, for each usable
system, I'd have about a trailer load of junk, old printers and monitors
etc. that I'd have to pay to get rid of. Also, I only have a tiny hatchback,
no trailer hitch. :-(

I need to get onto some of my friends in IT, that's where I got most of my
best 'junk'. I have a dual-CPU PIII 500 server coming but even that's of
little use these days.

Cheers, Gotta go, Top Gear's on.

 

[~misfit~]

Hehee! I just tinned the end off of a pin off a 486 (the actual pin
off the Coppermine was bent). The 486 pin had a wee 'mushroom' at the
end. then I dipped the soldering iron into the crater where the pin
was, leaving the tiniest bit of solder in the 'hole'. I then wiped
the soldering iron clean with a damp cloth, let it reach it's temp
again, held the pin in place with tweezers and help

That "help" should read "held". :-(