Low Level Format of 1.4 Mb Floppy Isn't Happening

[Martin McCormick]

The fdformat utility allows one to low-level format a
floppy disk. I need to convince an older PC that it should boot
from one of it's usb ports and the Plop Project has a boot
manager that fits on a 1.4Mb floppy. It is called plpbt.iso
so I got a 1.4Mb disk and copied the boot manager in to it and
then decided to start from scratch after it didn't work so I used
a bulk tape eraser and completely erased the disk which is
something I used to do back in the DOS days as it makes the
medium magnetically sterile. I would then format it again and use
it.

Floppies are a relic these days but the system I want
to usb boot on is a relic so the boot manager should work when I
install it correctly.

What happened after I zapped the floppy is that fdformat will not
run because it sees no pre-existing format information. I am not
surprised at this but I seem to remember being able to do
fdformat or something similar some time ago which would lay new
tracks in the magnetic dessert.

You then use mformat if you want a MSDOS disk but nothing
else happens until the initial low-level format is done.

What am I forgetting or what has changed?

I probably haven't messed with floppy disks in a few
years but one used to be able to format from a bare floppy
without much trouble.

Thanks for any constructive suggestions.

Martin McCormick WB5AGZ

[Henrique de Moraes Holschuh]

Post by Martin McCormick
then decided to start from scratch after it didn't work so I used
a bulk tape eraser and completely erased the disk which is

...

Post by Martin McCormick
What happened after I zapped the floppy is that fdformat will not
run because it sees no pre-existing format information. I am not

Did you read the fdformat manpage? Please do.

Then, tell us *exactly* what you are doing, and what the kernel reports
("dmesg | tail -n 30" should do). It is not impossible for a kernel bug
to be around, not many people use bare-hardware floppy devices nowadays
(the USB-attached ones don't count, they're hardware/firmware-mediated).

[Mirko Parthey]

Post by Martin McCormick
What happened after I zapped the floppy is that fdformat will not
run because it sees no pre-existing format information.
What am I forgetting or what has changed?

According to the fdformat man page, you should first set the disk
parameters with setfdprm, or use a device like /dev/fd0H1440 which
refers to a specific format.

If any of the device nodes are missing, you can create them with mknod,
the major and minor numbers are given in the fdformat man page as well.

Other options to boot an old machine could be
a Compact Flash card with a CF-to-IDE adapter,
or an Ethernet card with PXE support and a BOOTP/DHCP and TFTP server.

Regards,
Mirko

[Martin McCormick]

Post by Mirko Parthey

Post by Martin McCormick
What happened after I zapped the floppy is that fdformat will not
run because it sees no pre-existing format information.
What am I forgetting or what has changed?

According to the fdformat man page, you should first set the disk
parameters with setfdprm, or use a device like /dev/fd0H1440 which
refers to a specific format.
If any of the device nodes are missing, you can create them with mknod,
the major and minor numbers are given in the fdformat man page as well.

Thank you as I think the problem is a knowledge base problem on
my part. The device node for /dev/fd0 is there and works properly
if the disk contains so much as a low-level format. The man page
for fdformat says:

________________________________________
The generic floppy devices, /dev/fd0 and /dev/fd1, will fail to work
with fdformat when a non-standard format is being used, or if the for-
mat has not been autodetected earlier. In this case, use setfdprm(8)
to load the disk parameters.
________________________________________

Where I am confused is whether or not setfdprm must be
manually stuffed with the number of tracks, number of sectors,
etc or whether /etc/mediaprm can be made to fill in all those
data. I am not modifying the format so any list of normal
parameters for a 1.44 Mb floppy will do.

Martin

[Henrique de Moraes Holschuh]

Post by Martin McCormick
The generic floppy devices, /dev/fd0 and /dev/fd1, will fail to work
with fdformat when a non-standard format is being used, or if the for-
mat has not been autodetected earlier. In this case, use setfdprm(8)
to load the disk parameters.
________________________________________
Where I am confused is whether or not setfdprm must be
manually stuffed with the number of tracks, number of sectors,
etc or whether /etc/mediaprm can be made to fill in all those

Interestingly enough, I don't seem to have either setfdprm(8) or
setfdprm itself in my amd64 Debian Jessie system, so I can't check this
for you. /etc/mediaprm should be enough, but...

Post by Martin McCormick
data. I am not modifying the format so any list of normal
parameters for a 1.44 Mb floppy will do.

In that case, unless the kernel driver is having trouble, it should be
sufficient to:

fdformat /dev/fd0H1440 (for fd0).

Don't expect this to work on anything but a real FDD, connected directly
to a FDD controller in your motherboard (i.e. no USB). And don't do it
from a VM, it is unlikely to work.

If it fails to format (but does try, i.e. it turns the FDD on, spins the
floppy, and tries to write a track), please ensure the FDD heads are
clean[1]. Formatting a disk scrambled by too strong a magnet could take
quite a lot more magnetic flux from the FDD head to "sit".


[1] and don't get the cleaning procedure wrong, or you will kill the
FDD, the diskette, or both. Remeber to give enough time for the
cleaning solution to evaporate completely from the heads before
attempting to use the FDD.

[t***@tuxteam.de]

On Sun, Jan 15, 2017 at 12:45:54PM -0200, Henrique de Moraes Holschuh wrote:

[...]

Post by Henrique de Moraes Holschuh
Interestingly enough, I don't seem to have either setfdprm(8) or
setfdprm itself in my amd64 Debian Jessie system, so I can't check this
for you. /etc/mediaprm should be enough, but...

FWIW, setfdprm seems to be in a package called fdutils:

| ***@rasputin:~$ apt-file search setfdprm
| fdutils: /usr/bin/setfdprm
| fdutils: /usr/share/man/man1/setfdprm.1.gz
| manpages-es-extra: /usr/share/man/es/man8/setfdprm.8.gz
| manpages-ja: /usr/share/man/ja/man8/setfdprm.8.gz
| manpages-pl: /usr/share/man/pl/man1/setfdprm.1.gz

For the rest, good advice IMHO

regards
- -- t

[Gene Heskett]

Post by Martin McCormick

Post by Mirko Parthey

Post by Martin McCormick
What happened after I zapped the floppy is that fdformat will not
run because it sees no pre-existing format information.
What am I forgetting or what has changed?

According to the fdformat man page, you should first set the disk
parameters with setfdprm, or use a device like /dev/fd0H1440 which
refers to a specific format.
If any of the device nodes are missing, you can create them with
mknod, the major and minor numbers are given in the fdformat man
page as well.

Thank you as I think the problem is a knowledge base problem on
my part. The device node for /dev/fd0 is there and works properly
if the disk contains so much as a low-level format. The man page
________________________________________
The generic floppy devices, /dev/fd0 and /dev/fd1, will fail
to work with fdformat when a non-standard format is being used, or if
the for- mat has not been autodetected earlier. In this case, use
setfdprm(8) to load the disk parameters.
________________________________________
Where I am confused is whether or not setfdprm must be
manually stuffed with the number of tracks, number of sectors,
etc or whether /etc/mediaprm can be made to fill in all those
data. I am not modifying the format so any list of normal
parameters for a 1.44 Mb floppy will do.
Martin

To add to the confusion, Martin, windows won the battle of fdc
capabilities around a decade back with the fdc chip makers. This $300
ASUS motherboard, when I bought it to build this machine in 2007, one of
the first to support the AMD Phenom quad core chips, has an fdc that is
unhappy with a 256 byte sector size, and will crash/lock this machine,
so bad that it takes the long push on the power button, doing a
powerdown reset, if you tell it to use base zero sector numbering and
then try to format a disk. So while one might be able to build a
mediaparm file that should work to format an os9 disk so you can haul
something you have downloaded to one of your TRS-80 Color Computers I
have several of it the basement, I have had to resort in the early days,
of hooking up an rs-232 cable, firing up sz and restricting its speed to
4800 baud, and packet size to 256 bytes just to move the code. That was
somewhat like watching grass grow. But it worked. So if intending to use
these floppies for a sneakernet to a "legacy" machine, don't be
surprised if you only succeed in having problems. Its not your fault,
its the fdc chip in your machine.

You'll have better, if slower, luck with rs-232 and zmodem.

Cheers, Gene Heskett

[Henrique de Moraes Holschuh]

Post by Gene Heskett
To add to the confusion, Martin, windows won the battle of fdc
capabilities around a decade back with the fdc chip makers. This $300

Hmm, no. More like the fdc's all became "IP" ("intellectual property
module") verilog/vhdl to be added as an extra device in ASIC "super-io"
designs.

It is quite possible that one such base fdc "IP" module design used by
lots of such super-io ASICs cut corners (or someone screwed it up, no
way to know for sure), and did not implement the full spec, or did it
badly.

I would be hard pressed to tell you which super-io parts do have a
non-joke fdc. The only thing I have that uses them is a MSX2+, and it
has its own discrete FDC that has been working flawlessly for 25 years.

Post by Gene Heskett
ASUS motherboard, when I bought it to build this machine in 2007, one of
the first to support the AMD Phenom quad core chips, has an fdc that is
unhappy with a 256 byte sector size, and will crash/lock this machine,
so bad that it takes the long push on the power button, doing a

Are you sure this isn't a kernel driver issue? The Linux floppy driver
doesn't see much love, and the people who *do* still use it are *not*
good at helping. We don't see much posts of the "I will help" sort in
LKML.

Post by Gene Heskett
You'll have better, if slower, luck with rs-232 and zmodem.

While true, it won't help boot an old x86 box :p You'd have better luck
with PXE-based network boot, or something. Should be easy to find an
old NIC with a PXE boot ROM for US$ 10 or so on ebay, if your old box
lacks one :p

[Martin McCormick]

I must first say a big thank you to everybody who helped.

I got it working and here is what happened and how to
format a totally blank diskette.

The man page for setfdprm should have had at least one
example of how to use it since syntax, especially for something
one doesn't use every day, is not intuitive and can cause hours of prodding and
head-scratching/banging.

In my searching via Google and reading some of the
articles and discussion group postings, I learned that there is a
companion to setfdprm which is getfdprm. I put in a good and
formatted floppy and ran getfdprm on a 1.44 Mb disk. The
application produced

DS HD sect=18

This occurs after you have spun the disk to mount it or done
anything else to make the disk turn for reading.

If you eject the diskette, the value stored in setfdprm
vaporizes, assuming that one's floppy drive's disk-change switch
is working.

I figured that this value or

DS DD sect=9

for a 720 K diskette was worth passing to setfdprm so I wrote the
following shell script:

#!/bin/sh
#1.44Mb disk Uncomment line below.
#setfdprm /dev/fd0 DS HD sect=18
#720K disk Uncomment line below to use it.
#setfdprm /dev/fd0 DS HD sect=9
#fdformat /dev/fd0

It will begin by identifying that this is a double-sided disk and
will start laying down tracks from 0 to 79 and then verify their
presence.

You can skip the verification process in fdformat but I
don't advise it as I tried some old 720K disks that formatted all
80 tracks and then track 0 was dead on all of them.

Another warn-out disk formatted and verified until track
76 was reached and then it cratered with the same error that
befell the 720K disks.

A new (unused until now) 1.44 Mb disk formatted and
verified all 80 tracks so the drive appears to have no
electro mechanical issues on good media.

On the disk with a dead track 76, the error caused the
heads to seek all the way back to track 0 and then forward to 76.
One could hear it rip-sawing back and forth which was probably
good exercise for a drive that doesn't get much usage these days.
After trying to read track 76 many times, the verification
aborted with a message stating that this track was bad.

Anyway, this has been an interesting journey down memory
lane. Stuff just works a lot better these days than it used to.

Martin McCormick WB5AGZ

[Doug]

Post by Martin McCormick
I must first say a big thank you to everybody who helped.
I got it working and here is what happened and how to
format a totally blank diskette.
The man page for setfdprm should have had at least one
example of how to use it since syntax, especially for something
one doesn't use every day, is not intuitive and can cause hours of prodding and
head-scratching/banging.
In my searching via Google and reading some of the
articles and discussion group postings, I learned that there is a
companion to setfdprm which is getfdprm. I put in a good and
formatted floppy and ran getfdprm on a 1.44 Mb disk. The
application produced
DS HD sect=18
This occurs after you have spun the disk to mount it or done
anything else to make the disk turn for reading.
If you eject the diskette, the value stored in setfdprm
vaporizes, assuming that one's floppy drive's disk-change switch
is working.
I figured that this value or
DS DD sect=9
for a 720 K diskette was worth passing to setfdprm so I wrote the
#!/bin/sh
#1.44Mb disk Uncomment line below.
#setfdprm /dev/fd0 DS HD sect=18
#720K disk Uncomment line below to use it.
#setfdprm /dev/fd0 DS HD sect=9
#fdformat /dev/fd0
It will begin by identifying that this is a double-sided disk and
will start laying down tracks from 0 to 79 and then verify their
presence.
You can skip the verification process in fdformat but I
don't advise it as I tried some old 720K disks that formatted all
80 tracks and then track 0 was dead on all of them.
Another warn-out disk formatted and verified until track
76 was reached and then it cratered with the same error that
befell the 720K disks.
A new (unused until now) 1.44 Mb disk formatted and
verified all 80 tracks so the drive appears to have no
electro mechanical issues on good media.
On the disk with a dead track 76, the error caused the
heads to seek all the way back to track 0 and then forward to 76.
One could hear it rip-sawing back and forth which was probably
good exercise for a drive that doesn't get much usage these days.
After trying to read track 76 many times, the verification
aborted with a message stating that this track was bad.
Anyway, this has been an interesting journey down memory
lane. Stuff just works a lot better these days than it used to.
Martin McCormick WB5AGZ

About a year ago, or maybe two, I bought a pack of Radio Shack floppies.
I think ten or twelve in
the pack. Most of them had errors right off the bat! I don't know if
you can trust old floppies!
--doug

[Brian]

On Sun 15 Jan 2017 at 15:36:08 -0600, Martin McCormick wrote:

[A trip down Memory Lane snipped]

For those who may want to use Plop boot manager, it is far quicker and
more reliable to do so from GRUB on the hard disc. The Plop website has
details. Struggling with a pile of floppies is no fun and a floppy which
works today ivery likely not to do so next month.

[Gene Heskett]

Post by Martin McCormick
I must first say a big thank you to everybody who helped.
I got it working and here is what happened and how to
format a totally blank diskette.
The man page for setfdprm should have had at least one
example of how to use it since syntax, especially for something
one doesn't use every day, is not intuitive and can cause hours of
prodding and head-scratching/banging.
In my searching via Google and reading some of the
articles and discussion group postings, I learned that there is a
companion to setfdprm which is getfdprm. I put in a good and
formatted floppy and ran getfdprm on a 1.44 Mb disk. The
application produced
DS HD sect=18
This occurs after you have spun the disk to mount it or done
anything else to make the disk turn for reading.
If you eject the diskette, the value stored in setfdprm
vaporizes, assuming that one's floppy drive's disk-change switch
is working.
I figured that this value or
DS DD sect=9
for a 720 K diskette was worth passing to setfdprm so I wrote the
#!/bin/sh
#1.44Mb disk Uncomment line below.
#setfdprm /dev/fd0 DS HD sect=18
#720K disk Uncomment line below to use it.
#setfdprm /dev/fd0 DS HD sect=9
#fdformat /dev/fd0
It will begin by identifying that this is a double-sided disk and
will start laying down tracks from 0 to 79 and then verify their
presence.
You can skip the verification process in fdformat but I
don't advise it as I tried some old 720K disks that formatted all
80 tracks and then track 0 was dead on all of them.
Another warn-out disk formatted and verified until track
76 was reached and then it cratered with the same error that
befell the 720K disks.
A new (unused until now) 1.44 Mb disk formatted and
verified all 80 tracks so the drive appears to have no
electro mechanical issues on good media.
On the disk with a dead track 76, the error caused the
heads to seek all the way back to track 0 and then forward to 76.
One could hear it rip-sawing back and forth which was probably
good exercise for a drive that doesn't get much usage these days.
After trying to read track 76 many times, the verification
aborted with a message stating that this track was bad.
Anyway, this has been an interesting journey down memory
lane. Stuff just works a lot better these days than it used to.
Martin McCormick WB5AGZ

And any drive thats 10 years old & maybe even some newer ones are
desperately in need of being pulled out, and any covers that are
blocking good access to the carriage drive screw, need to be uncovered
for a serious cleaning with alcohol or even acetone so that no debris or
solidified grease can be found on the screw, followed by a pebble of
lithium chassis grease worked back into the groove of the screw. Put it
back together & back into the computer. It the q-tip isn't too dirty,
wipe a thin film of grease on the rails the head carriage slides on, but
more likely it would get a fresh & some clean grease. It might work
another for another ten years!

Cheers, Gene Heskett

[John Darrah]

Post by Gene Heskett
And any drive thats 10 years old & maybe even some newer ones are
desperately in need of being pulled out, and any covers that are
blocking good access to the carriage drive screw, need to be uncovered
for a serious cleaning with alcohol or even acetone so that no debris or
solidified grease can be found on the screw, followed by a pebble of
lithium chassis grease worked back into the groove of the screw. Put it
back together & back into the computer. It the q-tip isn't too dirty,
wipe a thin film of grease on the rails the head carriage slides on, but
more likely it would get a fresh & some clean grease. It might work
another for another ten years!
Cheers, Gene Heskett

Just curious if you looked at the 'ufiformat' command.

-- john

[Gene Heskett]

Post by John Darrah

Post by Gene Heskett
And any drive thats 10 years old & maybe even some newer ones are
desperately in need of being pulled out, and any covers that are
blocking good access to the carriage drive screw, need to be
uncovered for a serious cleaning with alcohol or even acetone so
that no debris or solidified grease can be found on the screw,
followed by a pebble of lithium chassis grease worked back into the
groove of the screw. Put it back together & back into the computer.
It the q-tip isn't too dirty, wipe a thin film of grease on the
rails the head carriage slides on, but more likely it would get a
fresh & some clean grease. It might work another for another ten
years!
Cheers, Gene Heskett

Just curious if you looked at the 'ufiformat' command.
-- john

No John, I haven't, not having seen that word on screen before.

And since I've better ways of doing that for the last 3 or 4 years,
called Drivewire, which is far more capable than a floppy based
sneakernet, I probably won't look it up. To demo the diff, if I
substitute the drivewire version of the Nitros9 p module while dw is
running on this machine, I have a b&w laser printer sitting here I use
for all my b&w printing needs, spits out a file sent to /p on the coco3,
this cheap brother printer spits it out at 19 pages a minute. The
fastest printer I ever had connected to it in the last 30 years, was a
Oki 24 pinner that could do a bit over a page a minute, feeding it from
a parport plugged into the coco3.

It does a decent imitation of a floppy drive, moving data to/from a file
on this linux boxes hard drive, at about the same speed as a standard
density floppy drive on the coco3.

Many more things it can do as the protocol itself has as high as 64
independent data channels.

Whats not to like?

Cheers, Gene Heskett

[Brian]

Post by Martin McCormick
The fdformat utility allows one to low-level format a
floppy disk. I need to convince an older PC that it should boot
from one of it's usb ports and the Plop Project has a boot
manager that fits on a 1.4Mb floppy. It is called plpbt.iso
so I got a 1.4Mb disk and copied the boot manager in to it and
then decided to start from scratch after it didn't work so I used
a bulk tape eraser and completely erased the disk which is
something I used to do back in the DOS days as it makes the
medium magnetically sterile. I would then format it again and use
it.
Floppies are a relic these days but the system I want
to usb boot on is a relic so the boot manager should work when I
install it correctly.

Does the machine you are booting use GRUB as the boot manager?

[Martin McCormick]

Post by Brian
Does the machine you are booting use GRUB as the boot manager?

Surprisingly enough, yes. I wasn't sure but it certainly does.

Here's an explanation with a hope for a workable
solution.

I don't know how common this is but the BIOS' of two Dell
Optiplexes plus the BIOS of another Dell Dimension don't stay set
the way one would like them to. As a computer user who happens to
be blind, the setup process is a royal pain in the backside
because since there is no OS booted, the only output device is
the VGA output. I have to hook it up to a monitor and have
someone like my kind and patient wife or, before I retired, a
coworker change the boot sequence order from floppy-C:-CDROM to
CDROM-floppy-C: or CDROM-C:-floppy which prevents the hard drive
from grabbing the boot sequence each time.

Of course you can make the hard drive unbootable and
maybe force it to CDROM but one should not kick Murphy in the
knee cap when things are already going South. In this particular
case, I want to try a ubuntu installation CD to see if the
"speakup" software synthesis system works on this box and if so,
I will install it. If not, the obsolete installation already on
the hard drive does work but it sure needs to be replaced if
possible.

What happens is that we set a desired boot order and save
the settings. Things go fine for several months and then I try to
boot from CDROM one day and it's all back wrong again.

A good solution would be what I call the yank-back. If
one could backup the settings when good and the box decides to
clobber them, force them back.

The true ideal solution is for it not to happen at all
but you fight the war with the weapons you have, not what you
wish they were.

I don't think the CMOS batteries are dead because the
clock still seems to keep time so I am guessing that something
happens that the BIOS code sees as an error maybe during bootup
so it decides to re-order the sequence. One doesn't discover this
until the next time the CDROM doesn't spin up when a bootable
disk is in it. By the way, the CDROM works fine when mounting a
CD or ripping one.

Martin WB5AGZ

[Felix Miata]

Post by Martin McCormick
I don't know how common this is but the BIOS' of two Dell
Optiplexes plus the BIOS of another Dell Dimension don't stay set
the way one would like them to. As a computer user who happens to

What models are they? If of approximately 2002 to 2007 vintage, they could be
victims of the bad capacitor plague:
http://en.wikipedia.org/wiki/Capacitor_plague

I've resurrected several, but also failed on several, by replacing bad caps.

Post by Martin McCormick
be blind, the setup process is a royal pain in the backside
because since there is no OS booted, the only output device is
the VGA output. I have to hook it up to a monitor and have
someone like my kind and patient wife or, before I retired, a
coworker change the boot sequence order from floppy-C:-CDROM to
CDROM-floppy-C: or CDROM-C:-floppy which prevents the hard drive
from grabbing the boot sequence each time.

I sympathize. I have a bunch of different models Optiplexes. I've had three
different ones waste an obscene amount of my time lately. Each of these is
installed in locations that make opening them up for maintenance a significant
inconvenience.

The first, a 780 with a Dell PCIe DVI port card and connected via DisplayPort,
when shut down from Linux puts my new Dell 21:9 display to sleep in a fashion
from which it cannot be awakened until after I remove and replace the 780's
power cable. Even after a BIOS update it took quite some time to determine that
the only way to acquire normal behavior was to remove the DVI port card.

The second, a 745, even with latest BIOS, has taken to reporting low battery
voltage, after the previous battery voltage was reduced to nil. I replaced it
with a brand new battery with measured voltage up to specification, but it still
pauses at every power up to report low voltage even though clock is correct.

The third, a 270, has decided to randomly report failed RAM. Each time it has
done so I've opened it, wiggled or removed and replaced the sticks, and then it
would work normally. As I write this I'm running Memtest86+ on it after having
replaced both sticks with matched pair of different brand but same 3-3-3-8 CL3
specification.

Post by Martin McCormick
I don't think the CMOS batteries are dead because the
clock still seems to keep time so I am guessing that something
happens that the BIOS code sees as an error maybe during bootup
so it decides to re-order the sequence. One doesn't discover this
until the next time the CDROM doesn't spin up when a bootable
disk is in it. By the way, the CDROM works fine when mounting a
CD or ripping one.

Clocks will keep time even with voltage quite low. Batteries are cheap enough,
as little as $2US for a card of 5 or 6, that they should be replaced, or at
least checked with a voltmeter, if ever there is any indication of a problem
that might be CMOS related. It's also a good idea when a battery has needed to
be changed, to clear CMOS, then restore any desired custom settings.

[Martin McCormick]

Post by Felix Miata
What models are they? If of approximately 2002 to 2007 vintage, they could
http://en.wikipedia.org/wiki/Capacitor_plague

Interesting article. These dells could possibly be in
that group as one has a BIOS date of August 1, 2001 and another
is July 12, 2000. Both are Dell desktops so they could possibly
be in that group. By the way, very few electronic malfunctions
are quite as spectacular as an electrolytic capacitor that
develops a sudden and catastrophic internal short.

It sounds like the caps discussed in the article have a
safety valve in the form of the vent but nothing short of high
explosives equals an electrolytic filter cap popping.

It sounds like a firecracker and throws more aluminum
foil shrapnel than one could think would fit in the case.

I had one about the size of a thumb blow up on me once
and I could still find pieces of plastic and foil 6 months later.

Thanks for the link to the article.

Martin McCormick WB5AGZ

[Joe]

On Mon, 16 Jan 2017 14:53:33 -0600

Post by Martin McCormick

Post by Felix Miata
What models are they? If of approximately 2002 to 2007 vintage,
http://en.wikipedia.org/wiki/Capacitor_plague

Interesting article. These dells could possibly be in
that group as one has a BIOS date of August 1, 2001 and another
is July 12, 2000. Both are Dell desktops so they could possibly
be in that group. By the way, very few electronic malfunctions
are quite as spectacular as an electrolytic capacitor that
develops a sudden and catastrophic internal short.
It sounds like the caps discussed in the article have a
safety valve in the form of the vent but nothing short of high
explosives equals an electrolytic filter cap popping.
It sounds like a firecracker and throws more aluminum
foil shrapnel than one could think would fit in the case.
I had one about the size of a thumb blow up on me once
and I could still find pieces of plastic and foil 6 months later.
Thanks for the link to the article.

It doesn't have the whole story. I've seen thousands (literally: some
television cameras I have repaired contained over 100 dodgy capacitors)
of surface-mount aluminium electrolytics leaking. The electrolyte used
in these [Japanese] capacitors was obviously very conductive, but it
also dissolved copper. There was literally no limit to the types of
fault caused by some tracks disappearing and some being bridged
together.

The mention of 'amine' rings a bell, when hot this electrolyte smelled
of fish, so the quick way to distinguish between harmless flux left on
a PCB after manufacture, and the brownish electrolyte, was to touch a
soldering iron to it. Evil stuff.

[John Hasler]

The electrolyte used in these [Japanese] capacitors

They were Chinese knockoffs of a Japanese design. However there was a
proprietary trick that the Chinese missed and so they made and sold a
vast number of caps that turned out to be defective.

Electrolytics usually have burst membranes intended to prevent them from
exploding when they evolve gas and they usually work on low-voltage
ones. They often work on high-voltage caps too, but other times you can
get a nice bang. Big motor starter capacitors can be fun that way.

[Joe]

On Mon, 16 Jan 2017 15:54:05 -0600

Post by John Hasler

The electrolyte used in these [Japanese] capacitors

They were Chinese knockoffs of a Japanese design. However there was a
proprietary trick that the Chinese missed and so they made and sold a
vast number of caps that turned out to be defective.

The ones I saw were Japanese components in Japanese television cameras,
many of them broadcast. This was from the mid-90s onwards.

Post by John Hasler
Electrolytics usually have burst membranes intended to prevent them
from exploding when they evolve gas and they usually work on
low-voltage ones. They often work on high-voltage caps too, but
other times you can get a nice bang. Big motor starter capacitors
can be fun that way.

And silent leakage of copper-dissolving electrolyte can stop things
working just as well.

[John Hasler]

Post by Joe
The ones I saw were Japanese components in Japanese television
cameras, many of them broadcast. This was from the mid-90s onwards.

Japanese companies can get suckered by something that looks like a good
deal just like anyone else.

Post by Joe
And silent leakage of copper-dissolving electrolyte can stop things
working just as well.

Better. Explosions lead to more lawsuits, though.

[Mirko Parthey]

Post by Martin McCormick
change the boot sequence order from floppy-C:-CDROM to
CDROM-floppy-C: or CDROM-C:-floppy which prevents the hard drive
from grabbing the boot sequence each time.

There may be a function key that brings up a BIOS boot menu.
For Dell machines, it should be F12. Other candidates are F8 and F11.
Choosing an entry in the menu takes effect exactly once,
then the previous boot order is restored.
This feature might be useful to you if the entries are the same each
time and appear in the same order.

Post by Martin McCormick
I don't think the CMOS batteries are dead because the
clock still seems to keep time

Changing the battery is certainly worth trying.

The BIOS clock could still have the correct time despite a dead battery
because the Linux system gets the time from an NTP server and writes it
to the BIOS clock on shutdown. The clock then runs on ATX standby power
and will only lose its memory when the power cable is removed. The same
applies to the BIOS settings.

Regards,
Mirko

[Felix Miata]

Post by Mirko Parthey
The BIOS clock could still have the correct time despite a dead battery
because the Linux system gets the time from an NTP server and writes it
to the BIOS clock on shutdown. The clock then runs on ATX standby power
and will only lose its memory when the power cable is removed. The same
applies to the BIOS settings.

That's the logical expectation, but not the way it always works. Just hours ago
on a Dell SFF Optiplex 745 I found a dead battery. After installing a new one I
decided to remove it after first setting the BIOS clock and rebooting. Without
having removed the power cable, only removing the battery and no other changes,
on first subsequent boot it reported BIOS settings reset, and sure enough, clock
was reset to its build date just short of 10 years ago.

[David Niklas]

On Sun, 15 Jan 2017 23:04:14 +0000

Post by Brian
[A trip down Memory Lane snipped]
For those who may want to use Plop boot manager, it is far quicker and
more reliable to do so from GRUB on the hard disc. The Plop website has
details. Struggling with a pile of floppies is no fun and a floppy which
works today ivery likely not to do so next month.

Actually, I've had to transfer data from an old windowz to another
windowz system and I found that all the spare floppies I had lying around
contained correct file systems (and I'm assuming correct files too).
So, long term storage seems to work fine for floppies, but transfering
lots of data does not work so well (the floppies wear out).

Sincerely,
David