Post any computer-hardware related questions including discussions of obsolete
computers.

205 responses total.

What is likely to be the cause of the following problems with a 486DX2 66:
Will not reboot with Ctl-Alt-Del, but reboots randomly while running things
like edit.  After running defrag crashes and screen blinks, won't reboot.
Occasional GPF error blamed on emm386.exe.  
Jim is thinking of reseating the BIOS and RAM and cpu.  Other ideas?
None of this is easily repeatable except for failure to reboot when desired.

Bill L says this is a software problem having to do with emm386 and buffers
and memory.  I will try removing emm386 to see if it will boot then.
It also sometimes does not accept any keyboard input.

Remarking out the emm386 line in config.sys allows me to reboot.
Has anyone else used DRDOS with emm386.exe?  frame=none dpmi=off
dos=high,umb  No himem.sys. Kermit sends much smaller packets
with only 1M RAM accessible.  

Might the problem be some conflict between DRDOS and the cpu, which identifies
as SGS-Thompson ST486D?  (Is this a Cyrix cpu?)  Can we change the cpu to an
Intel or are they board dependent?

Hoping to give this computer to someone out of town on Sunday, with more than
1M usable.  (A viewer will only show 1/4 of the image with 1M).

        It's probably a busted flux capacitor.

aw DAMN!

Jim wants to try emm386.exe from MSDOS on this DRDOS computer.  I will report
the results.  I would like to simply try MSDOS on it.

There's a switch you can give EMM386 that fixes the reboot problem.  I don't
remember what it is, but it's in the help.

We are using dpmi=off and frame=none, are those relevant to the problem?
Thanks for the clue (and for ghostview).

AltBoot 

Specifies that EMM386 use an alternate handler to restart your
computer when you press Ctrl+Alt+Del. Use this switch only if your computer
stops responding or exhibits other unusual behavior when EMM386 is loaded
and you press Ctrl+Alt+Del. 

That's for the DOS 7 version, I don't know when that switch was added.

I tried this as /altboot at the end of the line.  Got a list of possible
things you can put on the emm386 line as a result.  Altboot is not one of
them.  DRDOS 7.03.  We are trying to figure out what range to exclude and may
do exclude= the areas where the video ram is.
Perhaps it is just altboot without the /?  I am lost.
We just excluded c000-c7ff,e000-f3ff  and it won't boot again.
Next thing to try is getting rid of DRDOS 7.03 and using MSDOS 6.22, which
I wanted to do in the first place.

(man.  10 years ago I used to have to know that kinda shit)

We did scandisk and the file structure is okay.  Replaced the DRDOS system
files and emm386.exe.  Ran DRDOS from a boot disk with files from a known
working computer.  None of this helps.  The computer works with DRDOS unless
we use emm386.exe with it.  The emm386.exe is from DRDOS 7.03.

So we made an MSDOS boot disk with emm386 and himem.sys in config.sys and on
the disk and booted with that.  Then tried to look at the c: drive and all
the file names are in Greek characters.  Repeated this a few times.
This means the computer is unusable with MSDOS, I think.

Is this an extreme case of cpu incompatibility?  It is a Thompson cpu, not
an Intel (cyric?) ST486D.  Bad cpu, bad BIOS?

Time to move the hard drive (which is 80M and has only one B on it when we
run scandisk) to a new box?  We can set it up with 3M RAM in an ATT.

I was just offered a used Pentium 200MHz with CD-ROM drive and sound card.
Maybe I can give my old ATT to Jim's sister instead of the oddball.

Redoing the MBR also did not help.

Why would MSDOS do worse than DRDOS?  If we boot from a DRDOS floppy disk we
can read the file names in c:, but not if we boot with MSDOS.

Jim ran McAfee virus checker.  Found no viruses but said
Read access to file C;\ XXXXXXX denied. - the XXXX being a bunch of
upper ascii characters.  Six files denied access.  This was run under MSDOS.

If we put the hard drive in another computer and the problem follows, it is
software, otherwise probably hardware.  More later this morning.

McAfee run under DRDOS could access all the files and found no viruses.

We are about to try a replacement controller card.  The modem (which is
plugged into this one) also has problems - won't hang up when expected to with
one program (Termin).   Back soon.

With the new controller card it won't boot - it gets stuck on 
512K memory in the video card and won't boot the rest of the way.
The controller card worked somewhere else.  Jim wonders if the board is set
wrong for its Cyrix cpu and will try an Intel cpu.

We forgot to turn the computer off.  In the morning it would not boot from
the hard drive.  We tried replacing cpu, video card, and controller card and
concluded it was a bad motherboard.  We recycled the computer and saved the
battery, video card, controller card, jumpers (white and red ones) and the
hard drive.  We moved the hard drive to our computer with 2 72-pin slots
(which we have 2M SIMMs for) and the new computer would not boot from the hard
drive.  It would boot from a DR-DOS floppy and also reboot (one less problem
than before) but when we booted from MS-DOS floppy it had more problems -
instead of displaying 3.5G and Greek file names (on an 80M HD) it could not
access C: at all.  We made this drive slave and an identical one master and
could boot from C: and when running MS-DOS from C:  then the original HD (from
the other computer) displayed file names as Greek letters again.

We have a bad hard drive, obviously.  Since DR-DOS can at least read that
drive though won't boot from it we are reformatting c:, scandisk, and will
attempt to transfer the files to it and scrap the hard drive.

The other computer was no great loss - 4x30 pin slots (4M max) and was missing
the covers for two large bays (we put in cardboard instead).  Too many
computers.

Jim points out that the newer (post 1990) hard drives (IDE) have part of the
controller incorporated in them and that was apparently affecting the rest
of the computer even when we did not boot from the HD.  We should have pulled
the cable off the HD before booting from floppy to test.

We are formatting c: /s /u /x.  

Comments on hard drives and controllers welcome - what different types of hard
drives are there and when did they start having the controller included?
MFM IDE RLL ESDI SCSI  (from Build Your Own 486 and Save a Bundle - back when
a 486DX cpu was $600 for 50MHz and $1000 for 66MHx DX2.  We have thousands
of dollars worth of cpus!   What does a cpu cost now?

No problem in copying from the bad HD using DR-DOS, but MS-DOS was useless
in that respect.  Jim has convinced me to switch to DR-DOS.

The problem seems to have followed the software.  New hard drive won't boot
now.  ???????  We had done an unconditional format.  MS-DOS can now read the
hard drive, it just won't boot from it.  We will convert this back to an
MS-DOS computer.  Stay tuned.

The computer (the one we did not recycle) with the second hard drive (the
first would probably also have worked) is doing fine as an MS-DOS computer.
I had begged Jim to set it up that way in the first place.  I am still using
two temporary computers because he wanted to set up the good ones with DR-DOS
and in the meantime I put together something that I could understand, about
three years ago.  I have two pentiums not in use because of DR-DOS. Jim just
left without saying goodbye after I asked him to please not set up any more
computers for my use with DR-DOS on them. One full day wasted.  The whole DOS
directory deleted off the hard drive and the files have to be put back on via
floppy disk.
        Now I get to learn to replace RAM unsupervised.

Hmmm I love to recycle and everything but is it really worth putting 20
hours of effort into a 486 worth maybe 20 bucks, when a Pentium 133 class
machine can be had for 50 bucks.  Even at minimum wage it seems like it
would make more sense to work 10 hours and buy the persons the Pentium.
Of course you should make sure the 486 is disposed of properly and doesn't
wind up in a land fill, but it seems like a 486 on the blink isn't worth
the hassle to try and fix.

re #18 CPUS are pretty cheap now you can get a Duron 1 ghz CPU that is
about 30 (?) times as fast as those 486s for about a hundred bucks.

We don't need anything more powerful than a speeding bullet just to access
mnet from Warren and display photos from our camera.  We are learning
something this way, at least.

I am suspecting that Jim may have formatted the hard drive with MS-DOS and
then put DR-DOS on it and they don't like each other.  He thinks he used
DR-DOS format but is not sure.

I learned three ways to install a controller card wrong (dont' push it in far
enough, put the cable to the floppy drive on only half the way, and put the
cable to the serial ports on so that it does not cover the end pin or
alternatively so it does not cover one of the two rows of pins.  Hard to see
what I am doing in there. I pull the plug first.

I have PCPlus and Kermit working on Com1 and am trying to convince Lynx to
also behave.  It worked once but then could not find the nameservers.  It was
working yesterday - same software, controller card and modem.  I think I may
have messed up a setting (mru, whatever that is, should not be set equall to
mss).  

We don't use a mouse so I put in only one com port this time.  Attached to
Com2 plug, it does not get recognized as com2.  I made it com1.  Lynx was
working before as Com2....

I'm not saying don't tinker I love to tinker, just tinker with something
more usable, at this point you could probably dumpster dive a Pentium
166 level system pretty easily.  Another place to look is the dumpster
outside U of M property disposition.

We have 2.5 pentiums from Tim Ryan that we have not found a use for.
They tend to run hotter and need fans.
I solved the computer problem by starting over with an ATT (from Tim, one of
about 50 he donated from Borders' discards) and putting the files on that.
Lynx runs fine there, just like on my office computer which is the same.

On the second 4M RAM computer (it won't recognize a second 1M SIMM, if I add
it I get 1 total instead of 4+1) with one serial port (Com1), controller card
where we don't know how to disable Com2, and sometimes internal modem on Com3
(with Com1 unplugged) I could run PCPlus or Kermit but Lynx will not dial
(unless I change the init string from one that worked with it before) and if
it does dial it wont' connect, and the one time out of 30 it did connect it
would not find the nameservers.  The nonconnect problem occurs with lynx with
all my computers but only half the time.  Something timing out?

Anyway, I have decided this computer will never work with lynx and we will
use it at the building site to draw CAD drawings of the porch.

The ATTS seem to always work (don't have the printer problem that we had with
three other computers and numerous controller cards, work with Lynx, read
floppy disks that other computers won't).  The Pentiums from Tim are ATT.
The only thing I would appreciate more speed for is dealing with images, and
I can stand to wait 20 sec once in a while.  (One program insists on
converting all 2-color images to 8-bit 256 color while I wait, before it will
display them).

Are other people using computers older than pentiums, and what for?

I guess I ought to retitle this item 'obsolete computer hardware'.

I agree with raven.  People have been throwing away 486's like yours
(except that they are in perfect working order) for 5 years.  It appears
that you have put hundreds of $$ worth of your time into a computer that
will have basically no value even if you fix it.

That being said, I suggest staying with the most generic hardware and
software as possible.  Why mess around with DR DOS?

Also the hardware you are working with is so old that it may be to your
disadvantage. Newer stuff is more common and often cheaper.  For example,
30 pin memory is hard to find and very expensive per megabyte. On the other
hand I just bought a 128 MB DIMM for $0 after rebate. 

I can honestly say I know of nobody using a sub-pentium computer for anything
work or personal.  For a while I was using a P5-60 as a router/firewall, but
I replaced it with a router that I bought for $50 at best buy, and have 
probably saved that much in power consumption (or will sooner or later).
And that is coming from a known cheapskate -- my array of computers at 
home ranges from P5-120 to a Celeron 300.

Most early computer hard drives had one-off interfaces.  That is, the
interface would be designed for that particular model of drive, and the
next generation model would have a completely different interface.
Typically there would be a separate controller, in some cases occupying
a complete 6' tall 19" rack, which would contain the electronics needed
to go from the disk interface, to whatever peripheral interface the
computer supplied.  Typically, the interface on these early 1-off
designs would include separate "control" and "data" busses.  The control
bus would include signals to select drives, heads, plus signals to step
the arm in or out to the desired cylinder.  The data cable would include
the actual data being send to/from the drive and associated timing
signals.

One typical peripheral interface was the IBM channel interface, which
used 2 really big fat cables "bus and tag" to connect between the
channel interface, and individual device "controllers".  The channel
interface proper was capable of doing DMA into main memory independently
of the CPU, and ran short channel "programs" that directed the follow of
data and commands to and from devices.

Several 3rd party disk makers realized that 1-off disk interfaces was
not the way to go, and banded together to create the SMD (storage module
disk) standard.  One advance in SMD was that instead of supplying
stepper motor pulses, the drive itself would move the arm to the proper
position; the interface included a parallel interface on which the
cylinder could be supplied.  The data cable only transfered one bit at a
time, and this became something of a problem as drive capacities went
from 5 megs of removable storage, to 600 or more megs of non-removable
storage.  SMD drives were particularly popular in the minicomputer
market.

The shugart associates SA1000 series was one of the first hard disks to
reach the microcomputer world.  The SA1004 was an 8" 8M drive, which
used a fairly primitive interface design.  Based on earlier floppy drive
interfaces, it was also very typical of the "one-off" designs of earlier
mainframe disk technology.  It used a 50 pin ``A'' cable for control
signals, plus a 20 pin ``B'' cable for data.  Typically the control
cable was bussed in common to all disk drives, while the data cable
would be individual cables going from the controller to each drive
separately.  While the SA1004 was popular in its own right, it was also
important as the starting point for 2 disk interfaces, the ST-506 & SCSI
interfaces.

SCSI originally started off as SASI, and it was in effect the
realization of mainframe disk architecture.  The channel interface of
the mainframe was called a "host adapter" in the microcomputer, and the
device controller became a separate card which went between the SCSI
interface and one or more physical disk drives.  As LSI logic became
cheaper and better, it became possible to incorporate the SCSI interface
directly on the disk.  The original SCSI interface used a single 50-pin
cable, and allowed for an 8-bit transfer path, and up to 7 controllers
(plus the host CPU) on the SCSI bus.

The ST-506 drive came out rather soon after the SA1004, and was the
first popular 5.25" drive.  It only held 5 M.  It used essentially the
same signalling standard that had been designed for the SA1004, but on
only 34 pins, because 5.25" didn't leave enough room on the back of the
drive for anything more.  Like the SA1004, but unlike virtually all
drives that came after it, the ST-506 had a stepper motor interface and
no on-board CPU; the controller was responsible for scheduling stepper
motor pulses.  The ST-412 and successor drives used minor modifications
to what became known as the "ST-506" bus standard; one of those
modifications was the notion of "buffered" seeks, which essentially
meant the controller fired off a bunch of pulses all at once and the
controller would accumulate those in a counter and use those to drive
the actual stepper logic.  The ST-412 had a microprocessor that was
dedicated to just performing those stepper motor calculations.

The IBM PC XT used the ST-412.  The XT controller was capable of doing
DMA to main memory, necessary because of the slow speed of an 8088 at
4.77 Mhz.  The controller also came with its own "disk bios", which
patched into the main bios and extended the disk "INT" interface to
support the hard disk.

The IBM AT came originally with a 20 M drive.  It also came with a new
disk controller.  The 286 in the AT was capable of doing "string" I/O,
which was much faster than DMA on the XT.  The new disk controller used
these "string" I/O instructions instead of DMA, so was both faster and
cheaper.  (A DMA design would have been faster yet, but more expensive).
The original SA1004 interface and later ST-412 allowed for up to 4
drives per interface controller, but on the AT, this was further limited
to just 2 drives.

Later on, IBM and several other companies introduced ESDI, which was
basically an improvement on ST-506/412.  It used the same 34+20 pin
cable scheme, but (I believe) replaced the buffered seek logic with an
actual binary number sent in parallel - much like SMD.  ESDI was never
really popular, and I think the reason is right around this time, the
market became flooded with much cheaper RLL encoded drives using the
earlier ST-506 style interface.

To explain RLL, we have to go back a bit and explain FM and MFM.  The
earliest drives stored data as a series of interleaved timing and data
pulses.  Each pulse had a separate leading & trailing edge, and data
pulses were only present if the associated data was a logic ``1''.  This
was easy to decode in hardware, but expensive in terms of storage.  The
first advance was MFM -- in which each signal *transition* was used for
data and timing alternately.  This could record twice the storage of FM,
at the cost of only a few more gates for the "data separator", the gizmo
that separated the data from the timing signals.  Most early drives,
including SMD and most of the early ST-506 drives, used this.  RLL was
another step in advance of this -- instead of using half the signal
transitions for timing, a different scheme was devised by which 7 data
bits could be stored in 9 signal transitions - using a table in I think
guaranteed at least 2 signal transitions in any possible group of 9
signal transitions.  This was another 50% improvement over MFM, at the
cost of only a slightly more complex data separator.  This scheme could
in fact work with drives originally designed to be MFM, at the cost of a
slightly more expensive RLL controller, which had a data separator that
was slightly fancier than MFM.

RLL and buffered seeks were but minor improvements in storage
optimization.  The drive makers knew that they could get even more
improvement in storage by other tricks, such as storing more sectors in
the outer tracks, but this would have broken the ST-506 interface in
ways that would not have been easy to fix in software.  Rather than do
this, the PC disk world decided to instead move most of the
functionality of the disk controller onto the drive, leaving only the
minimum behind necessary to drive the I/O bus.  The interface these
drives would present to the world would instead be based on the IBM AT,
and so this was sometimes called the ATA interface rather than IDE.
Like the earlier AT drives, this interface would only support 2 drives,
and programmed I/O generally using the "string" input/output
instructions of the CPU.

So is IDE the latest in drive technology?  Marcus, you never cease to amaze
me with the amount of trivia that you retain and explain.  Wish I understood
more of the words so I could follow the details.

Re the previous response, I have not spent any money on any computer other
than my first (1985, no expansion slots or hard drive so I bought a
daughterboard with two slots so I could add a hard drive, and also a Herc Plus
card and 256K RAM, and two printers.)  Everything else was given to us by
people who had upgraded.
It is very cheap to be five years behind the times.
We are not spending hundreds of dollars on trying to fix these computers, we
are getting a free education and not feeling bad when we accidentally recycle
something that was working.  Now I know some dialing software does not work
with some computers.  You ask the purpose of DR-DOS?  It is non-MS (free and
legally so), it does multitasking or at least task-swapping, and it is useful
for reading files off computers that you cannot get to work with MS-DOS
because you set them up with DR-DOS, so that you can copy them to a hard drive
that has MS-DOS so it will work.

I forgot, we did spend $2 for a part to repair one of the TTL monitors after
I messed it up by accidentally loading a Thai VGA font.  We have others but
Jim wanted to learn to fix a monitor.  I have a Cyrillic font for TTL.

IDE has gone through several generations.  There is EIDE, and
"UltraDMA", and several other buzzwords.  My understanding is that the
latest drives in the IDE technology line are capable of DMA but not of
overlapping seeks.  For one drive, this is fine; for a server with 8
disk drives, IDE is still hopelessly outclassed by SCSI.

        I still use this IBM PS/2 386 25mhz to M-net and Grex on a daily
basis.  It serves the function.  In fact, while reading Grex and eathing
lunch today, on this computer, the good computer was burning two CD-Rs.
        Of course, I do not know to Grex on the new computer.  I don't
know how to set it up, it's still phone connection, and I don't have 
any local freinds that will come over and help me with it, taking me
thru the motions.

To each his or her own.  Personally, I could not revert to 10 year old
hardware, dialup modems and text based applications.  When they were
state of the art they were so much better than what preceeded them
(ie pencil and paper) but by todays standards are down right primitive.
You might want to rescue a landfill bound Pentium-133 or thereabouts.

We have in our hostel very different hardware and software:
from 486SX(IMB,white construction) to AMD Athlon1.2 and all
specter of different software programs(Mandrake8.1,Windows2k,95,98,NT4.0,
CorelDraw10,Adobe Photoshop6.0,ACDsee4.0,MSVisualC++5.0).
All old machines works with MSDOS and use HIMEM driver. We have no any 
problems because we use HIMEM. Also u can use very powerfull PTSDOS,
russian DOS OS(it is also free) with different external programs.
PTSDOS also have FAT16 filesystem.

And what about MSDOS? How many it costs? Is that a problem to format
hard disk and set active partition? In ukraine all babyes can do that.
Can u? Have u ever seen Power quest Partition magic or FDISK?

My home network setup contains the following:
1 486DX2/66 running as firewall/router
1 Mac clone (PPC603e/160)
1 Pentium-100 (currently in flux; destined to be Linux test-bed)
1 Pentium Pro-166 (Win NT)
1 Pentium-133 laptop (Win98)
1 brand-new (just received on Friday) Athlon-950 being set up as a FreeBSD
server which will be externally visible (replacing the dedicated server I
rent)

How much of that do most people consider "obsolete"?

Where does one find PTSDOS?  And at what age do Ukrainian babies learn to type
nowadays?

        I've got some seriously obsolete hardware at home.  As in a DEC
Rainbow.  But it'd be pointless to try to upgrade the machine to the point
to which it could actually format its' own floppy disks.

I threw out all my 486 stuff once I had enough Pentium machines for my 
needs.  I don't regret it.  I was always having problems with crashes 
due to things like mis-set memory cache timing on the 486 motherboards, 
which Pentium boards handle automatically.  Getting Pentium hardware 
running has, for me at least, been a lot less grief.

I have a TRS-80 Model I sitting a few feet away from me.  The character 
generator has a stuck bit, and the monitor needs some refurbishing, but 
otherwise it seems ok.  It has 16kb RAM.  Very obsolete.

        A TRS-80 just oozes character even if its' generator is broken.

For those on a semi-budget, I was at Property Dispo this morning and noticed
a few iMacs being sold "as-is" for $100/ea, no keyboards.  Most were labelled
"bad CD drive".

- Backtalk version 1.3.30 - Copyright 1996-2006, Jan Wolter and Steve Weiss