31 Jan
2006
31 Jan
'06
2:56 p.m.
I revived an old 6Mhz PC AT today by removing the original Western Digital based
hard disk controller and replacing it with a newer, generic IDE controller.
The large boat-anchor class 5.25" full height hard disk had died, and without a
suitable replacement a small IDE drive was chosen. The rest of the machine is
original, including the EGA card with the extra memory daughterboard and the
full slot monochrome card.
While discussing this beast, the topic of putting a CD-ROM on it came up. (I
would never do such a thing except for giggles.) The topic of the CD-ROM led
to a throughput question - could this old monster even do it? As in, is the
ISA bus up to reading from the CD-ROM at a reasonable speed?
Well, let's go throught some quick math. 6 million clock cycles per second is a
lot, but when you consider that each transation on the bus takes 4 to 5 cycles,
and instructions take 4 to 5 cycles (average) on a 80285, the throughput might
not be so hot.
Forget the CD-ROM and let's just pretend it's an I/O device on the bus, like a
fast hard disk. Given a PC AT running at 6Mhz, what can reasonably be expected
for bus throughput? (Obviously this depends on the loop and the instructions
used.) I would have guess 500K or so per second, but now I'm second guessing
myself and I'll actually have to write a small benchmark to try it.
When writing to a hard disk, would this machine have used a tight processor
loop, or would it have used DMA? Under what circumstances would it use DMA to
transfer data to a hard disk?
And one last question .. Unlike the PC and XT, the AT BIOS handles hard drives.
It didn't blink when I removed the crusty WD based controller and replaced it
with a no-name WinBond based controller. Does the new IDE controller really
look that much like the old controller that the BIOS can't tell?
-Mike
31 Jan
31 Jan
3:27 p.m.
On 1/31/2006 at 2:56 PM mbbrutman-cctalk at brutman.com wrote:
While discussing this beast, the topic of putting a
CD-ROM on it came up.
(I
would never do such a thing except for giggles.) The
topic of the CD-ROM
led
to a throughput question - could this old monster even
do it? As in, is
the
ISA bus up to reading from the CD-ROM at a reasonable
speed?
It doesn't matter--all IDE CD-ROMs are very generously buffered. I've run
CD-ROMs (albeit SCSI) on 8-bit 4.77MHz XTs.
Forget the CD-ROM and let's just pretend it's
an I/O device on the bus,
like a fast hard disk. Given a PC AT running at 6Mhz, what can reasonably
be
expected for bus throughput? (Obviously this depends
on the loop and the
instructions used.) I would have guess 500K or so per second, but now I'm
second
guessing myself and I'll actually have to write a
small benchmark to try
it.
Depends on the handshaking, etc. used. Clearly, if you have to read status
before reading data, the throughput is going to be half of what you can do
if you can read status alone. IIRC, doing nothing but I/O reads, the
16-bit throughput on a 286 bus is about 4 MB/sec.
When writing to a hard disk, would this machine have
used a tight
processor
loop, or would it have used DMA? Under what
circumstances would it use
DMA to transfer data to a hard disk?
PC-XTs used DMA, but most MFM AT controllers used PIO. Given that the
controller also buffers data, it makes sense to get a sector read and then
just use an INSW to get the data. Faster than DMA. I do recall a few
bus-mastering 16 bit controllers, however.
And one last question .. Unlike the PC and XT, the AT
BIOS handles hard
drives. It didn't blink when I removed the crusty WD based controller and
replaced it with a no-name WinBond based controller. Does the new IDE
controller
really look that much like the old controller that the
BIOS can't tell?
Pretty much. The IDE understands more commands, such as IDENTIFY, but the
basic command set's pretty much a superset of the old MFM controllers--and
for that matter, the ESDI controllers (which do have an IDENTIFY command,
so they look even more like IDE). And remember that your disk calls on a
286 will still go through the BIOS extension ROM on the WinBond. So it
doesn't matter.
--Chuck
3:30 p.m.
mbbrutman-cctalk at brutman.com wrote:
While discussing this beast, the topic of putting a
CD-ROM on it came up. (I
would never do such a thing except for giggles.)
I've successfully used my Sound Blaster Pro + 1X CDROM drive to play
Sherlock Holmes (a video title, no less) on a 12MHz 286, so it is
possible. 1X is, as I'm sure you remember, 150KB/s.
As for giggles, it's really useful. Most 286s don't have 650MB of
storage :-)
Well, let's go throught some quick math. 6
million clock cycles per second is a
lot, but when you consider that each transation on the bus takes 4 to 5 cycles,
and instructions take 4 to 5 cycles (average) on a 80285, the throughput might
not be so hot.
Don't forget memory timing. Different boards/manufacturers did
different things.
When writing to a hard disk, would this machine have
used a tight processor
loop, or would it have used DMA? Under what circumstances would it use DMA to
transfer data to a hard disk?
I believe the AT and later used tight processor loops as they beat the
speed of DMA but I could wrong.
--
Jim Leonard (trixter at oldskool.org) http://www.oldskool.org/
Help our electronic games project: http://www.mobygames.com/
Or check out some trippy MindCandy at http://www.mindcandydvd.com/
A child borne of the home computer wars: http://trixter.wordpress.com/
3:58 p.m.
I revived an old 6Mhz PC AT today by removing the original Western Digital based
hard disk controller and replacing it with a newer, generic IDE controller.
The large boat-anchor class 5.25" full height hard disk had died, and without a
suitable replacement a small IDE drive was chosen. The rest of the machine is
original, including the EGA card with the extra memory daughterboard and the
full slot monochrome card.
While discussing this beast, the topic of putting a CD-ROM on it came up. (I
would never do such a thing except for giggles.) The topic of the CD-ROM led
Why not? It works.
to a throughput question - could this old monster even
do it? As in, is the
ISA bus up to reading from the CD-ROM at a reasonable speed?
Depends on what you mean by 'reasonable speed'. AFAIK most CD-ROMs have
internal buffers, and will read quite well even on a slow machine (you
are limited by the bus speed, so don't try anything that _needs_ high
throughput). I wouldn't want to try a CD burner though.
You don't want ot know about the CD-ROM drive I have on this machine (a
much-hacked PC/AT with most of the original bits). It's an external
Philips unit, and it looks like a converted audio CD player. Darn it, it
_is_ a converted audio CD player. Most of the bits, including the
mechanism (CDM-4), decoder chips, case, etc are the same as those of a
contemporary Philips CD-plauer. They added an ASIC to grab the data
stream after the decoder and send it to the PC, and re-programmed the
microcontroller to accept comamnds from the PC.
It's just as well I realised it was a modified CD player. The mains
switch failed mechanically (wouldn't latch on), and the service mnaual
for the CD-ROM drive (yes, I have it) didn't list it. Fortunately the
manual for the audio player did, I ordered that, and fortunately again it
was the right part.
When writing to a hard disk, would this machine have
used a tight processor
loop, or would it have used DMA? Under what circumstances would it use DMA to
transfer data to a hard disk?
The PC/AT hard disk contoroller and the mainboard BIOS do not use DMA.
And one last question .. Unlike the PC and XT, the AT BIOS handles hard drives.
It didn't blink when I removed the crusty WD based controller and replaced it
with a no-name WinBond based controller. Does the new IDE controller really
look that much like the old controller that the BIOS can't tell?
YEs, that was the idea behind IDE. It's upward compatible with the ST506
controller (WD2001 or something like that). I have an IDE card in this
PC/AT. It's just a few TTL buffers and a PAL as the address deocder. It
uses the mainboard BIOS, there is no extension ROM on that card.
-tony
4:12 p.m.
While
discussing this beast, the topic of putting
a CD-ROM on it came up. (I
would never do such a thing except for giggles.)
The topic of the CD-ROM led
Why not? It works. I have an IDE card in this
PC/AT. It's just a few TTL buffers and a PAL as the
address deocder. It
uses the mainboard BIOS, there is no extension ROM
on that card.
It would be nice to obtain a schematic of one of the
more primitive IDE cards *whistles*. And wouldn't
artwork be nice too ;). Would an 8-bit card take the
place of a 16-bit card in an AT? Presumably yes,
albeit slower I guess. The Acculogic sIDE/16 or
something has all discrete logic, save for a pal or
gal (or 2).
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
2 Feb
2 Feb
6:44 a.m.
Quoting Chris M <chrism3667 at yahoo.com>:
You're not saying a modern (internal) ATAPI
cd-rom
(or dvd-rom?) would operate in such a machine with an
IDE card? I guess I could just try it - had my 5170
out not hours ago to test my old 80188 based graphics
card (beeeeep beep beep :( ), but I don't know where
my IDE card went.
I think it would work. It just seems out of character for a machine that old.
But then again, loading everything via 1.2MB floppies was a nightmare yesterday
. my data transfer machine is a PCjr with 360K floppies, which made things
interesting. I'll probably put an IDE CD-ROM on it just to make it a little
more useful instead of adding a 3.5" drive. (Or at least see how far I can get
with that configuration before breaking down and adding the 3.5" drive.)
It would be nice to obtain a schematic of one of the
more primitive IDE cards *whistles*. And wouldn't
artwork be nice too ;). Would an 8-bit card take the
place of a 16-bit card in an AT? Presumably yes,
albeit slower I guess. The Acculogic sIDE/16 or
something has all discrete logic, save for a pal or
gal (or 2).
Yes, this would be interesting. I had never realized that IDE was a superset of
ST506. Now I'm thinking of all of the old 8 bit machines I have that could use
an IDE hard disk. I've been trying to graft SCSI cards onto them, but 8 bit
SCSI cards with BIOS aren't exactly common either.
9:46 a.m.
On 2/2/2006 at 6:44 AM mbbrutman-cctalk at brutman.com wrote:
Yes, this would be interesting. I had never realized
that IDE was a
superset of ST506. Now I'm thinking of all of the old 8 bit machines I
have
that
could use an IDE hard disk. I've been trying to
graft SCSI cards onto
them, but 8
bit SCSI cards with BIOS aren't exactly common
either.
Careful--most recent IDE drives are not capable of 8-bit data transfers.
Commands and status are read and written from 8 bit registers, but the data
transfer part of the operation is 16 bit. There IS an 8-bit transfer
capability defined by the ATA standard, but it was deprecated in ATA-3 and
most drives don't handle it. There may be some old controllers around
that can do the translation, but ATA on 8 bit PC-XT was pretty rare.
However, it'd be easy to prototype up a controller. Alternatively, you
could do much of what you wanted in software--just ignore half the sector
data and read 2 sectors when one was requested. All in all, it'd probably
STILL be faster than the old 5.25 MFM drives.
Cheers,
Chuck
1:19 p.m.
It would be
nice to obtain a schematic of one of the
more primitive IDE cards *whistles*. And wouldn't
artwork be nice too ;). Would an 8-bit card take the
place of a 16-bit card in an AT? Presumably yes,
albeit slower I guess. The Acculogic sIDE/16 or
something has all discrete logic, save for a pal or
gal (or 2).
Yes, this would be interesting. I had never realized that IDE was a superset of
ST506. Now I'm thinking of all of the old 8 bit machines I have that could use
an IDE hard disk. I've been trying to graft SCSI
cards onto them, but 8 bit
SCSI cards with BIOS aren't exactly common either.
You will have a lot of fun (for suitable values of fun) trying to get
most old hard disk controllers or their hosts talking to IDE drives. Just
because an IDE drive looks the same to the host as one particular ST506
controller doesn't mean that all ST506 controllers look the same. They don't.
-tony
1:48 p.m.
On 2/2/2006 at 9:19 PM ard at p850ug1.demon.co.uk wrote:
ST506 is a low-level, raw, interface. What you see on
the interface
connectoer is essentially the pulse stream to/from the head. It's up to
the cotnroller to turn that into the user data bytes/words.
A small nit--and an itch when I hear folks talking about the "ST506
interface" on a PC.
I don't think a stock PC ever talked to an ST506 for the simple reason that
the ST506 lacks buffered seek. AFAIK, every PC controller shipped with
BIOS that assumed buffered seek. You could reprogram most (but not all)
controllers to deliver the timed(3 msec?) seek pulses that the 506
required, but it wasn't stock.
We should really be talking about the "ST412 Interface", no?
I think the original poster was talking about the software compatibility.
While an IDE drive will respond to commands intended for an AT-style
(WD1003?) controller, such is not the case for the old XT controller, which
is close, but still a different animal.
However, if we're talking 8 bit PCs, it doesn't really matter too much, as
the INT 13 interface is carried on the controller's BIOS extension
ROM--which is what you need with an XT-class system.
Cheers,
Chuck
1:57 p.m.
On Thursday 02 February 2006 04:19 pm, Tony Duell wrote:
It's not the _interface_ (as in hardware) that's got parts in common, it's
the command set. Which is one of the reasons we keep bumping into size
limitations.
<...>
Yes, this
would be interesting. I had never realized that IDE was a
superset of ST506. Now I'm thinking of all of the old 8 bit machines I
have that could use
It's not. IDE and ST506 are very different interfaces.
IDE ias a formatted interface. You transfer nice 8-bit bytes and 16 bit
words to the drive. The drive takes care of turning them into a suitable
pulse train to send ot thehead, etc.
ST506 is a low-level, raw, interface. What you see on the interface
connectoer is essentially the pulse stream to/from the head. It's up to
the cotnroller to turn that into the user data bytes/words.
The origianl PC/AT controller talked to an ST506 drive. But of course it
also had a host-side interface -- it plugged into an ISA slot and
appeared as a number of I/O ports to the 286 (or whatever) processor. The
controller card did the translation betweenn the user bytes/words and the
pulse stream on the ST506 interface connector.
Now you could also split the system up in a different way. Specifically,
move most of the controller over to the drive. The 'controller card' now
becomes little more than an address decoder which lets the host processor
talk to I/O port registers physically located in the drive. To the host,
it looks the same as the old ST506 controller (you send the same bytes to
the same ports to read a particular sector). Inside the drive, who knows
what goes on.
That's IDE. Just because an IDE drive looks the same to the host
as one particular ST506
controller doesn't mean that all ST506 controllers look the same. They
don't.
The way it looks to the host is exactly the point. And how different are
they? The main thing I remember having to do with those early HD cards is to
use the BIOS routines that were on the card (in XT-class stuff) to format the
drive. AT-class stuff and later didn't have any BIOS on the cards, the MB
BIOS took care of that, so there must have been a lot of it which was the
same.
--
Member of the toughest, meanest, deadliest, most unrelenting -- and
ablest -- form of life in this section of space, a critter that can
be killed but can't be tamed. --Robert A. Heinlein, "The Puppet Masters"
-
Information is more dangerous than cannon to a society ruled by lies. --James
M Dakin
10 Apr
10 Apr
9:16 a.m.
On Thu, 2006-02-02 at 21:19 +0000, Tony Duell wrote:
ST506 is a low-level, raw, interface. What you see on
the interface
connectoer is essentially the pulse stream to/from the head. It's up to
the cotnroller to turn that into the user data bytes/words.
So, and I do know this
is perverse :) - you could actually record, for
example, audio onto an ST hard drive, and play it back? Of course, it'd
just be a toy, but...
One thing I've done previously on a dead IDE drive is to hook a Hi-fi
amp to the voice coil actuator, and have listened to surprisingly
audible audio with that.
-toresbe :)
13 Apr
13 Apr
1:07 p.m.
On Thu, 2006-02-02 at 21:19 +0000, Tony Duell wrote:
Not easily. You don't, of course, get the raw head connections at the
interface connecotr, there's a write (recoding) driver and a read
(playback) amplifier in the drive. These circuits are strictly digital
(the interface signals are actually at RS422 levels IIRC). And there are
resrictions on the frequency of the pulses you can record (both minimum
and maximum from what I rememebr).
And of course, sioce the drive rotates at something like 3000rpm, you'd
only get a recording time of 1/50 of a second per track.
I do have a video hard disk in my collection. This is a head-per-track
hard disk rotating at 3000 rpm, such that 1 frame of TV video can be
recorded on each track. It uses FM (analogue) modulation to do this.
There are 3 indentical read/write circuits to handle RGB colour.
I am told that this sort of disk was oringinally designed for TV action
replays. The one I have has a DAC on the input to the FM modulator, with
a Unibus interface (via a DR11-B) on the front of that.
-tony
ST506 is a low-level, raw, interface. What you
see on the interface
connectoer is essentially the pulse stream to/from the head. It's up to
the cotnroller to turn that into the user data bytes/words.
So, and I do know this
is perverse :) - you could actually record, for
example, audio onto an ST hard drive, and play it back? Of course, it'd
just be a toy, but... 
1:27 p.m.
At 03:07 PM 4/13/2006, Tony Duell wrote:
I do have a video hard disk in my collection. This is a
head-per-track
hard disk rotating at 3000 rpm, such that 1 frame of TV video can be
recorded on each track. It uses FM (analogue) modulation to do this.
I saw a demo of this on an ordinary hard drive's mechanicals
in the late 80s. A friend recorded and played-back perhaps
ten or fifteen seconds of video on a (I think) 40 meg drive.
It was purely analog, not digital. (Amigas were early users
of IDE drives, so it might've been IDE, although that doesn't
really matter in this case, does it?) It was very futuristic at
the time, considering that comparable pro systems cost more
than a house. He soon developed the Amiga's Video Toaster and
even later, the Flyer, a hard-disk-based recording and editing
system.
- John
4:19 p.m.
John Foust wrote:
At 03:07 PM 4/13/2006, Tony Duell wrote:
<snip>
A friend recorded and played-back perhaps
ten or fifteen seconds of video on a (I think) 40 meg drive.
It was purely analog, not digital. (Amigas were early users
of IDE drives, so it might've been IDE, although that doesn't
really matter in this case, does it?)
I'm not aware of an analog way to access IDE, the interface is similar
to SCSI. The
SMD interface, Trident, or older 5440 drives had interfaces that
minicomputers had
to supply data separators and encoders for, and were able to record in
an "analog"
mode.
The object of the IDE and SCSI, and the earlier unsuccessful ANSI and
maybe SA1000
and SA4000 were to move the data separator and so forth into the drive.
The main
thing that was analog on any of these is that one could design an
interface which recorded
with different clock and data schemes, and could implement something
like PCM or
other phase modulation. The response was not really analog by any
means, but was
heavily tilted for digital, that is to get clear "true" and "false"
type
data to and
from the disk. If you had an interface to a media that
allowed for
analog level
responses, then you ran into problems with the signal moving up and down
with
less than desirable response, and therefore might get errors.
Hope this makes sense. I was involved in interfacing to a Western
Dynex, and
to some Trident and CDC drives, where other than mini computers were
using the media.
One of the earliest stop motion recorders used in broadcast used a
single drive and
a 5440 type head and positioner to record a few seconds of video, and could
be stopped and replayed easily. It used a different sort of interface
to get the
video to the media than a normal disk drive did, but the ones I saw
actually used
the same heads and media and positioners that our drives did (10mb capacity
on 4 heads, or 2.5mb / head).
Jim
4:35 p.m.
At 06:19 PM 4/13/2006, jim stephens wrote:
I'm not aware of an analog way to access IDE, the
interface is similar to SCSI. The
SMD interface, Trident, or older 5440 drives had interfaces that minicomputers had
to supply data separators and encoders for, and were able to record in an
"analog"
mode.
This fellow had ripped apart the drive quite severely. It wasn't a
question of squeezing analog in through the standard interface.
- John
4:50 p.m.
John Foust wrote:
At 06:19 PM 4/13/2006, jim stephens wrote:
This fellow had ripped apart the drive quite severely. It wasn't a
question of squeezing analog in through the standard interface.
- John
I'd love to hear that story for sure then. pretty interesting to do
that w/o really
makeing the logic in the drive really unhappy.
Jim
2:59 p.m.
Tony Duell wrote:
I do have a video hard disk in my collection. This is a head-per-track
hard disk rotating at 3000 rpm, such that 1 frame of TV video can be
recorded on each track. It uses FM (analogue) modulation to do this.
There are 3 indentical read/write circuits to handle RGB colour.
Back in the old days they were also used for computer animation.
You could record one frame at a time. At that time an 11/780 was the
single user workstation of choice and a single frame could take a while,
depending on what algorithms you were using. At least it had floating
point.
Once you (finally) got all the frames recorded you could see the result
in real time. Very nice for pencil tests. But rather expensive.
I am still floored everytime I watch someone playing WOW. So many
polygons, and shaded too.
-brad
11:37 p.m.
Tore S Bekkedal wrote:
On Thu, 2006-02-02 at 21:19 +0000, Tony Duell wrote:
Probably. Since for digital audio your data rate will be a lot lower
than normal IDE transfer speeds, you've got loads of time to get the
data on and off the disk. How much quicker is an IDE hard disk than a
single-speed CD-ROM?
Now, here you're talking about ST drives, which produce a string of
pulses. If you could read and write pulses at around 40kHz, and
pulse-width modulated them with the level of your signal (think 40kHz
sawtooth waves and comparators here) then you could do it. Add a bit of
logic to step the head on every rotation.
Gordon.
ST506 is a low-level, raw, interface. What you
see on the interface
connectoer is essentially the pulse stream to/from the head. It's up to
the cotnroller to turn that into the user data bytes/words.
So, and I do know this
is perverse :) - you could actually record, for
example, audio onto an ST hard drive, and play it back? Of course, it'd
just be a toy, but... 
14 Apr
14 Apr
3:09 a.m.
On Mon, 10 Apr 2006 18:16:33 +0200
Tore S Bekkedal <toresbe at ifi.uio.no> wrote:
One thing I've done previously on a dead IDE drive
is to hook a Hi-fi
amp to the voice coil actuator, and have listened to surprisingly
audible audio with that.
I have done this too. In addition I mounted a mirror on
the head and
pointed a laser to the mirror. Those voice coils are really good as
laser deflector. Gave me nice laser patterns on the wall.
--
tsch??,
Jochen
Homepage: http://www.unixag-kl.fh-kl.de/~jkunz/
2 Feb
2 Feb
11:18 a.m.
It's no news that cd-roms could operate in older
machines. I remember a friend with a Tandy 1000
something years ago who had one hooked up (can't
remember if it was internal or external though). And
he could listen to music while he was doing
*presumably* something useful with the machine.
Most older CD-ROM drives had built-in DACs and thus an audio output. You
could use that to play a music CD without using any of the host
computer's CPU cycles at all (the external Philips I have has
play/stop/skip track controls on the front, it can be used as a simple
audio CD player without connecting it to a computer).
There was often an audio connector that you linked to your soundcard,
but that just fed into the mixer stage of the soundcard. Again no host
CPU cyclers were needed.
It would be nice to obtain a schematic of one of the
more primitive IDE cards *whistles*. And wouldn't
I did trace out the schematics of the one in this machine. To be honest,
it was totally trivial. A few buffers and a PAL. I would thoink you could
design one by reading the IDE spec/
artwork be nice too ;). Would an 8-bit card take the
place of a 16-bit card in an AT? Presumably yes,
albeit slower I guess. The Acculogic sIDE/16 or
Most IDE drives expect a 16 bit data bus, at least for the data register.
The 16 data lines from the drive are buffered and fed to the 16 data
lines on the ISA slot.
-tony
31 Jan
31 Jan
4:44 p.m.
On 1/31/2006 at 11:58 PM ard at p850ug1.demon.co.uk wrote:
YEs, that was the idea behind IDE. It's upward
compatible with the ST506
controller (WD2001 or something like that). I have an IDE card in this
PC/AT. It's just a few TTL buffers and a PAL as the address deocder. It
uses the mainboard BIOS, there is no extension ROM on that card.
...and ATAPI is sideways-compatible with SCSI. :)
5:02 p.m.
On 1/31/2006 at 11:58 PM ard at p850ug1.demon.co.uk wrote:
You don't want ot know about the CD-ROM drive I
have on this machine (a
much-hacked PC/AT with most of the original bits). It's an external
Philips unit, and it looks like a converted audio CD player. Darn it, it
_is_ a converted audio CD player. Most of the bits, including the
mechanism (CDM-4), decoder chips, case, etc are the same as those of a
contemporary Philips CD-plauer. They added an ASIC to grab the data
stream after the decoder and send it to the PC, and re-programmed the
microcontroller to accept comamnds from the PC.
I used to have a couple NEC CD-ROM drives like that--still had all of the
holes in the case for plugging in the cigarette-lighter adapter and all.
They'd fitted it with a DB-25 for a SCSI interface. 1X speed.
What can I say? They were cheap.
Cheers,
Chuck
5:21 p.m.
I revived an old 6Mhz PC AT today by removing the
original Western Digital based
hard disk controller and replacing it with a newer, generic IDE controller.
While discussing this beast, the topic of putting a
CD-ROM on it came up. (I
would never do such a thing except for giggles.) The topic of the CD-ROM led
to a throughput question - could this old monster even do it? As in, is the
ISA bus up to reading from the CD-ROM at a reasonable speed?
Not a problem. Your ISA bus is running at 6 MHz (167 ns cycle) and
can do 16 bit transfers every couple four (667 ns) on a good day with
the wind at its back. The IDE specification for PIO mode 0 is really
spec'd for an 8 MHz ISA bus so it specifies 600 ns between transfers.
But ISTR that it's spec'd as a minimum time between transfers of 600
ns. The maximum can be much longer, and often is since PIO mode max
rates only happen if the processor isn't doing anything else. That's
while IDE drives always have a track buffer. So your 6MHz ISA bus
will max out at 3 MB/s (using 100% of CPU for I/O).
The original multimedia PC systems were suggested to use <15% CPU on a
single speed CD (153,600 bytes/s). On you system I'd estimate that a
single speed CD read would take 2.5% of the CPU time if the driver is
written well. You would theoretically max out your system with a 40X
CD-ROM drive, but I'd be very surprised if you can keep up with a 24X
drive.
When writing to a hard disk, would this machine have
used a tight processor
loop, or would it have used DMA? Under what circumstances would it use DMA to
transfer data to a hard disk?
The default BIOS would use polled I/O. CD-ROM drivers almost always
use polled I/O, because most IDE cards don't support DMA because a lot
of the drives don't support the ATA DMA modes. If by chance you do
have a DMA capable IDE card you will need to install a driver to use
it.
And one last question .. Unlike the PC and XT, the AT
BIOS handles hard drives.
It didn't blink when I removed the crusty WD based controller and replaced it
with a no-name WinBond based controller. Does the new IDE controller really
look that much like the old controller that the BIOS can't tell?
The BIOS shouldn't know the difference between the controllers. Just
as long as the I/O ports are in the right place (and they are designed
to be) there's no difference. The smarts on an IDE drive emulate the
old AT disk controller's communications with the BIOS. You will
probably have problems getting the BIOS to recognize the proper size
of any disks you install. You may need a disk manager in order to get
things running properly.
5:46 p.m.
Good, I'm learning a bit here ..
The superIO controller card I stuck in there did not look like it had a
BIOS ROM extension, so I'm assuming that the chipset emulated whatever
the hard disk interface was on the AT, and that the built-in BIOS on the
AT is still driving everything just as though it was the original
controller. It's hard to believe it is this simple, but I'm not going
to argue. I never considered IDE elegant until now ..
Chuck - you mentioned that the superio controller card would have a BIOS
extension. Are you certain? One card that I tried definitely did, as
it had the intelligence to handle larger drives. The particular card
that I used does not have any options for setting a BIOS ROM address,
doesn't put a splash message up at boot time, and doesn't have anything
that looks like an EPROM on it. (Even though it looks quite elderly.)
Eric Korpela's note indicates that it behaves more like my description
above ...
Another part of the equation is the CPU. If it's doing the I/O doing a
tight loop, it has to be fetching instructions from memory for the
instruction loop as well as doing the I/O to the bus.
I was thinking that if each instruction takes a few cycles that even the
tightest of loops would waste a lot of cycles, but I found a gem in the
286 user's guide - REP. Apparently you can use REP INx and REP OUTx
instructions to generate a tight loop that doesn't require subsequent
instruction fetches until the loop ends. So that would allow a 286 to
push the bus much harder than an 8088/8086 class machine would. (The
8088/8086 would have to keep fetching instructions, which would suck.)
Mike
6:11 p.m.
On 1/31/2006 at 7:46 PM Michael B. Brutman wrote:
Chuck - you mentioned that the superio controller card
would have a BIOS
extension. Are you certain? One card that I tried definitely did, as
it had the intelligence to handle larger drives.
Some controller cards have BIOS extensions, some don't. Most IDE ones that
were built to handle larger IDE drives do; garden-variety ones don't.
You may get in a bind with a larger IDE drive as the original AT was pretty
limited in its support for large drives.
I was thinking that if each instruction takes a few
cycles that even the
tightest of loops would waste a lot of cycles, but I found a gem in the
286 user's guide - REP. Apparently you can use REP INx and REP OUTx
instructions to generate a tight loop that doesn't require subsequent
instruction fetches until the loop ends. So that would allow a 286 to
push the bus much harder than an 8088/8086 class machine would. (The
8088/8086 would have to keep fetching instructions, which would suck.)
As I mentioned, the operative instructions are INS and OUTS--work just like
STOS and LODS, except the input is through port (DX). Even on a
Pentium-class machine, the original AT 16-bit bus timings are observed for
ISA devices, so at some point, processor speed just doesn't matter any
more.
Cheers,
Chuck
9:33 p.m.
Let me add a note on the BIOS-less IDE card for your PC-AT.
The biggest problem that I see is that the largest geometry that your PC-AT
BIOS tables allow is 1024 cylinders, 17 sectors and 16 heads, which gives
you a grand total of 139 MB total. While positively huge for the PC-AT
era, it's pretty puny when it comes to IDE drives.
An IDE drive in CHS mode can address 16384 cylinders, 63 sectors on 16
heads for a total of 8GB. While not huge by today's standards, it was
positively enormous for the time.
So, you've got two alternatives. Add a BIOS extension (usually on a disk
controller card, but doesn't have to be) to plug the necessary code in
during boot, or use an "overlay" (which is really the wrong word) which
loads some driver code from the first track of your hard drive, then plugs
itself in as sort of a TSR program (actually locates itself in high memory,
then adjusts the amount of available base RAM downwards) to intercept disk
requests and map them to the expanded geometry.
The "overlay" mode (such as that provided by "Disk Manager") has its
limitations, but should be okay on a 286. But I personally like the BIOS
ROM extension myself.
Cheers,
Chuck
1 Feb
1 Feb
7:11 a.m.
I like the BIOS ROM extension better myself, but what I grabbed is what I
grabbed ... It occurred to me last night that I had not changed the CMOS
settings for the drive, so the nice 600+MB IDE drive that I stuck in there was
being treated like the old full-height monster that it replaced.
Today's project will be to put a Promise EIDE MAX in a spare slot - that will
give it the BIOS extension that it needs to recognize larger drives. I'm also
going to run DIAGs on the beast and work out some other problems.
It's been a fun project so far. This AT is in a rack-mount case which is even
more overkill than the original AT case. The funny thing is that somebody
ditched the presumably industrial rack-mount coverings and built a plexiglass
cover for the case that gives you see-through views of the top, front, and two
sides. Horrible for inteference, but really fun to look at.
Besides the 5154 EGA display I have an original monochrome display, but it
doesn't work at all. I'm not sure that it is even getting power. I have the
Sams schematics for it, so when I get some free time I'll open it and start
looking for simple things like blown fuses.
I found this machine where I work and it hasn't come home yet, and may never due
to work policies. So whatever I get it to do is a bonus. It'll complement the
PCjr I keep in the office nicely.
9:19 a.m.
On 2/1/2006 at 7:11 AM mbbrutman-cctalk at brutman.com wrote:
I like the BIOS ROM extension better myself, but what I
grabbed is what I
grabbed ... It occurred to me last night that I had not changed the CMOS
settings for the drive, so the nice 600+MB IDE drive that I stuck in there
was being treated like the old full-height monster that it replaced.
Today's project will be to put a Promise EIDE MAX in a spare slot - that
will give it the BIOS extension that it needs to recognize larger drives.
I'm
also going to run DIAGs on the beast and work out some
other problems.
I remember well getting my first 1224 cylinder MFM drive (Miniscribe, I
think) to go with my Faraday ATease 286 mobo and realizing that the BIOS
wouldn't get me past the 1024 cylinder mark, leaving almost 20% of the
drive unusable. This was in the bad old MS-DOS days, so I eventually just
wrote a DOS device driver that gave me a partition with the remaining
capacity. With a WD-1006 RLL card, I was looking at a whopping 250 MB!
Seems kind of strange now that one can buy little bitty flash cartridges
with a gig or more on them that are smaller than the power connector on the
drive.
Remember, you can still run OS/2 1.x on that AT box.
I think a lot of very serviceable older hardware goes into the landfill
because of performance issues that aren't necessarily related to preocessor
speed. Often, just increasing the memory and adding a faster (ATA6) hard
disk controller can make a huge difference on older Pentium I and II
machines. And it's these older machines that still have a few ISA slots
for older peripherals. I've just finished setting up a 200 MHz PII box
this way with Win98, RH9 and Win2K. Performance for most mundame
programmer-type tasks is very adequate.
Cheers,
Chuck
9:48 a.m.
In article <200602010919510439.4D557645 at 10.0.0.252>,
"Chuck Guzis" <cclist at sydex.com> writes:
I think a lot of very serviceable older hardware goes
into the landfill
because of performance issues that aren't necessarily related to preocessor
speed. Often, just increasing the memory and adding a faster (ATA6) hard
disk controller can make a huge difference on older Pentium I and II
machines. And it's these older machines that still have a few ISA slots
for older peripherals. I've just finished setting up a 200 MHz PII box
this way with Win98, RH9 and Win2K. Performance for most mundame
programmer-type tasks is very adequate.
I think I found a good home for the remaining Pentium/Pentium II
machines they unloaded here at work. Yay!
--
"The Direct3D Graphics Pipeline"-- code samples, sample chapter, FAQ:
<http://www.xmission.com/~legalize/book/>
Pilgrimage: Utah's annual demoparty
<http://pilgrimage.scene.org>
12:10 p.m.
I think a lot
of very serviceable older hardware goes into the landfill
...
I think I found a good home ...
5:17 p.m.
Chuck Guzis wrote:
I think a lot of very serviceable older hardware goes
into the landfill
because of performance issues that aren't necessarily related to preocessor
speed. Often, just increasing the memory and adding a faster (ATA6) hard
disk controller can make a huge difference on older Pentium I and II
machines. And it's these older machines that still have a few ISA slots
for older peripherals. I've just finished setting up a 200 MHz PII box
this way with Win98, RH9 and Win2K. Performance for most mundame
programmer-type tasks is very adequate.
Seconded. Most Pentium Pro and later machines will take 256MB of RAM,
which greatly increases what they can run (although faster I/O helps too).
--
Jim Leonard (trixter at oldskool.org) http://www.oldskool.org/
Help our electronic games project: http://www.mobygames.com/
Or check out some trippy MindCandy at http://www.mindcandydvd.com/
A child borne of the home computer wars: http://trixter.wordpress.com/
31 Jan
31 Jan
4:23 p.m.
mbbrutman-cctalk at brutman.com wrote:
And one last question .. Unlike the PC and XT, the AT
BIOS handles hard drives.
It didn't blink when I removed the crusty WD based controller and replaced it
with a no-name WinBond based controller. Does the new IDE controller really
look that much like the old controller that the BIOS can't tell?
Well, consider the IDE hack for the PC/RT. As long as the IDE disk
honors all the ESDI commands (translate as "if it's a Seagate Medalist
<2GB"), you can replace the FDC/ESDI adapter in an IBM PC/RT with an
FDC/IDE card and the system firmware doesn't even notice.
Doc
6597
days inactive
6670
days old
31 comments
16 participants
participants (16)
-
allain＠panix.com -
ard＠p850ug1.demon.co.uk -
brad＠heeltoe.com -
cclist＠sydex.com -
chrism3667＠yahoo.com -
doc＠mdrconsult.com -
gordonjcp＠gjcp.net -
jfoust＠threedee.com -
jkunz＠unixag-kl.fh-kl.de -
jwstephens＠msm.umr.edu -
korpela＠ssl.berkeley.edu -
legalize＠xmission.com -
mbbrutman-cctalk＠brutman.com -
rtellason＠blazenet.net -
toresbe＠ifi.uio.no -
trixter＠oldskool.org