8086 powered IBM PC compatible
XT.8086
compatible
requirement:
1. using toroidal transformer as power supply instead of SMPS
2. using SRAM instead of DRAM
3. using super I/O chip
this is to replace aging 8088 system, and also used to teach pure 16
bit design.
thank you
petrus bitbyter
"XT.8086" <XT....@OVI.COM> schreef in bericht
news:929382f3-65d7-4921...@db5g2000vbb.googlegroups.com...
I'm sure it can be done. Even I can do it. I'm as sure it will become too
expensive.
The power supply will become too hot and too heavy. You may need a custom
designed toroid or more then one standard type.
It's not clear what super I/O chip you're talking about but I suppose it
will need at least a four layer PCB. Besides those chips are seldom
available in low volume.
You will need a BIOS. There may be an existing one for free. Otherwise it
may become more expensive then the hardware.
There are a lot of decisions to make before you can even start. About speed,
video, types and numbers of I/O channels, type and number of expansion slots
and so on.
So my guess - a wild guess I admit - is a pricetag of several thousend bucks
for a working prototype.
petrus bitbyter
Jon Kirwan
I'm guessing this is homework. The "to teach pure 16 bit
design" is a bit of a give-away. Chances are, that is how
the homework was sold to the students. The SRAM not DRAM
requirement is because the prof doesn't want to grade DRAM
designs -- pain in the ass. Super I/O is a all-in-one that
includes the 8259 interrupt controller and some standard
serial and parallel port items along with floppy control,
that got included into the South Bridge when PCI came along.
Again, it simplifies the student's work and the prof's review
time. I don't the BIOS is required here. Just a presumption
of one. It's school time, again.
Jon
Jon Kirwan
> I don't the BIOS is required here.
I don't _think_ the BIOS is required here.
XT.8086
> On Mon, 3 Oct 2011 10:27:52 +0200, "petrus bitbyter"
>
>
>
>
> >"XT.8086" <XT.8...@OVI.COM> schreef in bericht
thing, powered by single AA battery.
XT.8086
wrote:
> "XT.8086" <XT.8...@OVI.COM> schreef in berichtnews:929382f3-65d7-4921...@db5g2000vbb.googlegroups.com...
It would be mostly using solid state harddisk so it will conserve
power, and floppy drive is mounted only on the unit for teaching
purpose .
Do you really think it will exceed 2 ampere, give me your somewhat
detailed estimate.
Video type would be 8-bit CGA or 8-bit VGA, we allready have plenty
here and I want to reuse them (these one could suck up quiet alot of
ampere I guess)
I want just one each for 16-bit and 8-bit expansion slot.
What kind of BIOS is free which enable us to use existing IBM PCDOS
3.0
What do you suggest if Super I/O chip is not easily obtained.
thanks.
BTW I would like to know more of your company if you own one.
Winston
> On Oct 3, 5:08 pm, Jon Kirwan<j...@infinitefactors.org> wrote:
(...)
>> It's school time, again.
>>
>> Jon
>
> wooowww your so smart. ok this is your assignment, design similar
> thing, powered by single AA battery.
You got it in one try, Jon.
:)
--Winston
Nico Coesel
>On Oct 3, 3:27=A0pm, "petrus bitbyter" <petrus.bitby...@hotmail.com>
>wrote:
>> "XT.8086" <XT.8...@OVI.COM> schreef in berichtnews:929382f3-65d7-4921-a96=
>a-88f32...@db5g2000vbb.googlegroups.com...
>> > Could someone design for me a 8086 powered computer which is IBM
>> > compatible
>>
>> > requirement:
>>
>
>
>What kind of BIOS is free which enable us to use existing IBM PCDOS
>3.0
>
>What do you suggest if Super I/O chip is not easily obtained.
>
>thanks.
consumption would be small. IIRC I have a generic XT Bios source
somewhere which can be assembled into a working BIOS using MASM. About
15 years ago I used a modified XT as a simple means to control home
brew circuits.
--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------
tomcee
The HP200LX design may be of interest to you. It is a bit dated, but
it was a IBM compatible PC (NEC V70) with a PCMCIA slot that was
handheld and operated on 2 AA batteries.
TomC
Vladimir Vassilevsky
XT.8086 wrote:
> Could someone design for me a 8086 powered computer which is IBM
> compatible
Sure can. Do you have money?
>
> requirement:
>
> 1. using toroidal transformer as power supply instead of SMPS
> 2. using SRAM instead of DRAM
> 3. using super I/O chip
Long while ago a friend of mine built PC XT 8086 on the breadboard from
discrete parts. That board included I/O and CGA/HGS video. Worked fine.
> this is to replace aging 8088 system, and also used to teach pure 16
> bit design.
For any practical purpose it would be easier to emulate PC XT on a
modern computer.
Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com
Tom Del Rosso
> Could someone design for me a 8086 powered computer which is IBM
> compatible
Tech Manual with full schematics? It must be on ebay occasionally.
There was even an issue of Popular Electronics in the mid 80's with a
full schematic. It might have been called Computers and Electronics at
the time, and it might have been spread over 2 or 3 issues.
--
Reply in group, but if emailing add one more
zero, and remove the last word.
Tom Del Rosso
Tom Del Rosso
> What kind of BIOS is free which enable us to use existing IBM PCDOS
> 3.0
> What do you suggest if Super I/O chip is not easily obtained.
Tauno Voipio
The original PC and XT used 8088, which has an 8 bit data bus.
For a turbo version, use NEC V20. It can also run 8080 code.
--
Tauno Voipio
XT.8086
> This would be a project that could fit into an FPGA. Current
> consumption would be small. IIRC I have a generic XT Bios source
> somewhere which can be assembled into a working BIOS using MASM. About
> 15 years ago I used a modified XT as a simple means to control home
> brew circuits.
>
> --
> Failure does not prove something is impossible, failure simply
> indicates you are not using the right tools...
> nico@nctdevpuntnl (punt=.)
> --------------------------------------------------------------
dumber.
Can you program FPGA in place of Super I/O chip.
I have 2 tube of 8086 I want to use them.
I am very interested to your BIOS, let me know in detail.
XT.8086
>
> For any practical purpose it would be easier to emulate PC XT on a
> modern computer.
>
> Vladimir Vassilevsky
> DSP and Mixed Signal Design Consultanthttp://www.abvolt.com
an emulator.
XT.8086
wrote:
XT.8086
allready available, I decided not to waste my money on it.
I would be staying away from V70 , since I want it to be pure 8086 not
even 8088
Tauno Voipio
voltages with proper sequencing, and a weird clock.
For 8080 code in Intel hardware, think about 8085.
--
Tauno Voipio
XT.8086
wrote:
>
> The original PC and XT used 8088, which has an 8 bit data bus.
> For a turbo version, use NEC V20. It can also run 8080 code.
>
> --
>
> Tauno Voipio
alot.
Vladimir Vassilevsky
XT.8086 wrote:
>>The original PC and XT used 8088, which has an 8 bit data bus.
>>For a turbo version, use NEC V20. It can also run 8080 code.
>>
> alot.
convert 16 bit bus cycles to 2 x 8 bit cycles and back.
hamilton
>
> I want 16-bit CPU running 16-bit bus. that would simplify the design
> alot.
You can get an ARM processor with an external buss that can do the same
thing, and it would be more current.
If you school want s to replace a 8088 design trainer, train they
newbe's with something they can use when they get out of school.
my $,02
hamilton
Nico Coesel
>On Oct 3, 9:45=A0pm, n...@puntnl.niks (Nico Coesel)
>> This would be a project that could fit into an FPGA. Current
>> consumption would be small. IIRC I have a generic XT Bios source
>> somewhere which can be assembled into a working BIOS using MASM. About
>> 15 years ago I used a modified XT as a simple means to control home
>> brew circuits.
>>
>
>dumber.
>Can you program FPGA in place of Super I/O chip.
or Verilog files.
David Eather
support chips plus the 8080 itself. I'll put them on ebay if you want
Robert Baer
Those old boards (and compatibles) from the original PC/XT are still
around.
Find a bunch and offer to buy them from the scrap pile..
Robert Baer
1980 PC/XT.
Robert Baer
knows..
Robert Baer
Robert Baer
TunnelRat
<td...@verizon.net.invalid> wrote:
>XT.8086 wrote:
>> Could someone design for me a 8086 powered computer which is IBM
>> compatible
>
>I think someone did 30 years ago. Can't you find a copy of the IBM PC
>Tech Manual with full schematics? It must be on ebay occasionally.
>
>There was even an issue of Popular Electronics in the mid 80's with a
>full schematic. It might have been called Computers and Electronics at
>the time, and it might have been spread over 2 or 3 issues.
Michael A. Terrell
Robert Baer wrote:
>
> I think the original IBM PC/XT drew about 65W.
63 watt power supply.
--
You can't have a sense of humor, if you have no sense.
XT.8086
XT.8086
wrote:
XT.8086
XT.8086
teach hacking mobile phone. hehe
My department is where the dumb, gather around . :-)
XT.8086
> > I want 16-bit CPU running 16-bit bus. that would simplify the design
> > alot.
>
> That complicates the design alot. To be PC compatible, you have to
> convert 16 bit bus cycles to 2 x 8 bit cycles and back.
>
> Vladimir Vassilevsky
> DSP and Mixed Signal Design Consultanthttp://www.abvolt.com
bit design, running DOS , WordStar, Lotus, and...Turbo Pascal
XT.8086
Jon Kirwan
them. Such machines won't be listed on the web site, though.
And they may not exist there. So I don't know. Just a shot.
I have operational here 80386 running at 33MHz that provides
ISA 16 and 8 bit slots, which I use. Won't sell. I also
have 80486DX system -- same story, I keep it. Also, Pentium
running at 66MHz. I keep that one, too. All good. I like
having ISA around, as I design and build my own ISA boards.
No way do I want to do that with PCI or later. (I have two
editions of Solari's book on the ISA bus that I use as a
reference, plus
You might consider PC/104 boards. Although you are probably
going to be stuck with PCI support, some of the boards still
support the old ISA bus. Connectors won't be of the same
form factor -- they are PC-104, not ISA -- but the bus
transactions may be just fine. Need to read data sheets to
be sure. However, it's likely to work. For example:
http://www.winsystems.com/datasheets/PCM-VDX-1-256-DS.pdf
Supports DOS.
I happen to have some older PC-104 stacks here and they work
fine with DOS, Win98SE, etc. I can run old applications I
wrote back in the good old DOS days, just fine, on them.
But the following suggests to be that 8-bit and 16-bit
transactions are still supported, despite the deprecation of
the old ISA bus:
http://pc104.winsystems.com/products/PCM-WW.cfm
As you can see, prototype boards using the PC/104 connector
are still available.
These also provide support for my suggestion that they still
support the old ISA bus:
http://www.winsystems.com/manuals/ISA-PCM.pdf
http://www.winsystems.com/datasheets/ISA-PCM-DS.pdf
As they provide a board for tying in PC/104 boards into a PC
with genuine ISA slots in them -- for test and development.
Also, here is a short spec:
http://www.winsystems.com/specs/PC104Spec.pdf
You could reasonably build up a PC from these, a power supply
you provide (or buy, because they sell them too), and a case
(which you can also buy, but it won't likely have room for
the old ISA slot connectors... so you may need to do a little
packaging on your own.)
Those processors they sell now will do 16-bit stuff. They
will do 32-bit stuff, too, I think. And 8-bit. And I think,
regardless, they will work with old software you might have
in mind. Won't solve your "run on AA battery" query, won't
provide a teaching platform for 16-bit design, and won't
solve whatever it is you are really about here because I
still don't believe you've been frank and honest yet in the
group about that.
But if you are truly interested in buying an old computer,
then you could credibly fashion such a thing from modern
parts being sold today. If my 80386, 80486DX, and 6 slot ISA
based Pentium boards were to fail, I'd probably pick up one
of these and fabricate the ISA connector slots I want and
select an appropriate PC/104 system to drive all of that and
let that replace stuff.
I also have perhaps a dozen, still shrink-wrapped, DOS 5.0
retail packages from Microsoft sitting on the shelf. I
happen to still like DOS development for boards I build here
and keep a Micosoft vc++ 1.52c compiler (legal) for that
purpose (in fact, Microsoft still provides it for MSDN
licensees for download -- which I also happen to be.) I also
use x86 assembler, as needed, and have a 32-bit protected
mode O/S I've developed for play purposes.
But I don't think I really know what you are about, yet. From
the sound of things so far, and temporarily granting you have
some still-unstated reasonable purpose in mind, your best bet
might be to call up one or more of the PC/104 manufacturers
and ask them to help you out. Chances are one of them will
pass you over to a designer you can talk with and ask
questions regarding what is available for your purposes,
whatever they are. Those folks are in the middle of it and
are your better bet for the most up-to-date data. Make the
phone calls and ask for help. They may even have a design
you can have for free. Could happen. Get that research done
and get back to us about your results. Here's a short list
of folks to call:
http://www.pc104.com/
Jon
Jon Kirwan
>(I have two
>editions of Solari's book on the ISA bus that I use as a
>reference, plus
(I have two editions of Solari's book on the ISA bus that I
Michael A. Terrell
linnix
future of 1990?
RST Engineering
<m...@linnix.info-for.us> wrote:
>Why are you training users for these apps? Sending them back to the
>future of 1990?
before teaching them autorouter. A pyramid without a solid foundation
isn't going to stand up long.
Jim
linnix
Wanderer
> On Oct 3, 9:45 pm, n...@puntnl.niks (Nico Coesel)
>
>
> > This would be a project that could fit into an FPGA. Current
> > consumption would be small. IIRC I have a generic XT Bios source
> > somewhere which can be assembled into a working BIOS using MASM. About
> > 15 years ago I used a modified XT as a simple means to control home
> > brew circuits.
>
> > Failure does not prove something is impossible, failure simply
> > indicates you are not using the right tools...
> > nico@nctdevpuntnl (punt=.)
> > --------------------------------------------------------------
>
> dumber.
> Can you program FPGA in place of Super I/O chip.
>
>
> I am very interested to your BIOS, let me know in detail.
Open Cores has an 8080 and an 80186 core.
http://opencores.org/project,zet86
http://opencores.org/project,cpu8080
Warren
news:6bd7e894-d15e-4134...@d26g2000vby.googlegr
oups.com:
> On Oct 4, 3:17 am, Vladimir Vassilevsky
> <nos...@nowhere.com> wrote:
>
>> > I want 16-bit CPU running 16-bit bus. that would
>> > simplify the design alot.
>>
>> That complicates the design alot. To be PC compatible, you
>> have to convert 16 bit bus cycles to 2 x 8 bit cycles and
>> back.
>> Vladimir Vassilevsky
>
> clean 16- bit design, running DOS , WordStar, Lotus,
> and...Turbo Pascal
Now if only Plan-9 and Mac OSX ran on it-- maybe it's time
I tried it again. I see they're up to version 4.1.4..
Warren
Tauno Voipio
to nearly all you can use for conneting to the Internet.
--
Tauno Voipio
josephkk
<tauno....@notused.fi.invalid> wrote:
>On 3.10.11 9:29 , XT.8086 wrote:
>> Could someone design for me a 8086 powered computer which is IBM
>> compatible
>>
>> requirement:
>>
>> 1. using toroidal transformer as power supply instead of SMPS
>> 2. using SRAM instead of DRAM
>> 3. using super I/O chip
>>
>>
>> this is to replace aging 8088 system, and also used to teach pure 16
>> bit design.
>>
>> thank you
>
>
>The original PC and XT used 8088, which has an 8 bit data bus.
>For a turbo version, use NEC V20. It can also run 8080 code.
the 16 bit data bus use a V30.
?-)
josephkk
<nos...@nowhere.com> wrote:
>
>
>XT.8086 wrote:
>
>>>The original PC and XT used 8088, which has an 8 bit data bus.
>>>For a turbo version, use NEC V20. It can also run 8080 code.
>>>
>> alot.
>
>That complicates the design alot. To be PC compatible, you have to
>convert 16 bit bus cycles to 2 x 8 bit cycles and back.
>
>
>Vladimir Vassilevsky
>DSP and Mixed Signal Design Consultant
>http://www.abvolt.com
100% 16 bit memory however.
?-)
Martin Riddle
"Jon Kirwan" <jo...@infinitefactors.org> wrote in message
news:qbol87lhmcvpqt4io...@4ax.com...
Tern has a variety of x86 stuff.
Cheers
Vladimir Vassilevsky
josephkk wrote:
> On Mon, 03 Oct 2011 15:17:31 -0500, Vladimir Vassilevsky
> <nos...@nowhere.com> wrote:
>>>>The original PC and XT used 8088, which has an 8 bit data bus.
>>>>For a turbo version, use NEC V20. It can also run 8080 code.
>>>>
>>>I want 16-bit CPU running 16-bit bus. that would simplify the design
>>>alot.
>>
>>That complicates the design alot. To be PC compatible, you have to
>>convert 16 bit bus cycles to 2 x 8 bit cycles and back.
>>
> 100% 16 bit memory however.
peripherals should be accessible from 16 bit CPU bus in correct way.
The 8 bit only PC system is much simpler then 8/16 bit as you don't have
to split or merge cycles.
Ken S. Tucker
Reliable, easy to use. Machine code was easy
to write and access.
(I used a tape storage which is a pain though).
Ken
Tauno Voipio
Correctly interfaced 8086 memory supports both 8 and 16 bit transfers.
The condition is that there are separate write enables for odd and
even bytes. The CPU handles the necessary byte swapping.
--
Tauno Voipio
Jon Kirwan
keeping an eye out, too. I vaguely recall there wasn't any
80186 compatibles, either. Seems like the 80286 was the
first cpu that actually became a PC compatible.
Could you point me to anything that might refresh my memory
of an 8086 or 80186 based IBM PC compatible (not just MS-DOS)
that could actually run most IBM PC software without
modification? (I seem to recall some kind of OEM MS-DOS did.
Just not all IBM PC software.)
Jon
Jon Kirwan
boards that expand out the old ISA connectors, 8 or 16 bit.
(I think I knew about them before, but it's nice to get
reminded.) I do see the PC-Co, which requires an ISA slot.
Might be something with through-holes, but no connector, that
I missed. The search box doesn't come up with anything but
the PC-Co, though.
Interesting thing is that I do have the Paradigm tools (and a
few documents and software from other companies that support
DOS development, such as SSI embedded linker/locators and
Microsoft 16-bit compilers.)
But less call for that work, these days. I still dabble,
since I design and build ISA boards, still.
By the way, I got a complete Paradigm development kit for
free with a free license for as many machines as I'd like to
set up for the rest of my life, direct from the owner, some
years ago. My understanding is that the product I received
can be given away, though I'd have to go back and check my
emails first. It's the summer 2002 version, memory serving.
Which is quite adequate for embedded and operating system
work. (Paradigm development tools are based on the Borland c
compiler, but forked and modified, as I recall.)
Jon
Martin Brown
compatible machines which were the subject of an Intel vs NEC patent
infringement for reverse engineering of microcode. NEC might also have
used them - I never had the chance to dismantle an original NEC9801.
The V30 was an 8086 compatible chip was exactly pin compatible with the
Intel but about 20% faster (more on some ops). It could also emulate an
8080 in hardware very convincingly as a sideline. See for example:
http://www.cpu-collection.de/?l0=co&l1=NEC&l2=V20
http://en.wikipedia.org/wiki/NEC_V20
I had an Epsom machine at home with a V30 CPU & Hercules graphics. The
V30 was a heck of a lot faster at bit twiddling than the 8086.
Cyrix did the first 80x87 pin compatible FP chip and their scrupulous
black box analysis of the IEEE floating point specification found
several inconsistencies (ie bugs) in the Intel implementation of it.
The dying embers of Ollivetti M20 Zilog Z8000 machines were still
lingering around at about the same time. Our secretary had one.
Regards,
Martin Brown
JW
wrote in Message id: <ba7t87ltdv7qiidch...@4ax.com>:
The first Compaq Deskpro used an 8086 running at 7.14 MHz, but was capable
of running add-in cards designed for the 4.77 MHz IBM PC XT.
The IBM PS/2 models 25 and 30 were built with an 8 MHz 8086.
The Tandy 1000 SL-series machines used 8086 CPUs.
Martin Brown
And COCOM rules were set by lobbyists so that it was a serious criminal
offence to export BBC micros or Compaq PC compatibles to Russia whilst
IBM XTs were OK. It was particularly bad when in the same week an IBM
salesman got an industry award for selling 2000 IBM XTs (with hard
disks) to Moscow University and a West German businessman was jailed
for 5 years for selling 300 tape based BBC micros to East Berlin.
Ollivetti M24 was a full 8086 IBM PC compatible also sold as AT&T 6300.
http://en.wikipedia.org/wiki/Olivetti_M24
It was highly compatible with the IBM design. The earlier M20 (Z8000
based) was not although they produced a daughter board for it.
>
> The IBM PS/2 models 25 and 30 were built with an 8 MHz 8086.
>
> The Tandy 1000 SL-series machines used 8086 CPUs.
Many 8086 machines in that early era were of limited PC compatibility.
Almost everything had better graphics than the rubbish IBM offering.
Regards,
Martin Brown
k...@att.bizzzzzzzzzzzz
wrote:
>On Thu, 06 Oct 2011 19:06:02 -0700, josephkk
><joseph_...@sbcglobal.net> wrote:
>
>>On Mon, 03 Oct 2011 15:17:31 -0500, Vladimir Vassilevsky
>><nos...@nowhere.com> wrote:
>>
>>>
>>>
>>>XT.8086 wrote:
>>>
>>>>>The original PC and XT used 8088, which has an 8 bit data bus.
>>>>>For a turbo version, use NEC V20. It can also run 8080 code.
>>>>>
>>>> I want 16-bit CPU running 16-bit bus. that would simplify the design
>>>> alot.
>>>
>>>That complicates the design alot. To be PC compatible, you have to
>>>convert 16 bit bus cycles to 2 x 8 bit cycles and back.
>>>
>>>
>>>Vladimir Vassilevsky
>>>DSP and Mixed Signal Design Consultant
>>>http://www.abvolt.com
>>
>>I don't think so. There were 8086 based clones out there. You would need
>>100% 16 bit memory however.
>>
>>?-)
>
>I don't recall ever seeing an 8086 compatible. I remember
>keeping an eye out, too. I vaguely recall there wasn't any
>80186 compatibles, either. Seems like the 80286 was the
>first cpu that actually became a PC compatible.
for a PC type system because of interrupt (and I/O) conflicts. Using it
automatically made it non-comparable. Some did it anyway but anything written
"to the metal" had problems.
>Could you point me to anything that might refresh my memory
>of an 8086 or 80186 based IBM PC compatible (not just MS-DOS)
>that could actually run most IBM PC software without
>modification? (I seem to recall some kind of OEM MS-DOS did.
>Just not all IBM PC software.)
16-bit ISA bus.
Wanderer
> On Thu, 06 Oct 2011 23:42:14 -0700 Jon Kirwan <j...@infinitefactors.org>
> wrote in Message id: <ba7t87ltdv7qiidche2932i9kqjjso5...@4ax.com>:
http://www.redhill.net.au/iu.html
Jon Kirwan
http://en.wikipedia.org/wiki/IBM_Personal_Computer_XT
I owned a PC/XT. It used an 8088. It also didn't have the
16-bit ISA slots. They didn't happen until the IBM PC/AT.
Which I also bought (damned thing cost me $5495.)
I certainly remember these things.
Jon
FatBytestard
<k...@att.bizzzzzzzzzzzz> wrote:
>ALl "XT" class, and compatibles, were 8086 based. The '186 was rarely used
>for a PC type system because of interrupt (and I/O) conflicts. Using it
>automatically made it non-comparable. Some did it anyway but anything written
>"to the metal" had problems.
You're a fuckin' retard. The 80186 was not introduced for use in the
PCs because the 80286 silicon was already in production, and was a huge
step forward in design. The 80186 was dead before it hit the docks, and
the reason had NOTHING to do with its operation.
The 80186 was then relegated to industrial controls, where it worked
just fine (unlike you), interrupts and all.
STOP making shit up, you retarded twit. If you do not have citations
about "interrupt (and I/O) conflicts, shut the fuck up.
The 80186 was fully compatible, and some were actually used in a PC
design and it was successful in said design.
josephkk
<nos...@nowhere.com> wrote:
>
>
>josephkk wrote:
>
>> On Mon, 03 Oct 2011 15:17:31 -0500, Vladimir Vassilevsky
>> <nos...@nowhere.com> wrote:
>
>>>>>The original PC and XT used 8088, which has an 8 bit data bus.
>>>>>For a turbo version, use NEC V20. It can also run 8080 code.
>>>>>
>>>>I want 16-bit CPU running 16-bit bus. that would simplify the design
>>>>alot.
>>>
>>>That complicates the design alot. To be PC compatible, you have to
>>>convert 16 bit bus cycles to 2 x 8 bit cycles and back.
>>>
>> I don't think so. There were 8086 based clones out there. You would need
>> 100% 16 bit memory however.
>
>The ISA bus has to support for 8 bit operation. All 8 bit onboard
>peripherals should be accessible from 16 bit CPU bus in correct way.
>The 8 bit only PC system is much simpler then 8/16 bit as you don't have
>to split or merge cycles.
Of course that task was made massively easier by plenty of IO instructions
that operated on bytes (8 bits) instead of words (16 bits). Even the DMA
chip could operate on bytes.
David Harmon
Kirwan <jo...@infinitefactors.org> wrote,
>I don't recall ever seeing an 8086 compatible. I remember
>keeping an eye out, too. I vaguely recall there wasn't any
>80186 compatibles, either.
I could be vaguely misrecalling, but wasn't the PC JR a 186?
k...@att.bizzzzzzzzzzzz
No, it was an 8088, too.
k...@att.bizzzzzzzzzzzz
wrote:
You're right, the difference between the PC and /XT was the memory
(64Kx1-256/640KB max vs. 16Kx1-64KB max).
>I certainly remember these things.
There were 8086 fully compatible PC/XT class systems made, however.
k...@att.bizzzzzzzzzzzz
<PhatBy...@somewheronyourharddrive.org> wrote:
>On Fri, 07 Oct 2011 10:25:48 -0500, "k...@att.bizzzzzzzzzzzz"
><k...@att.bizzzzzzzzzzzz> wrote:
>
>>ALl "XT" class, and compatibles, were 8086 based. The '186 was rarely used
>>for a PC type system because of interrupt (and I/O) conflicts. Using it
>>automatically made it non-comparable. Some did it anyway but anything written
>>"to the metal" had problems.
>
> You're a fuckin' retard. The 80186 was not introduced for use in the
>PCs because the 80286 silicon was already in production, and was a huge
>step forward in design. The 80186 was dead before it hit the docks, and
>the reason had NOTHING to do with its operation.
How hard do you try to be wrong, AlwaysWrong? ...because you are, *always*.
The /AT came out in '84. The '186/8 was available, in quantity, in '81.
> The 80186 was then relegated to industrial controls, where it worked
>just fine (unlike you), interrupts and all.
>
> STOP making shit up, you retarded twit. If you do not have citations
>about "interrupt (and I/O) conflicts, shut the fuck up.
STOP being wrong, AlwaysWrong.
> The 80186 was fully compatible, and some were actually used in a PC
>design and it was successful in said design.
Wrong. <shock!>
Jon Kirwan
<k...@att.bizzzzzzzzzzzz> wrote:
And an MFM hard drive. Can't recall exactly, but I think it
was 10Mb. Slow. But better than the PC's floppies or the
PC's cassette tape drive interface, which I also remember
because I used a Radio Shack cassette recorder to save
programs.
>>I certainly remember these things.
>
>There were 8086 fully compatible PC/XT class systems made, however.
Yes, I'm getting that message loud and clear. I distinctly
remember 'difficulties' there and a general conclusion "not
to bother with them." But some did, apparently, and fewer
still actually managed to create something PC compatible from
what I've recently read posted here. Supplements and
corrects my recollections.
Jon
k...@att.bizzzzzzzzzzzz
wrote:
Sure. I was talking about the MB.
>>>I certainly remember these things.
>>
>>There were 8086 fully compatible PC/XT class systems made, however.
>
>Yes, I'm getting that message loud and clear. I distinctly
>remember 'difficulties' there and a general conclusion "not
>to bother with them." But some did, apparently, and fewer
>still actually managed to create something PC compatible from
>what I've recently read posted here. Supplements and
>corrects my recollections.
The '286 came out just a couple of years later, about the time the whole thing
took off. After the '286, there was no reason to use the '86 at all. The
'186 was still cost-effective, but it wasn't 100% compatible, without which a
system wasn't sellable.
Vladimir Vassilevsky
josephkk wrote:
To be PC compatible, all 8 and 16 bit instructions have to operate
correctly regardless. That 8/16 bit issue gets more complicated with
PC/AT, and then with 386 and later. Bus interface unit would be hard to
make from discrete logic parts.
Sjouke Burry
news:W6mdnXNtWeX3IAzT...@earthlink.com:
We bought 3 186 IBM pc's at the time, but about 6 months
later they switched to 286 cpu's.
We used IBM network cards in them, those worked without failure
for about 8 years, until replaced by ethernet.(3com).
Never noticed much difference between them, but they all
had math co-processors.
So yes, for a brief time they were on the market.
k...@att.bizzzzzzzzzzzz
>David Harmon <sou...@netcom.com> wrote in
>news:W6mdnXNtWeX3IAzT...@earthlink.com:
>
>> On Thu, 06 Oct 2011 23:42:14 -0700 in sci.electronics.design, Jon
>> Kirwan <jo...@infinitefactors.org> wrote,
>>>I don't recall ever seeing an 8086 compatible. I remember
>>>keeping an eye out, too. I vaguely recall there wasn't any
>>>80186 compatibles, either.
>>
>> I could be vaguely misrecalling, but wasn't the PC JR a 186?
>>
>
>We bought 3 186 IBM pc's at the time, but about 6 months
>later they switched to 286 cpu's.
the only thing they used the 8086 in was the System/23 DataMaster (a.k.a.
DataToaster).
Gerhard Hoffmann
Olivetti M20 / M24??? were 8086 based. It was really fast,
better than early 286s. XTs were 8088 based and slooooow.
We had a lot of the Olivettis running QNX. Was a good thing.
80186 was announced at the very same same time as the 286.
It was not really compatible to an AT or XT, some timer /
DMA / interrupt addresses were different and that could not be healed.
I also got a datasheet for an 80187 under NDA, but that chip
never saw the light of day. Instead they proposed a hack with a 8087 and
a PAL IIRC.
At that time I have built a lot of 80186 machines for Multibus I
and AMS/SMP bus, also with _very_ early 186 silicon.
Gerhard
p.s.
I suppose, Intel gave on the iAPX[1234]86 type numbers, because now they
would have to pay royalties to Apple.
1/2 :-) only.
Sjouke Burry
news:jn84971kbj664ne1s...@4ax.com:
> On 09 Oct 2011 22:10:08 GMT, Sjouke Burry <bu...@planet.nl> wrote:
>
>>David Harmon <sou...@netcom.com> wrote in
>>news:W6mdnXNtWeX3IAzT...@earthlink.com:
>>
>>> On Thu, 06 Oct 2011 23:42:14 -0700 in sci.electronics.design, Jon
>>> Kirwan <jo...@infinitefactors.org> wrote,
>>>>I don't recall ever seeing an 8086 compatible. I remember
>>>>keeping an eye out, too. I vaguely recall there wasn't any
>>>>80186 compatibles, either.
>>>
>>> I could be vaguely misrecalling, but wasn't the PC JR a 186?
>>>
>>
>>We bought 3 186 IBM pc's at the time, but about 6 months
>>later they switched to 286 cpu's.
>
> No you didn't. IBM never used the 186 or 188 in PCs (couldn't have).
> AFAIK, the only thing they used the 8086 in was the System/23
> DataMaster (a.k.a. DataToaster).
Curious. You know better, what we bought???????
I have had years in which to read the label on the
front of the case, showing 186.
That was in the Netherlands, and I dont think the label
was lying.
k...@att.bizzzzzzzzzzzz
>"k...@att.bizzzzzzzzzzzz" <k...@att.bizzzzzzzzzzzz> wrote in
>news:jn84971kbj664ne1s...@4ax.com:
>
>> On 09 Oct 2011 22:10:08 GMT, Sjouke Burry <bu...@planet.nl> wrote:
>>
>>>David Harmon <sou...@netcom.com> wrote in
>>>news:W6mdnXNtWeX3IAzT...@earthlink.com:
>>>
>>>> On Thu, 06 Oct 2011 23:42:14 -0700 in sci.electronics.design, Jon
>>>> Kirwan <jo...@infinitefactors.org> wrote,
>>>>>I don't recall ever seeing an 8086 compatible. I remember
>>>>>keeping an eye out, too. I vaguely recall there wasn't any
>>>>>80186 compatibles, either.
>>>>
>>>> I could be vaguely misrecalling, but wasn't the PC JR a 186?
>>>>
>>>
>>>We bought 3 186 IBM pc's at the time, but about 6 months
>>>later they switched to 286 cpu's.
>>
>> No you didn't. IBM never used the 186 or 188 in PCs (couldn't have).
>> AFAIK, the only thing they used the 8086 in was the System/23
>> DataMaster (a.k.a. DataToaster).
>
>Curious. You know better, what we bought???????
I know it wasn't an IBM, as described. Yes.
>I have had years in which to read the label on the
>front of the case, showing 186.
>That was in the Netherlands, and I dont think the label
>was lying.
Someone is.
XT.8086
> On Tue, 4 Oct 2011 10:08:15 -0700 (PDT),
>
> For the same reason I teach my kids basic tape and donuts layout
> before teaching them autorouter. A pyramid without a solid foundation
> isn't going to stand up long.
>
> Jim
Hoorah, cavalry coming to the rescue. Thanks Buddy....:-)
XT.8086
> josephkk wrote:
> > On Mon, 03 Oct 2011 15:17:31 -0500, Vladimir Vassilevsky
> > <nos...@nowhere.com> wrote:
> >>>>The original PC and XT used 8088, which has an 8 bit data bus.
> >>>>For a turbo version, use NEC V20. It can also run 8080 code.
>
> >>>I want 16-bit CPU running 16-bit bus. that would simplify the design
> >>>alot.
>
> >>That complicates the design alot. To be PC compatible, you have to
> >>convert 16 bit bus cycles to 2 x 8 bit cycles and back.
>
> > I don't think so. There were 8086 based clones out there. You would need
> > 100% 16 bit memory however.
>
> The ISA bus has to support for 8 bit operation. All 8 bit onboard
> peripherals should be accessible from 16 bit CPU bus in correct way.
> The 8 bit only PC system is much simpler then 8/16 bit as you don't have
> to split or merge cycles.
>
> DSP and Mixed Signal Design Consultanthttp://www.abvolt.com
I do remember 386 board with 8-bit CGA on it. while all of its bus is
16-bit wide.
Could you explain more about correctly accessing 8-bit bus in 16 bit
ISA bus, particularly for 8086 processor.
separate document or web site would be better I think.
Thanks
XT.8086
wrote:
we can obtain 16-bit wide SRAM today, does byte swapping still any use.
XT.8086
> better than early 286s. XTs were 8088 based and slooooow.
>
> We had a lot of the Olivettis running QNX. Was a good thing.
>
> 80186 was announced at the very same same time as the 286.
> It was not really compatible to an AT or XT, some timer /
> DMA / interrupt addresses were different and that could not be healed.
>
> I also got a datasheet for an 80187 under NDA, but that chip
> never saw the light of day. Instead they proposed a hack with a 8087 and
> a PAL IIRC.
>
> At that time I have built a lot of 80186 machines for Multibus I
> and AMS/SMP bus, also with _very_ early 186 silicon.
>
> Gerhard
>
> p.s.
> I suppose, Intel gave on the iAPX[1234]86 type numbers, because now they
> would have to pay royalties to Apple.
>
> 1/2 :-) only.
How come 8086 machine is faster than 80286 machine, was it clocked
faster, which I doubt.
Berlin wall is Kaput, could you share the NDA document with us now.
Danke Schoon
XT.8086
<k...@att.bizzzzzzzzzzzz> wrote:
> On Thu, 06 Oct 2011 23:42:14 -0700, Jon Kirwan <j...@infinitefactors.org>
I believe 8088 is crippled version of 8086, with only 8-bit data bus
XT.8086
wrote:
> "XT.8086" wrote:
>
> > On Oct 4, 9:04 am, Robert Baer <robertb...@localnet.com> wrote:
> > > XT.8086 wrote:
> > > > Could someone design for me a 8086 powered computer which is IBM
> > > > compatible
>
> > > > requirement:
>
> > > > 1. using toroidal transformer as power supply instead of SMPS
> > > > 2. using SRAM instead of DRAM
> > > > 3. using super I/O chip
>
> > > > this is to replace aging 8088 system, and also used to teach pure 16
> > > > bit design.
>
> > > > thank you
>
> > > Why re-invent the wheel or re-design something available?
> > > Those old boards (and compatibles) from the original PC/XT are still
> > > around.
> > > Find a bunch and offer to buy them from the scrap pile..
>
> > where is it. anywhere nearby here is at least Pentium 500 Mhz board.
>
> Original XT motherboards aren't hard to find.
>
> --
> You can't have a sense of humor, if you have no sense.
hmm , I would like the one with cassette port in it.
Lots of fun :-)
Michael A. Terrell
"XT.8086" wrote:
>
> On Oct 4, 11:29 pm, "Michael A. Terrell" <mike.terr...@earthlink.net>
> wrote:
> > "XT.8086" wrote:
> >
> > > On Oct 4, 9:04 am, Robert Baer <robertb...@localnet.com> wrote:
> > > > XT.8086 wrote:
> > > > > Could someone design for me a 8086 powered computer which is IBM
> > > > > compatible
> >
> > > > > requirement:
> >
> > > > > 1. using toroidal transformer as power supply instead of SMPS
> > > > > 2. using SRAM instead of DRAM
> > > > > 3. using super I/O chip
> >
> > > > > this is to replace aging 8088 system, and also used to teach pure 16
> > > > > bit design.
> >
> > > > > thank you
> >
> > > > Why re-invent the wheel or re-design something available?
> > > > Those old boards (and compatibles) from the original PC/XT are still
> > > > around.
> > > > Find a bunch and offer to buy them from the scrap pile..
> >
> > > where is it. anywhere nearby here is at least Pentium 500 Mhz board.
> >
> > Original XT motherboards aren't hard to find.
>
>
> Lots of fun :-)
Do you mean the original five slot PC motherboard?
k...@att.bizzzzzzzzzzzz
k...@att.bizzzzzzzzzzzz
four-byte prefetch queue, while the 8086 had a six-byte prefetch queue. This
difference was what was used to ID the particular processor.
The 80188 and 80186 were the same die (bond-out difference), however.
k...@att.bizzzzzzzzzzzz
It wasn't. For the 8086 to be faster, it had to be a pretty horrid
implementation of the '286 system. The '86 was a *lot* cheaper, however.
k...@att.bizzzzzzzzzzzz
TheGlimmerMan
Jon Kirwan
There was one problem with the 80286, solved in the 80386 and
beyond, that greatly slowed it down when using memory beyond
the 20 address limit (in other words, the upper four address
lines of the 80286 were non-zero.) [Obviously, this does NOT
relate to the above discussion about the 8086 vs 80286, but
it made me remember things so I'm writing it anyway. :P ]
The 80286 could not shift back from protected mode into real
mode without a processor reset. So the BIOS routines that
supported accessing memory past the 1Mb boundary would set up
a protected mode environment, shift to protected mode,
perform the indicated operation, and then ... well, then the
BIOS used the clock calendar chip with a very tiny battery
backed memory, stored a special code there so that on power
reset it could look and realize that it was coming back from
a protected mode call, and then would return to the caller in
real mode rather than reboot the O/S from disk.
Which reminds me of yet another oddity. Anyone remember (I
do) that code can be downloaded into a motherboard via the
keyboard interface -- no disks required? A feature used to
download diagnostic/test code for motherboard manufacturing.
Jon
Michael A. Terrell
looked at the pile of PC/XT stuff in 10 years.
Gerhard Hoffmann
real MMU and the address translation took time, cycle-wise. Also, 8086s
were available at 10 and 12 MHz from Siemens and other 2nd sources,
while initial 286 clock rates were 6 and 8 MHz. (guess what IBM choose)
In addition, the first board designs did not use the
address pipelining feature of the 286.
But then I had an ALR Fastboard-286 that reliably ran at 10 MHz with
an overclocked Intel sample processor. The ALR was one of the
first "good ones". It even ran "Interactive Unix". I tried to
compile Berkeley spice 2G6 on it, still the Fortran version, and it
was a fiasco with the large memory model.
I also had to patch the disk descriptor tables in the BIOS to
make that _HUGE_ 70 MB disk fit.
Gerhard
k...@att.bizzzzzzzzzzzz
wrote:
>On Mon, 10 Oct 2011 10:12:21 -0500, "k...@att.bizzzzzzzzzzzz"
;-)
Yes, but the keyboard trick (patented by IBM, IIRC) worked reasonably well.
>The 80286 could not shift back from protected mode into real
>mode without a processor reset. So the BIOS routines that
>supported accessing memory past the 1Mb boundary would set up
>a protected mode environment, shift to protected mode,
>perform the indicated operation, and then ... well, then the
>BIOS used the clock calendar chip with a very tiny battery
>backed memory, stored a special code there so that on power
>reset it could look and realize that it was coming back from
>a protected mode call, and then would return to the caller in
>real mode rather than reboot the O/S from disk.
>Which reminds me of yet another oddity. Anyone remember (I
>do) that code can be downloaded into a motherboard via the
>keyboard interface -- no disks required? A feature used to
>download diagnostic/test code for motherboard manufacturing.
Part of (side effect of) the cassette interface, IIRC.
k...@att.bizzzzzzzzzzzz
<mike.t...@earthlink.net> wrote:
Mine are complete units, minus some case screws, perhaps. I don't have the
original SS diskette drive, however.
Gerhard Hoffmann
> On Mon, 10 Oct 2011 09:04:06 -0700, Jon Kirwan<jo...@infinitefactors.org>
> wrote:
>> The 80286 could not shift back from protected mode into real
>> mode without a processor reset. So the BIOS routines that
>> supported accessing memory past the 1Mb boundary would set up
>> a protected mode environment, shift to protected mode,
>> perform the indicated operation, and then ... well, then the
>> BIOS used the clock calendar chip with a very tiny battery
>> backed memory, stored a special code there so that on power
>> reset it could look and realize that it was coming back from
>> a protected mode call, and then would return to the caller in
>> real mode rather than reboot the O/S from disk.
Yes! Xilinx place & route did that once a millisecond to keep
the mouse happy.
Gerhard
Jon Kirwan
And it was entirely done over the keyboard interface. I
don't know why the cassette interface would be required. I
think that one was supported by the extended BIOS that IBM
provided, which included BASIC. I'd have to go back to my
BIOS listings (IBM provided them) to be sure, though.
Jon
k...@att.bizzzzzzzzzzzz
wrote:
Well, it was 30 years ago... ;-)
I seem to remember that it was. The cassette port was part of extended BIOS,
though you're right it didn't exist after the 5150.
josephkk
<nos...@nowhere.com> wrote:
>
>
>josephkk wrote:
>
>> On Fri, 07 Oct 2011 00:20:51 -0500, Vladimir Vassilevsky
>> <nos...@nowhere.com> wrote:
>>
>>
>>>
>>>josephkk wrote:
>>>
>>>
>>>>On Mon, 03 Oct 2011 15:17:31 -0500, Vladimir Vassilevsky
>>>><nos...@nowhere.com> wrote:
>>>
>>>>>>>The original PC and XT used 8088, which has an 8 bit data bus.
>>>>>>>For a turbo version, use NEC V20. It can also run 8080 code.
>>>>>>>
>>>>>>
>>>>>>I want 16-bit CPU running 16-bit bus. that would simplify the design
>>>>>>alot.
>>>>>
>>>>>That complicates the design alot. To be PC compatible, you have to
>>>>>convert 16 bit bus cycles to 2 x 8 bit cycles and back.
>>>>>
>>>>
>>>>I don't think so. There were 8086 based clones out there. You would need
>>>>100% 16 bit memory however.
>>>
>>>The ISA bus has to support for 8 bit operation. All 8 bit onboard
>>>peripherals should be accessible from 16 bit CPU bus in correct way.
>>>The 8 bit only PC system is much simpler then 8/16 bit as you don't have
>>>to split or merge cycles.
>>
>>
>> Of course that task was made massively easier by plenty of IO instructions
>> that operated on bytes (8 bits) instead of words (16 bits). Even the DMA
>> chip could operate on bytes.
>
>To be PC compatible, all 8 and 16 bit instructions have to operate
>correctly regardless. That 8/16 bit issue gets more complicated with
>PC/AT, and then with 386 and later. Bus interface unit would be hard to
>make from discrete logic parts.
>
>
>Vladimir Vassilevsky
>DSP and Mixed Signal Design Consultant
>http://www.abvolt.com
Thus it is pretty convenient that the 8086 chips have proper interface
electronics onboard for that. Then you have a much easier time designing
the matching circuitry for memory and IO.
?-)
josephkk
The difference in memory bandwidth made for a rather noticeable difference
in performance.