Searching \ for '[EE]: Scratchbuilt computer?' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=scratchbuilt+computer
Search entire site for: 'Scratchbuilt computer?'.

Exact match. Not showing close matches.
PICList Thread
'[EE]: Scratchbuilt computer?'
2002\04\09@005314 by Tony Goetz

picon face
Hey, I was talking to a friend earlier and we were wondering something. Has
anyone built a PC from scratch? i.e. design and build a motherboard,
incorporating a pentium class processor or similar? (Still probably using
standard memory, IDE drives, cards, etc.) Okay, yes, I suppose IBM could fall
into that category, as well as the geeks of the 80's that got us where we are
today. But nowadays, has anyone taken a commercial processor and built a
custom motherboard for their own PC? Reprogrammed a pentium/athlon/whatever?
Seems like it would be a heck of a project, but quite a rewarding one! I did
some searches on the web but, as expected, ran into little more than how to
build your own computer using commercial motherboards.

Just a bit curious. If nothing like this is out there - why don't someone get
to work?? ;)

-Tony

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@013935 by Nick Veys

flavicon
face
To get anything even mildly useful would be pretty damn tough for a
person without pretty impressive connections to manufacturers.  The
purchase quantities on chipsets and stuff is no doubt insane.  Not to
mention multilayer boards aren't exactly a hobbiest-realm deal, despite
being able to comprehend the design, many of us just simply couldn't do
it in a cost-effective way seeing that a very very good motherboard for
a P4 or Athlon chip is in the $50-$150 range depending on your
stingyness and feature desire.

And re-inventing the wheel isn't all that fun when you could be making
cool gadgets that don't exist yet! :)

spam_OUTnickTakeThisOuTspamveys.com | http://www.veys.com/nick

> {Original Message removed}

2002\04\09@020412 by Vit

picon face
Tony,

What is the point of this project?  What are the "rewards" you are talking
about?  To be honest with you, this idea sounds to me like a waste of time
and resources.

Sincerely,

Vitaliy

> Hey, I was talking to a friend earlier and we were wondering something.
Has
> anyone built a PC from scratch? i.e. design and build a motherboard,
> incorporating a pentium class processor or similar? (Still probably using
> standard memory, IDE drives, cards, etc.) Okay, yes, I suppose IBM could
fall
> into that category, as well as the geeks of the 80's that got us where we
are
> today. But nowadays, has anyone taken a commercial processor and built a
> custom motherboard for their own PC? Reprogrammed a
pentium/athlon/whatever?
> Seems like it would be a heck of a project, but quite a rewarding one! I
did
> some searches on the web but, as expected, ran into little more than how
to
> build your own computer using commercial motherboards.
>
> Just a bit curious. If nothing like this is out there - why don't someone
get
> to work?? ;)

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@080659 by Olin Lathrop

face picon face
> Hey, I was talking to a friend earlier and we were wondering something.
Has
> anyone built a PC from scratch? i.e. design and build a motherboard,
> incorporating a pentium class processor or similar? (Still probably using
> standard memory, IDE drives, cards, etc.) Okay, yes, I suppose IBM could
fall
> into that category, as well as the geeks of the 80's that got us where we
are
> today. But nowadays, has anyone taken a commercial processor and built a
> custom motherboard for their own PC? Reprogrammed a
pentium/athlon/whatever?
> Seems like it would be a heck of a project, but quite a rewarding one! I
did
> some searches on the web but, as expected, ran into little more than how
to
> build your own computer using commercial motherboards.

Today's motherboards are assembled automatically in such vast quantities
that building an individual one will be much more expensive than buying one
off the shelf.  If you can even get the parts in one-off quantities, they
will be difficult if not impossible to assemble by hand.

It depends on what your goal is.  If you want to have a computer, just buy
one and move on.  If you want build your own operating system and software
from the ground up, you can still start with a standard PC.  If you want to
build the hardware for a computer system yourself, then you might be better
off with a different processor.  How about one of the 18 family PICs with
external memory?  That will be a lot easier to assemble, but still powerful
enough to run real programs on.  The Harvard architecture of PICs makes them
less than ideal for a general purpose computer, however.  There may be
variants of the PowerPC that can be assembled by hand.  I would look at the
various microprocessor families to see what might be suitable for a home
built general purpose computer.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@093146 by michael brown

flavicon
face
> The Harvard architecture of PICs makes them
> less than ideal for a general purpose computer, however.  There may be
> variants of the PowerPC that can be assembled by hand.  I would look at
the
> various microprocessor families to see what might be suitable for a home
> built general purpose computer.

This is why I like the 8052 arch.  It's still a harvard arch (more or less)
but the newer "expanded" and "high speed" parts have allot of nice features.
AFAIK, all of the parts have the ability to combine the code and data
segments.  This lets you do all the "really fun" stuff that you can't
normally do on a harvard arch; like writing self modifying code. ;-)))
Another benefit is that you can start small and add more capability later
(such as external RAM and ROM up to 4Mb on Dallas 80c390).  The 8051/52
arch. is over 20 years old now and keeps on going like the Energizer bunny.
;-)  It has it's weaknesses of course, but then all platforms do.

As Olin says, there may be PPC, MIPS, 68K, Strong-Arm etc.... parts that you
could work with more easily than constructing an i86 based system from
scratch.

Some more reasons to lower your ambitions on a full fledged computer:

You can build 8052's up on breadboards and they will work fine.  A pentium
(or faster) is "never" going to run on a breadboard.  Interfacing to DRAM
memory is not fun, SRAM's are way easier to talk to.  The PC bus and chipset
is extremely complex, what with the ISA, PCI and AGP busses and their
associated bridges.  The test equipment needed to analyze and debug a PC bus
and memory interface would most probably not be available to you.  Once you
have everything connected correctly, you would need a physicist to help
explain why things didn't work as expected.

I'm not trying to "rain on your parade", but you may wish to first start
with a somewhat more "workable" set of hardware.  Regardless of what others
may think, it would *not* be a "waste of time" to attempt such a project.
The learning experience would be more than worthwhile.  And, I *can* see the
personal reward in accomplishing such a project, even if, in the end, it was
only able to display the traditional "hello world" on a display screen.  ;-)

If you still wish to proceed with your original idea, I wish you much luck
and fun.

michael brown

michael brown

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@100022 by Eoin Ross

flavicon
face
www.zfmicro.com/zfx86.html
These may simplfy your task a whole lot - never used them myself

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@101859 by M. Adam Davis

flavicon
face
The timing is so critical for these faster processors that without
expensive design tools (or a nice calculator and a few years of
computations) you simply won't be able to make a reliable board.

It's not so much connecting the chips together, but routing the board so
the wires on a bus are close enough in length that the signal arrives at
the next device within certian parameters.

Something to think about:  At 1GHz light travels just under 30cm, often
less than the depth of your case, in one clock cycle.  At these speeds
you have to take into account the shape of the signal when it arrives at
the chip it's intended for.  You may see squigly lines (lines that
wander back and forth for no apparent reason) on motherboards you deal
with - these aren't just to look cool...

-Adam

-Adam

Tony Goetz wrote:

{Quote hidden}

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@103358 by Herbert Graf

flavicon
face
Probably the biggest reason is that it is so complex few have the resources
and time to do it. You can't simply do this on perf board, at least you
can't if you want to get anywhere near designed speed, and these types of
processors don't like running too slowly. Most motherboards out there are
simply point to point wired, this pin on the processor goes directly to the
north bridge, this pin goes directly to the south bridge, this pin goes
directly to the ISA slot. There simply isn't very much "fun" going on,
mostly everything is point to point. About the only "fun" area left is the
regulator on the board, that usually has a chunk of discrete components.

Aside from this is the fact that most package types used these days are NOT
hobbiest friendly, while you could hand solder a TSOP package I don't see
how you could solder a BGA package by hand, and AFAIK all currrent chipsets
are in BGA packaging.

TTYL

> {Original Message removed}

2002\04\09@103837 by Vit

picon face
> I'm not trying to "rain on your parade", but you may wish to first start
> with a somewhat more "workable" set of hardware.  Regardless of what
others
> may think, it would *not* be a "waste of time" to attempt such a project.
> The learning experience would be more than worthwhile.  And, I *can* see
the
> personal reward in accomplishing such a project, even if, in the end, it
was
> only able to display the traditional "hello world" on a display screen.
;-)

Michael,

You missed my point!  :)

<quote>
"i.e. design and build a motherboard,
incorporating a pentium class processor or similar? (Still probably using
standard memory, IDE drives, cards, etc.) "
</quote>

The thing would cost *at least* several thousands when it is completed, and
the project will take several years of hard work.  Seeing "hello world" on
the screen is *not* worth it!  :)

Last summer I started working on a project that I thought was fun.  I
estimated it would take me two months to build a working prototype and then
I would try to sell it.  I estimated the cost of parts to be around $50.  So
far, the project cost me and my partner over a thousand dollars, and
hundreds of hours spent doing research, soldering, coding, plotting,
etching, drilling, troubleshooting...

I don't mean to sound negative or pessimistic, but in my opinion, building a
PC clone is just not worth it.  I simply wanted to keep Tony and his friend
from wasting their time and money on something that is doomed to fail.

As Jesus has put it,  "Suppose one of you wants to build a tower.  Will he
not first sit down and estimate the cost to see if he has enough money to
complete it?  For if he lays the foundation and is not able to finish it,
everyone who sees it will ridicule him, saying, 'This fellow began to build
and was not able to finish'".

Best wishes to ya'll.

Vitaliy

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@111050 by Olin Lathrop

face picon face
> Something to think about:  At 1GHz light travels just under 30cm, often
> less than the depth of your case, in one clock cycle.

And that's the propagation speed in vacuum (30cm or about 1 foot per nS).
Figure about half that for an electric signal on a transmission line, like a
PC board bus trace.  So 6 inches of PC board trace is one clock cycle at
1GHz.  Still want to build this yourself?


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@111717 by Kevin Blain

flavicon
face
Other way round, if the propogation speed is half (which sounds about
right, Er of FR4 is about 3.5 to 4 at 1GHz) then the wavelength with be
double. V = f * lambda. So at 1Ghz wavelength with velocity factor of
0.5 will be about 600mm.

Regards, Kevin

Olin wrote:

{Quote hidden}

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@114214 by michael brown

flavicon
face
> > I'm not trying to "rain on your parade", but you may wish to first start
> > with a somewhat more "workable" set of hardware.  Regardless of what
> others
> > may think, it would *not* be a "waste of time" to attempt such a
project.
{Quote hidden}

and
> the project will take several years of hard work.

I was trying to say the same thing, but with less impetus.  ;-)  Kind of
like the difference between planting a seed and hitting someone with a tree.
;-D  They would probably have to invest that much money in just getting a
working motherboard, but if they have the money and tools, more power to
them.  ;-)  At some point they will be able to accomplish the whole project
using just a single FPGA with no "special" requirements other than a funky
socket and  some way to attach to it (probably thru a blue-tooth like short
distance rf interface ;-).  Meanwhile, they will have learned invaluable
lessons about how electricity works in the "real" world.  ;-)

>Seeing "hello world" on the screen is *not* worth it!  :)

After all the hard work, it would be much better than not seeing it.  ;-D

> Last summer I started working on a project that I thought was fun.  I
> estimated it would take me two months to build a working prototype and
then
> I would try to sell it.  I estimated the cost of parts to be around $50.
So
> far, the project cost me and my partner over a thousand dollars, and
> hundreds of hours spent doing research, soldering, coding, plotting,
> etching, drilling, troubleshooting...

Yes, I can see how that happens.  ;-)  BTDT  ;-)

> I don't mean to sound negative or pessimistic, but in my opinion, building
a
> PC clone is just not worth it.  I simply wanted to keep Tony and his
friend
> from wasting their time and money on something that is doomed to fail.

I said the same thing, I thought???

> As Jesus has put it,  "Suppose one of you wants to build a tower.  Will he
> not first sit down and estimate the cost to see if he has enough money to
> complete it?  For if he lays the foundation and is not able to finish it,
> everyone who sees it will ridicule him, saying, 'This fellow began to
build
> and was not able to finish'".

Be carefull, there are those here who will smite you for saying that.  It
has nothing to do with the meaning of the message, only He who said it.  ;-)

BTW, please take out your reply-to address.

> Best wishes to ya'll.
>
> Vitaliy

Take care,

michael brown

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@120115 by john

flavicon
face
Just my 2 cents worth.
   1...You will NOT bring up a P class CPU with a FPGA.
   2...One no longer 'designs. an MB. That is done by the chipset
makers. All you do is make                       decisions about sockets
& peripherals.
   3...You are going to throw high 4 figure money after a low 3 figure
product. The PCB CAD tools               alone would buy you a box full
of MBs. Trying to manually route a board of that
             complexity would try the patience of  a monk.
   4...The most fun you can have with your clothes on, does not usually
involve self immolation !
          I have spent a lot of time and money on non-profit projects.
I.E. 'just for fun'. But if you have
          the resources that this project would require available and
can use them for fun...buy a
          sports car.

But hey...that's just my opinion, I could be wrong.
John M

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@121553 by Dal Wheeler

flavicon
face
Not that I don't disagree that the project is ill advised; but what current
home level PC has a 1Ghz bus?

{Original Message removed}

2002\04\09@121802 by Olin Lathrop

face picon face
> Other way round, if the propogation speed is half (which sounds about
> right, Er of FR4 is about 3.5 to 4 at 1GHz) then the wavelength with be
> double. V = f * lambda. So at 1Ghz wavelength with velocity factor of
> 0.5 will be about 600mm.

Oh, yeah.  Duh!


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@121946 by Douglas Butler

flavicon
face
>            I have spent a lot of time and money on non-profit
> projects.
> I.E. 'just for fun'. But if you have
>            the resources that this project would require available and
> can use them for fun...buy a
>            sports car.
>
... or build a sports cars ;-)

Sherpa Doug

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@122625 by Mark J. Dulcey

flavicon
face
Dal Wheeler wrote:
> Not that I don't disagree that the project is ill advised; but what current
> home level PC has a 1Ghz bus?

You're right. There aren't any 1 GHz off-chip signals in any home PC.
The fastest signals likely to be found in most computers are 266 MHz
(PC2100 DDR SDRAM, 4x AGP), or whatever the off-chip speed used with
RDRAM is. That's still fast enough to be difficult to work with,
however. And even faster signals are in the works; 8x AGP and DDR-II
(533 MHz) are on the way by next year.

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@123422 by john

flavicon
face
Actually P4 FSB is 400MHZ, and the 850 chipset has a 400MHZ RAM IF

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@123431 by Sean H. Breheny

face picon face
No, Olin was right the first time around. Your formula, V=f*lambda is
correct. If we re-write this, we get
lambda=V/f. Halving V also halves the wavelength. The slower it goes, the
less distance it can go in a cycle.

Sean

At 12:16 PM 4/9/02 -0400, Olin Lathrop wrote:
{Quote hidden}

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@131143 by Olin Lathrop

face picon face
> > Other way round, if the propogation speed is half (which sounds about
> > right, Er of FR4 is about 3.5 to 4 at 1GHz) then the wavelength with be
> > double. V = f * lambda. So at 1Ghz wavelength with velocity factor of
> > 0.5 will be about 600mm.
>
> Oh, yeah.  Duh!

No, wait.  If the propagation speed is half that of vacuum, a wave front
will travel half as far in the same time.  If it travels 12 inches in 1nS in
vacuum, then it will travel 6 inches during the same 1nS at half the speed
of light.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@135850 by Dale Botkin

flavicon
face
Don't let the naysayers talk you out of what could be a very educational
experience.  (only half joking here...)  Just consider a 386-33 about your
top end for homebrewing.  Actually, you could do worse than some of the
x86 embedded vaiations on a small single-board computer, and some of them
are low-tech enough and integrated enough that it might be do-able.

In the end -- assuming enoughtime and money invested -- you'll have a very
limited, very low-end, very expensive and labor-intensive system.  On the
other hand, look at the cost of tuition, it's probably not a bad deal for
what you'll learn in the process.  Just don't expect to build the
equivalent of a $500 consumer PC on your workbench.

Dale
(who has scratchbuilt before, but the last one was an 8085!)
--
"Curiosity is the very basis of education and if you tell me that
curiosity killed the cat, I say only the cat died nobly."
         - Arnold Edinborough

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@155105 by Micro Eng

picon face
Company I work for DID...used the DEC chip and custom form factor...


_________________________________________________________________
Join the world s largest e-mail service with MSN Hotmail.
http://www.hotmail.com

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@201026 by Tony Goetz

picon face
I got home, signed on, and looked at the mail box. Quite a response!

Don't get me wrong, guys. I have (and had) little intention of doing this
myself. Didn't realize just how good that mindset was until reading the
replies, though! Heck, I have to be careful just so I can go out on Friday's
with friends, much less buy a processor to mess around with, much much less
start buying chipsets and manufacturing boards. I figured someone out there
would have done something like this though, whether buying chipsets or makign
their own (I have no idea what's involved in the process, but I get the idea
it's...intensive). Someone with a ton of time, plenty of money, little life,
and lots of patience. Looks like you need to amplify all those requirements a
fair amount... There's a site from Stanford U that shows a 486-SX based
computer they built that's the size of a matchbox. Appropriately named the
matchbox PC and, likewise, their matchbox webserver each running either Win9x
or Red Hat 5.2 Linux. I thought if they could come up with this, maybe it had
been taken a step or two further at some point. If you want to see the 486
system, go to http://wearables.stanford.edu/

It was said you'd have to buy chipsets costing up in the thousands, and part
of the thread got into the actual physical considerations to account for in
lengths of traces, etc. I realize the limits you start to push when dealing
with today's gigahertz machines. The hardware to support these processors is
no doubt some pretty high performance stuff.

Probably starting another arm of the thread, what about yesterday's
performance processors? Somewhere in between the 486 matchboxPC and, say,
dual P4's? For the sake of argument, a Pentium 133 or 233, whose support
hardware, while maybe complex compared to the Pic's we're used to, is
possibly worth very little to anyone but hobbyists nowadays. With SX uC's up
to 100Mhz now (I think that's right. correct me if it's not), we probably
have a good idea of what to do with speeds really getting away from the
20Mhz/5Mips systems that we use. Okay, I guess I can see the difference
between 8-bit uC's and 32-bit PC procs. Just trying to dig the hole deeper
here! Even so, the idea intrigues me, even if I don't have the
time/money/skill to do it.

It was just a passing thought. Michael said exactly what I was thinking, with
the reward being a piece of hardware that you can take pride in and what is
apparently a great deal of knowledge learnt. That would be worth a great deal
to me, though I doubt I could actually follow through and come up with
something like this myself.

And BTW, as always thanks for everyone replying. Always lots of enlightening
responses.

-Tony

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@203943 by Jon Baker

flavicon
face
> It was just a passing thought. Michael said exactly what I was thinking,
with
> the reward being a piece of hardware that you can take pride in and what
is
> apparently a great deal of knowledge learnt. That would be worth a great
deal
> to me, though I doubt I could actually follow through and come up with
> something like this myself.

I always liked the motorola 6800 series.. especially the 6809. I designed my
own computer based on a 6809 but havent got around to building it until now.
I recently replaced most of the IO logic with PICs to allow for a bit more
flexibility once I assemble it. I'm really looking forward to powering it up
and seeing it work- even though it won't be particularly spectacular to
others, I can't wait to write an email with it or something- It doesnt have
a real purpose other than for me to be proud of.

Unfortunatley the older processors are getting hard to find so its going to
get more difficult and ultimately impossible to do this kind of thing in
future-for a hobbyist at least.

--
Jon Baker

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@205140 by Peter Wintulich

flavicon
face
Hi,

A few years ago I built some slower boards, 386SX-25 and a 486 DX (5V only) board. We made about 1500 of the 386's and 10 prototypes of the 486's. The 486 took about 6 months to do the circuit design & Pcb layout about %90 SMT. These boards were some of the first ones to integrate the I/O.

One thing I wanted to do was try and integrate a 68000 processor into the system to play with. Probably end up being a bus mastering AT card ?

Peter
>>> .....NaB25JKILLspamspam@spam@AOL.COM 04/09/02 01:59PM >>>
Hey, I was talking to a friend earlier and we were wondering something. Has
anyone built a PC from scratch? i.e. design and build a motherboard,
incorporating a pentium class processor or similar? (Still probably using
standard memory, IDE drives, cards, etc.) Okay, yes, I suppose IBM could fall
into that category, as well as the geeks of the 80's that got us where we are
today. But nowadays, has anyone taken a commercial processor and built a
custom motherboard for their own PC? Reprogrammed a pentium/athlon/whatever?
Seems like it would be a heck of a project, but quite a rewarding one! I did
some searches on the web but, as expected, ran into little more than how to
build your own computer using commercial motherboards.

Just a bit curious. If nothing like this is out there - why don't someone get
to work?? ;)

-Tony

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@211101 by Andrew Warren

flavicon
face
Jon Baker <PICLISTspamKILLspammitvma.mit.edu> wrote:

> I always liked the motorola 6800 series.. especially the 6809.

   Me, too, and we're not the only ones; the Apple Macintosh was
   originally to have been built with a 6809, and 6809-based
   prototypes were actually built.

   -Andy

=== Andrew Warren -- .....aiwKILLspamspam.....cypress.com
=== Principal Design Engineer
=== Cypress Semiconductor Corporation
===
=== Opinions expressed above do not
=== necessarily represent those of
=== Cypress Semiconductor Corporation

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\09@215131 by Nick Veys

flavicon
face
Someone mentioned StrongARM processors in the beginning here.  Might I
suggest looking into those?  Many people have made Linux machines out of
a home-made Intel ARM processor and associated hardware, it's definately
more in your grasp, and the Linux kernel supports the processor already.

EraseMEnickspam_OUTspamTakeThisOuTveys.com | http://www.veys.com/nick

> {Original Message removed}

2002\04\09@223306 by michael brown

flavicon
face
>
> > I always liked the motorola 6800 series.. especially the 6809.
>
>     Me, too, and we're not the only ones; the Apple Macintosh was
>     originally to have been built with a 6809, and 6809-based
>     prototypes were actually built.

Weren't these the first micro's back then to have a hardware divide
instruction?

michael

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\04\10@032824 by Kevin Blain

flavicon
face
Oh, yeah, Double Duh!

What was I thinking!!!

Apologies.

{Quote hidden}

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email KILLspamlistservKILLspamspammitvma.mit.edu with SET PICList DIGEST in the body


2002\04\10@040029 by Jon Baker

flavicon
face
----- Original Message -----
From: "michael brown" <RemoveMEn5qmgTakeThisOuTspamAMSAT.ORG>
To: <spamBeGonePICLISTspamBeGonespamMITVMA.MIT.EDU>
Sent: Wednesday, April 10, 2002 3:28 AM
Subject: Re: [EE]: Scratchbuilt computer?


> >
> > > I always liked the motorola 6800 series.. especially the 6809.

> Weren't these the first micro's back then to have a hardware divide
> instruction?

Unfortunately not, they have a hardware multiply though - and have the
ability to run in multiprocessor systems, so I suppose you could add a co
processor if you wanted a challenge :)

--
Jon Baker

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email TakeThisOuTlistservEraseMEspamspam_OUTmitvma.mit.edu with SET PICList DIGEST in the body


2002\04\10@044314 by Kathy Quinlan

flavicon
face
Yes this can be done :o)

I designed and built a PC for my previous employer for embedded apps.

It was based on an AMD "all in one chip" It has 640K OR 1Mb of on board ram,
all peripherals were integrated, We had PC104 bus, 2 serial (needed MAX 232
chips) floppy and IDE port, PC keyboard, LCD display connector (for graphics
LCD) 2 connectors for matrix keyboards (we used 4 *1 and 4 * 4 pads).


All I did was point to point and a little bit of logic (max 232 and
regulator, some pull up networks)

Regards,

Kat.


____________________________________________________________________________
/"\   ASCII Ribbon Campaign  |        K.A.Q. Electronics
\ / - NO HTML/RTF in e-mail  | Software and Electronic Engineering
 X  - NO Word docs in e-mail |      Perth Western Australia
/ \                          |        Ph +61 419 923 731
____________________________________________________________________________

---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.344 / Virus Database: 191 - Release Date: 2/04/2002

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email RemoveMElistservspamTakeThisOuTmitvma.mit.edu with SET PICList DIGEST in the body


2002\04\10@054547 by Vit

picon face
> Yes this can be done :o)
>
> I designed and built a PC for my previous employer for embedded apps.
>
> It was based on an AMD "all in one chip" It has 640K OR 1Mb of on board
ram,
> all peripherals were integrated, We had PC104 bus, 2 serial (needed MAX
232
> chips) floppy and IDE port, PC keyboard, LCD display connector (for
graphics
> LCD) 2 connectors for matrix keyboards (we used 4 *1 and 4 * 4 pads).

Kathy,

I am curious.  :)  Can you tell us what was the speed of the processor?  And
what was the purpose of this computer?

Vitaliy

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listservEraseMEspam.....mitvma.mit.edu with SET PICList DIGEST in the body


2002\04\10@090330 by Douglas Butler

flavicon
face
If you can still get the parts, it might be fun to build your own
bit-slice CPU.  Design and build the registers, ALU, instruction
decoders, etc. yourself.  My friend did this in 1982 as his senior
project in college.  He used 4 bit bit-slice parts to make a 36 bit CPU.
The logic was static so he could single step the clock phases for
debugging.  I think his max clock speed was about 100kHz.  It could do a
36x36 multiply in the blink of an eye - if you didn't blink too fast!

Even simpler would be to make a 4004 class 4 bit CPU.  Maybe a good warm
up for the above.

Sherpa Doug

> {Original Message removed}

2002\04\10@092430 by Mark J. Dulcey

flavicon
face
Douglas Butler wrote:
> If you can still get the parts, it might be fun to build your own
> bit-slice CPU.  Design and build the registers, ALU, instruction
> decoders, etc. yourself.  My friend did this in 1982 as his senior
> project in college.  He used 4 bit bit-slice parts to make a 36 bit CPU.
>  The logic was static so he could single step the clock phases for
> debugging.  I think his max clock speed was about 100kHz.  It could do a
> 36x36 multiply in the blink of an eye - if you didn't blink too fast!

Neat idea - but I think all those parts are made of unobtanium now.
Bit-slice CPUs have certainly gone the way of the dodo. Even a lot of
the MSI TTL stuff that you would have used in a CPU design - ALUs and so
forth - have been discontinued. (You might be able to make replacements
for some of the functions out of PALs.) Bipolar PROMs - another mainstay
of such designs - are history, too, though CMOS parts have gotten fast
enough that they might be able to fill in - or, if not, you could use
SRAM that was loaded from ROM at boot time.

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email EraseMElistservspammitvma.mit.edu with SET PICList DIGEST in the body


2002\04\10@094451 by Alan B. Pearce

face picon face
>Neat idea - but I think all those parts are made of unobtanium now.
>Bit-slice CPUs have certainly gone the way of the dodo. Even a lot of
>the MSI TTL stuff that you would have used in a CPU design - ALUs and so
>forth - have been discontinued. (You might be able to make replacements
>for some of the functions out of PALs.) Bipolar PROMs - another mainstay
>of such designs - are history, too, though CMOS parts have gotten fast
>enough that they might be able to fill in - or, if not, you could use
>SRAM that was loaded from ROM at boot time.

I think you will find the most recent versions of the 2901 family are still
available. Great things to have around. In one of my past incarnations I
serviced to chip level CPU cards that had processors built around these. It
was interesting to see how the different generations of the CPU family had
things tweaked to get around the chip speed limitations to get the CPU cycle
time down.

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email RemoveMElistservEraseMEspamEraseMEmitvma.mit.edu with SET PICList DIGEST in the body


2002\04\10@103957 by John Dammeyer

flavicon
face
Truth be told if youi wanted to scratch build now you'd do it in VHDL or
Verilog in one of the FPGAs.

John

{Quote hidden}

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email RemoveMElistservTakeThisOuTspamspammitvma.mit.edu with SET PICList DIGEST in the body


2002\04\10@104732 by ted_rossin

flavicon
face
source= http://www.piclist.com/postbot.asp?id=piclist\2002\04\09\223306a

No divide but it had a multiply which was new at that time for cheap
microprocessors.


---

www.piclist.com/member/ted_rossin-agilent-com
PIC/PICList FAQ: http://www.piclist.com

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email EraseMElistservspamspamspamBeGonemitvma.mit.edu with SET PICList DIGEST in the body


2002\04\10@105139 by Kathy Quinlan

flavicon
face
> -----Original Message-----
> From: pic microcontroller discussion list
> [RemoveMEPICLISTKILLspamspamMITVMA.MIT.EDU]On Behalf Of Vit

> I am curious.  :)  Can you tell us what was the speed of the
> processor?  And
> what was the purpose of this computer?
> Hi all,

Here is the PDF for the AM486dx serries :

www.amd.com/epd/processors/6.32bitproc/11.am486fami/12.am486dx/20736/
20736.pdf

I think it was a DX40 we used, great chips, AMD still make alot of great PC
embedded uP's :o)


The purpose we had was for a front end to a 3 channel DSP system. It has
been replaced by an AVR with a 4 * 40 LCD, we chopped right back on the
final design, although I wish I still had one of these chips :o(


Regards,

Kat.


____________________________________________________________________________
/"\   ASCII Ribbon Campaign  |        K.A.Q. Electronics
\ / - NO HTML/RTF in e-mail  | Software and Electronic Engineering
 X  - NO Word docs in e-mail |      Perth Western Australia
/ \                          |        Ph +61 419 923 731
____________________________________________________________________________

---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.344 / Virus Database: 191 - Release Date: 2/04/2002

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listservSTOPspamspamspam_OUTmitvma.mit.edu with SET PICList DIGEST in the body


2002\04\10@110042 by Sean H. Breheny

face picon face
I'm not familiar with "bit-slice parts". What are these?

It would be a lot easier if you built your own CPU using an FPGA. That way,
you would eliminate the need to buy all the various different logic
functions and assemble them, you could make changes just by editing your
file in the design software, AND if you have decent software, you could
even simulate part or all of your design and see what is going on for
debugging purposes.

We did this for a simple 8 bit CPU way back in our first class on logic
design. It was fun and not all that difficult. You might even be able to
build something useful that you would want to use later on!

Sean

At 09:22 AM 4/10/02 -0400, you wrote:
{Quote hidden}

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email KILLspamlistservspamBeGonespammitvma.mit.edu with SET PICList DIGEST in the body


2002\04\10@112501 by Douglas Butler

flavicon
face
Bit-slice parts are parts that are meant to be stacked together to
create a larger function.  They generally came 4 bits wide.  If you
wanted a 24 bit ALU you would line up six of these 4 bit ALU chips.
They would have pins for the appropriate carries and borrows so they
would work together.  You could make a CPU of whatever width your data
required.

A Google search of "bit slice chips" will give lots more info.

Sherpa Doug

> {Original Message removed}

2002\04\10@115501 by Mark J. Dulcey

flavicon
face
Sean H. Breheny wrote:
> I'm not familiar with "bit-slice parts". What are these?

Bit slices were 4-bit logic units that were designed to be combined into
CPUs with wider word lengths. You also had to add PROM memory for the
microcode, and some glue to make it all stick together. Slices were
popular for a while, before it was possible to design high-power CPUs on
a single chip. Nobody designs computers that way now; if any bit slice
parts are still available, it's to allow continued production and repair
of old designs.

{Quote hidden}

If your goal is to learn how to design FPGAs, this sounds great, and
should be a useful exercise for people who are learning current system
design practice.

On the other hand, if you want a CPU that you can dig into with a scope,
it won't do at all, because all the workings will be hidden inside the
FPGA. You're no better off than you are if you build a system around a
stock microprocessor or microcontroller.

The lesson: first decide what your goals are. Then act accordingly.

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email EraseMElistservspamEraseMEmitvma.mit.edu with SET PICList DIGEST in the body


2002\04\10@123645 by Alan B. Pearce

face picon face
>Bit slices were 4-bit logic units that were designed to be
>combined into CPUs with wider word lengths.

Most of them seem to be 4 bits, but Intel had a range of 2 bit ones, using
the 3xxx part number range. A number of parts in this range were the same as
the 82xx peripheral parts, but bearing the 32xx part number, and possibly
some subtly different dynamic specs.

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email @spam@listserv@spam@spamspam_OUTmitvma.mit.edu with SET PICList DIGEST in the body


2002\04\10@130529 by Benjamin Bromilow

flavicon
face
From: "Douglas Butler" <spamBeGonedbutlerspamKILLspamIMETRIX.COM>
> the resources that this project would require available
> and can use them for fun...buy a  sports car.

>> ... or build a sports cars ;-)
>
> Sherpa Doug

Wow!
You read my mind. Ordering the bits today :)
I must agree with the answers along the lines of "why?". I agree that a
challenge is a challenge, "because it's there" etc. but after all the
motherboard companies have spent years if not decades perfecting their
knowledge and/or technique. The bits involved are also now very specialised
and complicated in terms of packaging. Perhaps a m/board for something
simpler like a XT would be possible but after recently retiring my old 286 I
can confirm that rumaging around with old computers and writing software for
them just to see "what can be done" looses it's attraction after a while....
I personally believe that making this m/board would
a) be virtually impossible
b) be very expensive
c1) require specialist PCB production services or else
c2) result in such a huge PCB that you'ld need a new house to store the
resultant PCB (the size of which alone would probably introduce problems
itself)...
After all, the joy of hobby EE/uC work  is either inventing new things or
making things cheaper than you can buy them. Surely this would be neither!
Just my 1p,

Ben

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email .....listservspam_OUTspammitvma.mit.edu with SET PICList DIGEST in the body


2002\04\10@130948 by gacrowell

flavicon
face
{Quote hidden}

Need to find a copy of Mick & Brick tho.  (Bit-Slice Microprocessor Design)

GC

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email .....listservspamRemoveMEmitvma.mit.edu with SET PICList DIGEST in the body


2002\04\10@133045 by John Dammeyer

flavicon
face
I have one of those on my bookshelf.  Great book.  Published by McGraw
Hill

>
> Need to find a copy of Mick & Brick tho.  (Bit-Slice
> Microprocessor Design)
>
> GC
>
> --
> http://www.piclist.com#nomail Going offline? Don't AutoReply us!
> email RemoveMElistservspamspamBeGonemitvma.mit.edu with SET PICList DIGEST in the body
>
>
>
>

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email spamBeGonelistserv@spam@spamspam_OUTmitvma.mit.edu with SET PICList DIGEST in the body


More... (looser matching)
- Last day of these posts
- In 2002 , 2003 only
- Today
- New search...