Re: POST Probes -> Was -> Re: Here's a plan

[Phil King]

> > MicroPose (MicroProse?) - do the Micro2000, which is a rather useful
> > POST probe.....
> >
> > It's not cheap, but well worth it for the usefulness, it is listening
> > address configurable, and will provide extra diagnostics like Osc check,
> > clock check tri state logic probe, and of course all the voltages....
>
> We *could* make our own. A POST card is an ultra-simple device, It
> simply latches data from a specific port reserved for POST (I forget
> which one off-hand though) and displays it via either a bank of 8 LEDs
> in binary or two 7-segment LED displays in hex.

...

If you're looking for a basic functionality POST card, I did a quickie web
search and came up with several.  Check out:

http://megacode.com/index.htm

http://www.datadepo.com/mpost.htm

http://www.micro2000.com/postprobe.htm

http://www.quicktech.com/qppcplus.htm

They vary wildly in price, but I found a local distributor for the megacode
ISA-Bus only product and went out this afternoon to pick one up for US$30
(+ $2.45 local sales tax.).  A lot less painful than rolling my own...

And the standard POST port is apparently 0x80.

Cheers!

Phil King
    				[e-mail]

POST Probes -> Was -> Re: Here's a plan

[David J. Coffin]

> If you're looking for a basic functionality POST card, I did a quickie web
> search and came up with several.  Check out:
>
> http://megacode.com/index.htm
> http://www.datadepo.com/mpost.htm
> http://www.micro2000.com/postprobe.htm

These three products are hardwired to show only port
0x80.  That's good enough for technicians, but engineers
need more.  You might be able to change the port number by
cutting address lines and inserting inverters.
> http://www.quicktech.com/qppcplus.htm

This card can be set for *any* ISA port 0x0 to 0x3ff,
although the web page doesn't explain how.  Unfortunately,
this company's phone number has been disconnected.

Here's why having POST cards on two or three different
ports is so important:

%macro		print	1-2
    	push	eax
    	push	dx
    	mov	ax, %1
    	out	84h, al
    	mov	dx, 300h
    	mov	al, ah
    	out	dx, al
%if %0 > 1
    	mov	al, %2
    	out	80h, al
%endif
    	mov	eax, 600000	; Change this to adjust delay
%%loop:		dec	eax
    	jnz	%%loop
    	pop	dx
    	pop	eax
%endmacro
--

When writing code, you insert something like this:

print ax,0xaa		; Print 16-bit values
print dx,cl
print si,0x51
print ds,0xd5
print [memval],0x22

ror ebx,16		; Print a 32-bit register
print bx,0xb1
ror ebx,16
print bx,0xb2
--

It takes some trial and error to get the delay count
right.  It depends on your CPU, and whether the BIOS is
running from ROM or RAM (RAM is about 30 times faster).

Early POST code runs with no stack, because DRAM is
not yet ready.  Any print statements made in stackless
code will destroy EAX and DX.

Dave Coffin  3/11/99
--
David Coffin		Evening: 781-397-9932
967 Salem Street	Daytime: 978-686-6468x341
Malden, MA 02148-4515	E-mail:  [e-mail]

POST Probes -> Was -> Re: Here's a plan

[James Oakley]

"David J. Coffin" wrote:
> > http://www.quicktech.com/qppcplus.htm
>  This card can be set for *any* ISA port 0x0 to 0x3ff,
>
> although the web page doesn't explain how.  Unfortunately,
> this company's phone number has been disconnected.
>  Here's why having POST cards on two or three different
>
> ports is so important:

Another 2 reasons why building our own would be a good idea.

1) Company not available/reliable

2) a microcontroller can monitor multiple ports and display them all to
an lcd or multiple 7-seg displays mounted on front of the PC so we won't
have to look inside the case (I have a small flashlight next to my
computer for checking jumper settings, part numbers, etc..) all the
time.

Also, Matthew Sullivan mentioned it would be a good idea to have a "dual
bios" on the card that could be switched back and forth.  What if we
flash into the main BIOS code that would jmp to whatever bios is enabled
on the card, a backup bios and a test bios. That would *really* speed up
development time. The switch could go on front of the pc with the
display. It all could be mounted in a 5 1/4" drive bay cover.

This is all, of course, very simple design.

It would be great to have a post card with such a level of intelligence.

James Oakley

POST Probes -> Was -> Re: Here's a plan

[Matthew Sullivan]

Hi,

James Oakley wrote:
> "David J. Coffin" wrote:
> > > http://www.quicktech.com/qppcplus.htm
> >  This card can be set for *any* ISA port 0x0 to 0x3ff,
> >
> > although the web page doesn't explain how.  Unfortunately,
> > this company's phone number has been disconnected.
> >  Here's why having POST cards on two or three different
> >
> > ports is so important:
>
> Another 2 reasons why building our own would be a good idea.
>
> Company not available/reliable
>
> a microcontroller can monitor multiple ports and display them all to
>
> an lcd or multiple 7-seg displays mounted on front of the PC so we won't
> have to look inside the case (I have a small flashlight next to my
> computer for checking jumper settings, part numbers, etc..) all the
> time.
>
> Also, Matthew Sullivan mentioned it would be a good idea to have a "dual
> bios" on the card that could be switched back and forth.  What if we
> flash into the main BIOS code that would jmp to whatever bios is enabled
> on the card, a backup bios and a test bios. That would *really* speed up
> development time. The switch could go on front of the pc with the
> display. It all could be mounted in a 5 1/4" drive bay cover.

Whilst this is a good idea, it wasn't quite what i was thinking about, for
the only reason ... wouldn't this give us flashing problems, or can you
also arrange to be able to flash to the '2nd' BIOS ..? (I said I was not
much good with digital electronics design ;-))

However, assuming that one BIOS can be programmed, and the other maybe
read-only? any design ideas/plans etc...  If this works, I'd like to get my
hands on a plan ASAP cause this would be just what I need...
> This is all, of course, very simple design.

I hope so ;-)
> It would be great to have a post card with such a level of intelligence.

If you're carful you could even make a business plan & money from it... but
more on that when we see what we've got....

Yours

Matthew

AW: POST Probes -> Was -> Re: Here's a plan

[Stephan Müller]

> Whilst this is a good idea, it wasn't quite what i was thinking about, for
> the only reason ... wouldn't this give us flashing problems, or can you
> also arrange to be able to flash to the '2nd' BIOS ..? (I said I was not
> much good with digital electronics design ;-))
>
> However, assuming that one BIOS can be programmed, and the other maybe
> read-only? any design ideas/plans etc...  If this works, I'd like
> to get my
> hands on a plan ASAP cause this would be just what I need...
> > This is all, of course, very simple design.
>
> I hope so ;-)

The simplest method to switch between 2 Flash-memories is to connect both of
them and switch the CHIPENABLE-pin between the two flashs. Only the Flash
with the CE enabled will respond to requests. So there can be more than one
flash at the same address. Switching is done by connecting the
mainboard-CE-Connector to either one of the two flashs (CE-pin of them). A
simple switch can do this. And probably there can be more than 2 flashs the
same time connected (not enabled).
If somebody is interested in developing such a system, he can contact me, we
can try to do this, cheap and efficient for openBIOS-developers.

Ciao,
    Stephan [SpaceNet-Systems]

AW: POST Probes -> Was -> Re: Here's a plan

[James Oakley]

Stephan Müller wrote:
> The simplest method to switch between 2 Flash-memories is to connect both of
> them and switch the CHIPENABLE-pin between the two flashs. Only the Flash
> with the CE enabled will respond to requests. So there can be more than one
> flash at the same address. Switching is done by connecting the
> mainboard-CE-Connector to either one of the two flashs (CE-pin of them). A
> simple switch can do this. And probably there can be more than 2 flashs the
> same time connected (not enabled).

Definitely, not a problem at all.
> If somebody is interested in developing such a system, he can contact me, we
> can try to do this, cheap and efficient for openBIOS-developers.

What we need to find out is exactly what everybody wants and probably
make one "OpenBIOS Development Device" that has the most popular
features.

What we need to know is what everybody wants, do we want just the
"multiple flash" device, or do we want POST functionality as well. If we
want POST, do we want regular or our own "enhanced" POST method.

I really do think we all need POST if we plan to do any programming. I
do embedded systems which often do not have displays on them, I usually
hook up LEDs, LCDs, or serial ports for debugging purposes only. They
tend to save one's behind.

James Oakley
[e-mail]

AW: POST Probes -> Was -> Re: Here's a plan

[Qiwei Xiao]

Stephan M�ller wrote:
> The simplest method to switch between 2 Flash-memories is to connect both of
> them and switch the CHIPENABLE-pin between the two flashs. Only the Flash
> with the CE enabled will respond to requests. So there can be more than one
> flash at the same address. Switching is done by connecting the
> mainboard-CE-Connector to either one of the two flashs (CE-pin of them). A
> simple switch can do this. And probably there can be more than 2 flashs the
> same time connected (not enabled).
> If somebody is interested in developing such a system, he can contact me, we
> can try to do this, cheap and efficient for openBIOS-developers.

I actually made such a board and it has been working pretty well. The schematic 
is exactly the same as the above. It uses a 32-pin cable connect to the BIOS
socket on the MB. The cable has to be short (less than 2 inches), especially
for fast Pentium. The switch has to be mechanical. (You might have thought
using some kind of TTL gate to switch the CE pin, this doesn't work, at least
for me, because of the delay it introduces). If anybody would like more 
information, let me know so I can post details. (actually, there isn't too
much detail. Stephan said it all. )

As you still remember, I posted the EPROM simulator circuit to this list 
some months ago. I actually managed to build the simulator and made it to 
work!!! I did made some change to the original circuit. Again, let me know
if you are interested.

Cheers,
Qiwei Xiao 
[e-mail]

EPROM-simulator, FLASH evaluation

[Stephan Mueller]

> As you still remember, I posted the EPROM simulator circuit to this list
> some months ago. I actually managed to build the simulator and made it to
> work!!! I did made some change to the original circuit. Again, let me know
> if you are interested.

i'm interested in the circuit.

Ciao,
    Stephan [SpaceNet-Systems]

EPROM-simulator, FLASH evaluation

[James Oakley]

Stephan Mueller wrote:
> > As you still remember, I posted the EPROM simulator circuit to this list
> > some months ago. I actually managed to build the simulator and made it to
> > work!!! I did made some change to the original circuit. Again, let me know
> > if you are interested.
>
> i'm interested in the circuit.

Me too!

Sorry, just had to do that :-)

We should probably put stuff like this in a central place, or at least
link to it from a centralised place. This would eliminate a lot of
confusion to newcomers who may think something hasn't been done yet when
it is.

POST Probes -> Was -> Re: Here's a plan

[James Oakley]

Matthew Sullivan wrote:
> Whilst this is a good idea, it wasn't quite what i was thinking about, for
> the only reason ... wouldn't this give us flashing problems, or can you
> also arrange to be able to flash to the '2nd' BIOS ..? (I said I was not
> much good with digital electronics design ;-))

Does anybody know how, electrically, the flash is erased and
reprogrammed? None of my books with BIOS stuff in them mention this. Is
it a MEMW or an IOW line? Either way, if you remove the BIOS and place
two of them on an appropriate ISA card with a switch on their enable
pins, it should be fine (just be careful of the switch position before
doing any flashing, of course).
> However, assuming that one BIOS can be programmed, and the other maybe
> read-only?

Simple, don't tie whatever line it is (MEMW/IOW) that enables writes to
the flash to the backup bios. And clearly label which socket is which on
the silk screen :-)
> any design ideas/plans etc...  If this works, I'd like to get my
> hands on a plan ASAP cause this would be just what I need...

Once a list of features is finalised, it's probably a matter of a couple
of hours to design the schematic and board (I have an autorouter at
work, we've done PCI boards before). Then we can figure how much one of
these babies will cost us each (most expensive is the board itself,
though there are some cheap solutions out there now, such as ExpressPCB
(~US$100 for a couple of boards, most of which is the setup fee,
additional boards cost much less)).

We also could etch the boards ourselves, but that way lies madness. It
sounds cheap at first but you don't want to try to draw an edgecard
connector with a marker (Rat Shack style), and not everyobody has access
to a UV oven or a CNC machine.
> If you're carful you could even make a business plan & money from it... but
> more on that when we see what we've got....

Nah, no market for that outside of people like us, I prefer the
non-profit approach anyway (why do you think I hang around here?). I
propose that any additional electronic design be part of the OpenBIOS
project, like what is done by many Linux packages that have specific
hardware.

Is anybody else interested? Speak now... please :-)

James Oakley
[e-mail]

POST Probes -> Was -> Re: Here's a plan

[Stephen Rust]

Hi James,

You wanted to know if other people are interested.  I am one that
is interested.  I just recently joined the list and am interested in
becoming involved in the coding.  I will soon have a test machine to play
with.  I do'nt have much experience in this sort of thing but am very
interested in learning and trying to help out where I can.  So as far as
what the board needs on it, your idea with needing POST in order to help
program it, makes sense to me.

Steve

On Sun, 14 Mar 1999, James Oakley wrote:
> Matthew Sullivan wrote:
> > Whilst this is a good idea, it wasn't quite what i was thinking about, for
> > the only reason ... wouldn't this give us flashing problems, or can you
> > also arrange to be able to flash to the '2nd' BIOS ..? (I said I was not
> > much good with digital electronics design ;-))
>
> Does anybody know how, electrically, the flash is erased and
> reprogrammed? None of my books with BIOS stuff in them mention this. Is
> it a MEMW or an IOW line? Either way, if you remove the BIOS and place
> two of them on an appropriate ISA card with a switch on their enable
> pins, it should be fine (just be careful of the switch position before
> doing any flashing, of course).
> > However, assuming that one BIOS can be programmed, and the other maybe
> > read-only?
>
> Simple, don't tie whatever line it is (MEMW/IOW) that enables writes to
> the flash to the backup bios. And clearly label which socket is which on
> the silk screen :-)
> > any design ideas/plans etc...  If this works, I'd like to get my
> > hands on a plan ASAP cause this would be just what I need...
>
> Once a list of features is finalised, it's probably a matter of a couple
> of hours to design the schematic and board (I have an autorouter at
> work, we've done PCI boards before). Then we can figure how much one of
> these babies will cost us each (most expensive is the board itself,
> though there are some cheap solutions out there now, such as ExpressPCB
> (~US$100 for a couple of boards, most of which is the setup fee,
> additional boards cost much less)).
>
> We also could etch the boards ourselves, but that way lies madness. It
> sounds cheap at first but you don't want to try to draw an edgecard
> connector with a marker (Rat Shack style), and not everyobody has access
> to a UV oven or a CNC machine.
> > If you're carful you could even make a business plan & money from it... but
> > more on that when we see what we've got....
>
> Nah, no market for that outside of people like us, I prefer the
> non-profit approach anyway (why do you think I hang around here?). I
> propose that any additional electronic design be part of the OpenBIOS
> project, like what is done by many Linux packages that have specific
> hardware.
>
> Is anybody else interested? Speak now... please :-)
>
> James Oakley
> [e-mail]

AW: POST Probes -> Was -> Re: Here's a plan

[Stephan Mueller]

> Does anybody know how, electrically, the flash is erased and
> reprogrammed? None of my books with BIOS stuff in them mention this. Is
> it a MEMW or an IOW line? Either way, if you remove the BIOS and place
> two of them on an appropriate ISA card with a switch on their enable
> pins, it should be fine (just be careful of the switch position before
> doing any flashing, of course).

Take a look at www.amd.com
They sell flash-memories (and of course, CPUs :) There you can get a
PDF-File for their Flash-chips.

Ciao,
    Stephan [SpaceNet-Systems]

AW: POST Probes -> Was -> Re: Here's a plan

[James Oakley]

Stephan Mueller wrote:
> > Does anybody know how, electrically, the flash is erased and
> > reprogrammed? None of my books with BIOS stuff in them mention this. Is
> > it a MEMW or an IOW line? Either way, if you remove the BIOS and place
> > two of them on an appropriate ISA card with a switch on their enable
> > pins, it should be fine (just be careful of the switch position before
> > doing any flashing, of course).
>
> Take a look at www.amd.com
> They sell flash-memories (and of course, CPUs :) There you can get a
> PDF-File for their Flash-chips.

Ok, I just went over a motherboard with a continuity tester. The /WE
line on the flash is connected to /MEMW. The /OE line is connected to
/MEMR. The flash functions as simple RAM memory. Vpp (on this
motherboard, a Eurone 486 board) is connected to the middle pin of a
3-pin header. The other two pins are 5V and 12V, a jumper is installed
to connect it to either voltage. The /CE line, however, is connected to
the UMC UM8886 ISA bridge controller. I assume this is the case with all
these chipsets. I'll check to make sure.

So, with any board we make, we're probably going to have to take that
line from the flash socket (it's not available on the bus, obviously).
No prob, just use a socket with all the appropriate pins soldered to a
ribbon cable, which connects to whatever board we use, or make a board
which plugs into the flash socket (bad idea, the flash is often in a
rather tight place).

So this is not a problem, we could switch the /CE line for selecting
each chip, connect /MEMW and Vpp only to the test flash, and connect the
rest of the lines (Ax, Dx, /MEMR, Vss, Vcc) to both.

Told you it was easy.

James Oakley
[e-mail]

connection of the flash-devices

[Stephan Mueller]

> Ok, I just went over a motherboard with a continuity tester. The /WE
> line on the flash is connected to /MEMW. The /OE line is connected to
> /MEMR. The flash functions as simple RAM memory. Vpp (on this
> motherboard, a Eurone 486 board) is connected to the middle pin of a
> 3-pin header. The other two pins are 5V and 12V, a jumper is installed
> to connect it to either voltage. The /CE line, however, is connected to
> the UMC UM8886 ISA bridge controller. I assume this is the case with all
> these chipsets. I'll check to make sure.

The CE, if i can remember, is connected to the 74 254 (an address-decoder).
The 74 254 is connected to the ISA-Bus. However, perhaps the UMC has the 254
integrated. I dunno. :)
> So, with any board we make, we're probably going to have to take that
> line from the flash socket (it's not available on the bus, obviously).
> No prob, just use a socket with all the appropriate pins soldered to a
> ribbon cable, which connects to whatever board we use, or make a board
> which plugs into the flash socket (bad idea, the flash is often in a
> rather tight place).
> So this is not a problem, we could switch the /CE line for selecting
> each chip, connect /MEMW and Vpp only to the test flash, and connect the
> rest of the lines (Ax, Dx, /MEMR, Vss, Vcc) to both.
> Told you it was easy.

I've got an experimention-Board (prototype-board) to put devices
directly in a socket and build connections. I'll try to build a flash-switch
myself. If i'll be successfull i'll put the wire-diagram on my page. :)

Ciao,
    Stephan [SpaceNet-Systems]

Flash'n'chips

> I've got an experimention-Board (prototype-board) to put devices
> directly in a socket and build connections. I'll try to build a flash-switch
> myself. If i'll be successfull i'll put the wire-diagram on my page. :)

Or even better: The OpenBIOS page! :)

Just want to drop in a 'me too'...:) I'm also very interested in a dual
flash-board, and a post-monitor...

My motherboard is a Asus tx97-x (intel 430-tx?)
I have a symbios ncr-scsiadapter on it. (+ a ide disk) I found the scsi bios on
the net, maybe that one could be used... Anyway, the ide disk will be the first
one to get to work.
I also have an old P90, but I don't remember which chipset it have, I'll check
it out. Think it is a noname MB with intel chipset.

Karl Erik

-----------------------------------------
Karl Erik Asbjornsen  [e-mail]
    http://www.samfundet.no/karlea
   P: [phone]  M: [phone]