Re: POST Probes -> Was -> Re: Here's a plan
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 pking@netschools.net
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.
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: -- ; Print a word value to ports 300h/84h and an optional ; tag to 80h. Give the user two seconds to read it. %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: -- %include "debug.mac" 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: dcoffin@shore.net
"David J. Coffin" wrote:
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
Hi, James Oakley wrote:
"David J. Coffin" wrote:
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.
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
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]
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 jfunk@roadrunner.nf.net
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 qiwei@lucent.com
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]
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. -- James Oakley jfunk@roadrunner.nf.net
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 jfunk@roadrunner.nf.net
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 jfunk@roadrunner.nf.net
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]
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 jfunk@roadrunner.nf.net
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]
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 karlea@samfundet.no http://www.samfundet.no/karlea P: +47 64874397 M: +47 90738663
participants (9)
-
David J. Coffin -
James Oakley -
karl-erik.asbjornsen＠thomson-csf.no -
Matthew Sullivan -
Phil King -
Qiwei Xiao -
Stephan Mueller -
Stephan Müller
-
Stephen Rust