bad flash corrupted bios

[kc8adu]

Re: bad flash corrupted bios

Sigh... I like the old boards with a nice big DIP socket. As if PLCC's weren't annoying enough, now they don't even put a socket on them anymore. I still don't see the logic in soldering those things directly to the board - those sockets in quantity are worth just a few cents, and save lots of labor when it comes time to repair a bad flash.

you think the beancounters care?

[starfury1]

Re: bad flash corrupted bios

yeah few cents saved does add up across a few thousand MB's

Not that I know for sure,
but guess the life cycle of a MB in the manufacture eyes is maybe 3 years
1 year to out live the warranty
2 years for it to fail at some point so you go and buy a new one...
by that time things have changed in the computer world enough and you would probably need one.

just my thoughts

cheers

[gdement]

Re: bad flash corrupted bios

you think the beancounters care?

It just seems that, one way or another, the money saved on labor would come back to the manufacturer's benefit. Brand name clients like Dell, for example, will find themselves repairing customer's bad flashes fairly often. How easily it can be fixed doesn't really change what the customer will be willing to pay for the service. So if it's a quick fix, that leaves more money in their pocket from the repair bill. So why don't they spec PLCC sockets from the board OEM?

Ironically, it's the well-supported and expensive boards from brand name PC's which most often don't have sockets. Hell, some of them even use those stupid SOIC things. Maybe they honestly think they'd lose repair business if people could fix their own bad flashes, but I doubt there's more than a tiny fraction of people that are capable of doing that. The business logic just escapes me - no sockets only raises their labor costs, it doesn't really change what the customer will pay them to fix it.

[zandrax]

Re: bad flash corrupted bios

It just seems that, one way or another, the money saved on labor would come back to the manufacturer's benefit. Brand name clients like Dell, for example, will find themselves repairing customer's bad flashes fairly often. How easily it can be fixed doesn't really change what the customer will be willing to pay for the service. So if it's a quick fix, that leaves more money in their pocket from the repair bill. So why don't they spec PLCC sockets from the board OEM?

Dell would be so happy to charge the customer for a refurbished board
A cent saved is a cent earned, and some hundreds of cents saved sparing on "not so necessary" components such as bios sockets, ps2 ports, ram slots etc. reduce manifacturing costs (but not always sale price...): if a customer screws something, he knows he has to bring his pc to a technician, wait a couple of days and pay a lot for getting it back.
I did it once with a Compaq Presario when I was green in pc world (second brand pc, the first one was an Olivetti M240): now no more.

Zandrax

[davmax]

Re: bad flash corrupted bios

Asus Mobo warranty is 3 years. I have not yet discovered the Gigabyte warranty.

[starfury1]

Re: bad flash corrupted bios

ok 4 years then for Asus
yes your right Davmax

even if socketed you still need a programmer and the bios firmware to flash
sockets can sometimes cause problems too.
(not all that often but I have had it happen)

but I see your point. gdement

Cheers

[Tom41]

Re: bad flash corrupted bios

I was thinking that you could maybe use an external EEPROM programmer, and rig up something to connect it to the chip while it's still on the board. I once made something similar with a chip tester (one that clips onto the chip and shows you the signals through LEDs) and wired it into an EEPROM reader/programmer. I could then dump and rewrite chips without desoldering them from the board.

Though I'm not sure they make a chip tester like that for PLCC chips, and even if they did soldering all those pins would be a nightmare!

[starfury1]

Re: bad flash corrupted bios

I thought about that but normally with the old stuff you need to take a pin high or low to enable programing ...then you got all the circuits hanging off it so this could pose a problem or even possible damage but I suppose it depend on whats what really as it is programmed on board as its hooked up

I wouldn't have a clue on the pinouts of the chip so no idea.

the above method sound like the best option if you can get the chip and program the BIOS into it in a programmer then just follow the above.

I guess its a matter of a programmer accepting the BIOS file...hopefully it just a ascii type file the Bios file its self I meant not the flasher program...then that open the next question is it supplied as seperate file and flash program...if its rolled into one then I guess you some what stuffed unless as you say, read in on board or you can pull one so you can take a dump of the file from a working bios.

still it would be nice to be able just to clip on and flash from a programmer
I know very little about programmers and chip of today

cheers

[JorgeHernan]

Re: bad flash corrupted bios

You can solder two PLCC socket`s together, so you can put one over the soldered chip, and into the other you can put in a reprogrammed chip.
After booting the system up with a DOS BIOS flashing tool, remove the adapter and flash the on board chip again.

ECS support has provided such a kit for people with this kind of problem.

That's what i was thinking about but didn't know how to do it.Excellent

[gdement]

Re: bad flash corrupted bios

You can solder two PLCC socket`s together, so you can put one over the soldered chip, and into the other you can put in a reprogrammed chip.
After booting the system up with a DOS BIOS flashing tool, remove the adapter and flash the on board chip again.

ECS support has provided such a kit for people with this kind of problem.

I didn't get it at first, but I think now I understand what you're saying. I don't see how this would work though - the 2 chips would be in conflict with each other and the one that's "right" wouldn't always win the argument. You might just end up with something like a logical "OR" of the 2 ROMs. Not sure which logic level "1" or "0" overpowers the other, but either way it would just lead to a corrupted data output from the combined ROMs.
But if ECS is doing this, I must be missing something.

[Harvey]

Re: bad flash corrupted bios

^^ Thats what was/is confusing me too. I'm guessing its somehow handled by using the Chip-Select line to tristate the onboard BIOS.. but exactly how is a mystery.

[starfury1]

Re: bad flash corrupted bios

ok
In my haste I forgot that...one very important part...how does it use one then the other with the top hat flash...so good catch

If you look at the other pdfs Paul S provided there is some mode for this so maybe I just got that stuck in my mind

heres (what little) wiki there is on Top Hat flash

http://www.coreboot.org/Top_Hat_Flash

what it references to is BIOS Saviour

here what that Wiki says

A BIOS Saviour is a device that plugs into a PLCC socket, and provides another one on top. In between there is an additional Flash ROM chip, and a switch to select either that chip or the one sitting in the provided top socket. The purpose of this construct is to minimize the amount of hot swapping required, to reduce mechanical and electrical stress on the BIOS chips.

So I guess the TOP HAT must have some why of doing this
(again it probably depends on the "type" of BIOS chip and may also depend on the MB)

anyway here is the Install BIOS Saviour

so now go have a look at the Selection chart

So its possible but you may have to mod the idea to what your true situation is

The above still doesn't explain exactly how the "Top Hat Flash" does it
but it must do something along these lines..if not something more.

found comment on it at OCW

Top-Hat Flash is ECS's patented version of of Dual BIOS. The top hat flash is a handy external BIOS recovery that sits on top just like a hat. For example when you flashed a BIOS and the system can't boot, the top hat flash can be inserted on top of the existing soldered on BIOS (as shown below). Power up the machine and it should boot as per normal. Once it is in DOS prompt, remove the top hat flash and flash the on board bios with the correct BIOS code and reboot machine. Now you have a working board again.

Page Here

So guess by the use of the word "patented" your not going to easily find exactly how they do it.

so its "Ask not how the trick was done but how would you do it?"

so yeah it must disable the dud on BIOS while booting and once THF is removed the dud bios is re-enabled and now you are running DOS for the flasher your set to go

how you do this is specific to the actual chip its self I suppose
(and as I said may also depend on the MB)

Seems this device is specific to that MB
(or possibly range of MB's)

seems to be possible if you can nut it out with that MB.

anyway me no idea...would have to start with the specific chip info I guess.

Cheers

[starfury1]

Re: bad flash corrupted bios

Seems the Above is WRONG!!!!
with regard to how the TOP HAT FLASHER WORKS

Put it on the board BIOS chip aligned correctly, power system up, for a few seconds until it beeps, power it down, remove the unit, and you have a restored BIOS. .

It actually "Flashes" the dud onboard BIOS

SEE this page first update, bout 1/3 down page

As Max would say Sorry about that Chief!!!!

still it must have some way to disable dud so it can boot itself up into the program then re enable (in programming mode) for flashing

this of course puts a whole new slant on it
and your back to ripping the bios chip off the MB!

OK back to the drawing board

[kikkoman]

Re: bad flash corrupted bios

It actually "Flashes" the dud onboard BIOS

no, it doesn't.

from the ECS KN1 manual:



(i have that board, but they forgot to include the TopHat Flash, so I can't tell you more about it.)

The RD1 BIOS Saviour needs the switch because it is left in the socket during the flash (I guess to minimize risk of damaging anything). Since hotflashing works perfectly if you're careful, it's just a matter of convenience in my opinion.

I am 99% sure that it's simply a paralleled chip. as long as the original EEPROM isn't physically damaged, its looks like its I/O pins are left floating, otherwise it wouldn't work.

[starfury1]

Re: bad flash corrupted bios

Humm OK then that must have been a brief..like he didn't mention the full procedure.

I must say that threw me when I read it and I didn't stop to re-read the bit I had posted above on booting to DOS in the quote


So you got a manual.....(more or less what I was trying to find) thanks
that basically confirms the first post

it boots it to the point of loading Dos

good, then we are back to how does it do this while in parallel,
I agree it must be
but now is a pin being switched some how ? to force as harvey said a tristate state or a disable of some kind?

Had I had look at the REF 1 on the wiki for the THF its shows the MB marked as THF Here

so has the MB been designed to enable this switching with the use THF
possibly by using a pull up (or down) resistor ?

it appears there is a pcb between the 2 sockets this would enable them to have a high or low on the right pin, as you say Harvey ..chip select, chip enable or disable

Once the THF is removed the resistor just takes it to the opposite state
(its an old trick)

normally if your never intending to disable something you would just tie it were it needs to go ground or +
(and here is were you could have a problem with any other MB)

anyway just my thoughts and pure guess work

Have to see if I can ID the chip and get a PDF

cheers

[kikkoman]

Re: bad flash corrupted bios

think about it again. pull-up/-down resistors are useless if you use two chips in parallel because they would affect both chips.

it would definitely require some active logic circuitry to determine which chip to use if it's no parallel solution. however all i can see is the usual bunch of resistors.

i attached the best pic i could find, which still makes me believe that there's no magic behind it.
if i can get the vendor to send me a replacement, i'll take it apart.

edit: the EEPROm on my MB is a PM49FL-004
(datasheet)

[kikkoman]

Re: bad flash corrupted bios

google is your friend. here's part of the solution:

http://www.mail-archive.com/linuxbio.../msg07266.html

if you search that site for "top-hat flash", there are some more results that may be interesting.

[Paul S]

Re: bad flash corrupted bios

So, what I have gathered, this method still requires a chip with a working BIOS on it.

How can you flash the second chip with the BIOS? Use another MB and piggy-back on that one?

[Harvey]

Re: bad flash corrupted bios

while the ECS mainboard would have like a voltage divider feeding */OE-/Init*

Hmmm. That could work.

Feed the onboard BIOS chips ENABLE pin from a resistor on the motherboard. That would make no difference as its a high impedance input.

The Tophat could then hold that pin in Disable mode simply by it being tied to a low resistance drive (i.e. direct to 5V assuming its active low). The resistor on the PCB stops the line being shorted and burning out whatever is driving it normally.

The ENABLE pin on the tophat's chip is then tied LOW to make that chip active.

Et voilà! - the new BIOS chip appears and the old one disappears.

Boot PC, rip off tophat, flash old chip, have beer


However - it all assumes that shorting the ENABLE pin isn't going to cause damage... it may be that ECS always do that - I wouldn't like to say other manufacturers do.

[Harvey]

Re: bad flash corrupted bios

So, what I have gathered, this method still requires a chip with a working BIOS on it.

How can you flash the second chip with the BIOS? Use another MB and piggy-back on that one?

You buy it from ECS, with adaptor, ready flashed.

Or you could flash it in another board I suppose - boot working PC, fit Tophat, flash, remove tophat.