VOGONS


First post, by amigasith

User metadata
Rank Newbie
Rank
Newbie

As I wrote in this thread here, I'm a proud owner of a Biostar MB-8433UUD-A mainboard since a couple of weeks. I am now experimenting with different BIOSes and came across something that I cannot explain. Perhaps one of the gurus on this board can help me out.

I flashed feipoa's latest 2014 BIOS onto a 39SF010A flash chip from Microchip. It's brand new and not a new old stock one. Here is a picture:

39SF010A.jpg
Filename
39SF010A.jpg
File size
51.03 KiB
Views
599 views
File license
Fair use/fair dealing exception

UNIFLASH recognizes the flash chip type correctly and both flashing and verifying work just fine. All the rectangles on the progress bar are green. And the funny things is that board also POSTs correctly with the newly flashed chip, but only up to a certain point. It always hangs at

Updating ESCD ...

and that's it. It doesn't even react to Ctrl-Alt-Del anymore and I have to hard-reset. It's like the ESCD area on the chip contains garbage and the BIOS wants to update ESCD at every boot. But then it can't write to the ESCD area for some reason (?). That's my theory at the moment.

feipoa mentions in his great 8433 manual revision 2.0 that depending on the flash type, the "...memory address where the Plug and Play ESCD data is stored" can be different between different flash chip types. But this one here is definitely a 5V one according to its datasheet and as I said, flash and verify works just fine.

I tried some old versions of AWDFLASH that I found that still support the /f switch to force updating the BIOS with whatever file you provide to it. And I also added the /cp switch in order to clear PnP / ESCD, but for some reason, all old AWDFLASH versions that I tried just hang and don't even start writing to the flash chip, although they are recognizing the type of flash chip correctly (39SF010A).

--> Does anybody have an idea of what could be going wrong here?

Perhaps a tip regarding a specific version of AWDFLASH? Perhaps using MODBIN to change the memory address of the ESCD area, if that's possible at all?

Reply 1 of 10, by Horun

User metadata
Rank l33t++
Rank
l33t++

Unfortunately I have a board that does similar (FIC KA31) and halts on "updating ESCD..." and not been able to diagnose it yet.
Curious if you have a speaker hooked up ? Mine gives a constant beep sound (more like a whine ;p) about one second after the updating ESCD.

Hate posting a reply and then have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor. Stuff: https://archive.org/details/@horun

Reply 2 of 10, by maxtherabbit

User metadata
Rank l33t
Rank
l33t

I read somewhere the hanging at ESCD thing can be caused by trying to use EDO ram on a chipset that does not support it.

OP's chipset does not support EDO afaik

Reply 3 of 10, by Horun

User metadata
Rank l33t++
Rank
l33t++

Thanks Max ! Next time I play with mine will try a bunch of diff ram, maybe is a ram issue. The KA31 uses standard pc100/133 DIMMS but maybe I was using some incompatible dimms.

Hate posting a reply and then have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor. Stuff: https://archive.org/details/@horun

Reply 5 of 10, by amigasith

User metadata
Rank Newbie
Rank
Newbie

In my case, I fear it's not as easy as this...

I am already using FPM RAM and not EDO. And unfortunately, there is no "Reset Configuration Data" option or anything similar in the BIOS menus that you could use to suppress updating ESCD.

I tried another flash chip - an SST PH29EE010-150-3CF - and this one works just fine. So it must be something related to the 39SF010A...

Reply 6 of 10, by snufkin

User metadata
Rank Oldbie
Rank
Oldbie
amigasith wrote on 2021-05-21, 06:08:

In my case, I fear it's not as easy as this...

I am already using FPM RAM and not EDO. And unfortunately, there is no "Reset Configuration Data" option or anything similar in the BIOS menus that you could use to suppress updating ESCD.

I tried another flash chip - an SST PH29EE010-150-3CF - and this one works just fine. So it must be something related to the 39SF010A...

I've only read the device summaries, but the 29EE010 looks to have a 128 bytes pages and handles the erase/write cycle internally. The 39SF010 has 4kB sectors and the user has to erase a sector before writing. At a guess the 39SF010 never signals the write cycle has completed, so the boot process hangs. Even if it did continue, the ESCD data would likely be corrupted.

Reply 7 of 10, by amigasith

User metadata
Rank Newbie
Rank
Newbie

D'oh, that could really be the culprit, thank you very much snufkin. Seems like I haven't read the datasheets thoroughly enough.

Does anybody know a way to suppress the updating of the ESCD data? MODBIN to the rescue? As I said, unfortunately there is no "Reset Configuration Data" option or anything similar in the BIOS menus.

Reply 8 of 10, by amigasith

User metadata
Rank Newbie
Rank
Newbie

I managed to get by the "Updating ESCD ... " hang when using a 39SF010A flash chip, but it's more of a hack than a solution 😉 But see for yourself...

I used UNIFLASH to save the BIOS from a working system (using the SST PH29EE010-150-3CF flash chip) into a file. Then I flashed this saved BIOS (and not the original UUD2014.BIN file) onto the 39SF010A flash chip and it worked like a charm. No "Updating ESCD ... " message anymore and therefore also no hang. Apparently, when UNIFLASH saves a BIOS to file, it also saves the ESCD data alongside the actual BIOS.

This approach, however, only works if nothing related to your Plug'n'Play data ever changes, since this inevitably leads to an update of the ESCD. So for example, if I only put the VGA card into another PCI slot, then the BIOS sees a change in P'n'P data and consequently wants to update ESCD. This - of course - again hangs when using the 39SF010A flash chip.

So all in all, you can only use the described method when (a) you have a working setup that you can save to file using UNIFLASH and (b) you're not going to change anything regarding P'n'P. Obviously not very practical...

Nevertheless, this was a nice little side project 😀

Reply 9 of 10, by snufkin

User metadata
Rank Oldbie
Rank
Oldbie
amigasith wrote on 2021-05-23, 21:15:

So all in all, you can only use the described method when (a) you have a working setup that you can save to file using UNIFLASH and (b) you're not going to change anything regarding P'n'P. Obviously not very practical...

Really putting the 'not' into PnP.

Any reason (other than the obvious one of cost) of not getting a 29EE010? There seem to be a few available on ebay.

Reply 10 of 10, by amigasith

User metadata
Rank Newbie
Rank
Newbie
snufkin wrote on 2021-05-23, 22:08:

Any reason (other than the obvious one of cost) of not getting a 29EE010? There seem to be a few available on ebay.

Not really 😀

This whole little project here started just because I got hold of the 39SF010A flash chip at a decent price, it was brand new, and I thought "let's give it a try". I think the best way to look at this thread now is to see it as a kind of "tried it, doesn't work" experience in case anybody else is thinking about giving it a try.