Mamba wrote on 2024-12-15, 20:06:

I do not think so.
The chips seems original, it is not easy to replace the writings on it.
Plus, the programmer would spot differences or errors. It is not a basic ch341a thing.

If these used intel EEPROMs from China are genuine and exactly the same model as the original chip, then there should be no POSTing problems related to the freshly flashed BIOS.

Once I thought "there's no reason in all these EEPROM relabeling fraud as these chips costs pennies..." until I got some sanded and relabeled Winbond chip instead of purchased SST one - at least it turned out they're interchangeable models. Recently some of these chinese sellers have started to successfully relabel even laser engraved labels/markings.

As I'm looking at the picture of the original chip (the one in your first post) and the picture of the chinese chip (from your last post), I see a couple of differences in the labeling style.

Mamba wrote on 2024-12-15, 20:06:

There is not use in trying a different bios.

There is, but you have to get the Atmel chip. Or just in case if one of those chinese intel chips is relabeled ATMEL AT49LH004 EEPROM.

This thing is making me mad.

Tried Uniflash, hotflashing.
I know it is a nonsense since I have the programmer but since the situation is so strange, I try strange things.

Uniflash is not able to recognise the chip, all three of them.
This could an evidence that they are false.
But I read on the net that Supermicro boards have some stubborn system to “protect” firmware.
So I tried using -force parameter with related code for the chip.

Of course I obtained an error in flashing.

The PM49FL004T (that works) differs from the Intel for the pin 1. At least this is what my programmer says. It is possible that the pin 1 status is the source of protection for Intel?

So there are two possibilities:

- All three chips are false, no matter what my Xgecu reads.

- The Supermicro board has sone protection system of the firmware that is beyond my comprehension and not documented at all.

Can someone with experience with this board help?
I know is a really rare junk.

analog_programmer wrote on 2024-12-15, 21:13:

If these used intel EEPROMs from China are genuine and exactly the same model as the original chip, then there should be no POST […]
Show full quote

Mamba wrote on 2024-12-15, 20:06:

I do not think so.
The chips seems original, it is not easy to replace the writings on it.
Plus, the programmer would spot differences or errors. It is not a basic ch341a thing.

If these used intel EEPROMs from China are genuine and exactly the same model as the original chip, then there should be no POSTing problems related to the freshly flashed BIOS.

Once I thought "there's no reason in all these EEPROM relabeling fraud as these chips costs pennies..." until I got some sanded and relabeled Winbond chip instead of purchased SST one - at least it turned out they're interchangeable models. Recently some of these chinese sellers have started to successfully relabel even laser engraved labels/markings.

As I'm looking at the picture of the original chip (the one in your first post) and the picture of the chinese chip (from your last post), I see a couple of differences in the labeling style.

Mamba wrote on 2024-12-15, 20:06:

There is not use in trying a different bios.

There is, but you have to get the Atmel chip. Or just in case if one of those chinese intel chips is relabeled ATMEL AT49LH004 EEPROM.

Let’s say that all three are false. How can they pass the Xgecu ID and pin check?

Mamba wrote on 2024-12-15, 21:26:

Let’s say that all three are false. How can they pass the Xgecu ID and pin check?

I do have a Xgecu T48 programmer and I also have a list of programmer's supported chips in which some of the (E)EPROMs from a different brands/makes use one and the same chip ID.

I also know that some Xgecu users successfully managed to "flash" a 128 kB BIOS image on a 64 kB EEPROM without a sign of any errors (the values are just an example for different capacity). And of course, there is no way to run a half flashed BIOS in half EEPROM capacity, but the ghost chasing begins because of the improper EEPROM model and the "silent" Xgecu software 😁 This is just an example that you can't rely on some chinese software to protect you from such an elementary errors.

I always think logically (or at least I'm trying to), so if there's no problems with the chinese EEPROM chips, the BIOS dump file and your Xgecu flashing routine, then I have a bad news for your non-working motherboard.

Mamba wrote on 2024-12-15, 21:25:

The PM49FL004T (that works) differs from the Intel for the pin 1. At least this is what my programmer says. It is possible that the pin 1 status is the source of protection for Intel?

Intel N82802AB pin 1:
"Block Erase/Program Power Supply. For erasing array blocks or
programming data. V PP = 3.3 V or 12 V V PP. With V PP ≤ V PPLK,
memory contents cannot be altered. Attempting a block erase or
program with an invalid V PP (see DC Characteristics) will produce
spurious results and should not be attempted. V PP may only be held
at 12 V for 80 hours over the lifetime of the device."

PMC PM49FL004T pin 1 is marked as "not connected". Are you sure that PM49FL004T is compatible with N82802AB?

analog_programmer wrote on 2024-12-15, 21:50:

I do have a Xgecu T48 programmer and I also have a list of programmer's supported chips in which some of the (E)EPROMs from a di […]
Show full quote

Mamba wrote on 2024-12-15, 21:26:

Let’s say that all three are false. How can they pass the Xgecu ID and pin check?

I do have a Xgecu T48 programmer and I also have a list of programmer's supported chips in which some of the (E)EPROMs from a different brands/makes use one and the same chip ID.

I also know that some Xgecu users successfully managed to "flash" a 128 kB BIOS image on a 64 kB EEPROM without a sign of any errors (the values are just an example for different capacity). And of course, there is no way to run a half flashed BIOS in half EEPROM capacity, but the ghost chasing begins because of the improper EEPROM model and the "silent" Xgecu software 😁 This is just an example that you can't rely on some chinese software to protect you from such an elementary errors.

I always think logically (or at least I'm trying to), so if there's no problems with the chinese EEPROM chips, the BIOS dump file and your Xgecu flashing routine, then I have a bad news for your non-working motherboard.

Mamba wrote on 2024-12-15, 21:25:

The PM49FL004T (that works) differs from the Intel for the pin 1. At least this is what my programmer says. It is possible that the pin 1 status is the source of protection for Intel?

Intel N82802AB pin 1:
"Block Erase/Program Power Supply. For erasing array blocks or
programming data. V PP = 3.3 V or 12 V V PP. With V PP ≤ V PPLK,
memory contents cannot be altered. Attempting a block erase or
program with an invalid V PP (see DC Characteristics) will produce
spurious results and should not be attempted. V PP may only be held
at 12 V for 80 hours over the lifetime of the device."

PMC PM49FL004T pin 1 is marked as "not connected". Are you sure that PM49FL004T is compatible with N82802AB?

The datasheet says that it is read compatible with Intel firmware.

And with the bios flashed in that chip, the motherboard boots.
I have several problems (no agp, seems to work with a Pentium II only, probably more), that is why I was searching for an Intel chip.
It is the meaning of this topic (see first page).

I see. Let's assume that those chinese intel EEPROMs are (most likely) relabeled non-intel chips, so you don't have another compatible EEPROM chip except this PMC PM49FL004T and the mobo is kinda alive while using it. Then if there is no some compatibility conflict with your AGP videocard (as PC Hoarder Patrol mentioned here), it seems like the problem is not only in the original damaged EEPROM chip - maybe there's also damaged chipset or some other component(s). It is quite possible that in the past this motherboard has suffered from higher voltage due to a faulty PSU and as a result not only the original BIOS chip has been damaged.

analog_programmer wrote on 2024-12-15, 22:50:

I see. Let's assume that those chinese intel EEPROMs are (most likely) relabeled non-intel chips, so you don't have another compatible EEPROM chip except this PMC PM49FL004T and the mobo is kinda alive while using it. Then if there is no some compatibility conflict with your AGP videocard (as PC Hoarder Patrol mentioned here), it seems like the problem is not only in the original damaged EEPROM chip - maybe there's also damaged chipset or some other component(s). It is quite possible that in the past this motherboard has suffered from higher voltage due to a faulty PSU and as a result not only the original BIOS chip has been damaged.

I am more inclined to think that the bios is not correctly loaded or/from the bios chip since there is not pin1 compatibility.
In fact the system says every time “update nvram…..failed”.

Maybe sone Rom sections are not read and this is why I cannot put a coppermine or use an agp card.
Tried a voodoo 3000 also. No luck

I need a reliable bios chip

Do not know if it is related or not but here another one with problems properly flashing (hot flashing) a Intel firmware:

https://www.wimsbios.com/forum/motherboard-de … erts-t9577.html

It seems like there is sone fail safe process that prevents correct flashing if the correct flash program is not used.

Still, if someone know about a reliable Intel 80802ab source, PLEASE message me.

Mamba wrote on 2024-12-16, 05:56:

In fact the system says every time “update nvram…..failed”.

This happens when BIOS is not 100% compatible with the EEPROM chip. You're using PMC PM49FL004T (ID: 9D6E) instead of intel N82802AB (ID: 89AD) and obviously their chip IDs differ.

Mamba wrote on 2024-12-16, 06:42:

It seems like there is sone fail safe process that prevents correct flashing if the correct flash program is not used.

It doesn't matter when dedicated EEPROM chip programmer is used.

That is why I need a proper Intel chip.
But I only see them from China.
I do not want to waste my time again.

analog_programmer wrote on 2024-12-16, 09:23:

This happens when BIOS is not 100% compatible with the EEPROM chip. You're using PMC PM49FL004T (ID: 9D6E) instead of intel N828 […]
Show full quote

Mamba wrote on 2024-12-16, 05:56:

In fact the system says every time “update nvram…..failed”.

This happens when BIOS is not 100% compatible with the EEPROM chip. You're using PMC PM49FL004T (ID: 9D6E) instead of intel N82802AB (ID: 89AD) and obviously their chip IDs differ.

Mamba wrote on 2024-12-16, 06:42:

It seems like there is sone fail safe process that prevents correct flashing if the correct flash program is not used.

It doesn't matter when dedicated EEPROM chip programmer is used.

Exactly what I thought.

As for a good source, it's lottery.
I bought pristine new 256kbit chips that are sanded and reprinted. Work on some boards, don't on others.
Guess what? ID doesn't match.

Mamba wrote on 2024-12-16, 09:38:

That is why I need a proper Intel chip.

Or you may kindly ask kmeaw to edit once more the EEPROM chip ID check procedure in the BIOS, so that it doesn't give an error when using PMC PM49FL004T (ID: 9D6E).

analog_programmer wrote on 2024-12-16, 14:31:

Mamba wrote on 2024-12-16, 09:38:

That is why I need a proper Intel chip.

Or you may kindly ask kmeaw to edit once more the EEPROM chip ID check procedure in the BIOS, so that it doesn't give an error when using PMC PM49FL004T (ID: 9D6E).

It seems he cannot

Maybe not much relevant, but did anyone else notice the ES (Engineering Sample) label on the Intel chip from China?

Mamba wrote on 2024-12-16, 14:58:

It seems he cannot

He made it for the Atmel chip, so he can also make it for the PMC. A matter of BIOS unpacking-repacking and four bytes alteration for the chip ID.

PcBytes wrote on 2024-12-16, 18:01:

Maybe not much relevant, but did anyone else notice the ES (Engineering Sample) label on the Intel chip from China?

Good catch. It didn't cross my mind that the two letters "ES" are an abbreviation and it could mean "engineering sample". Actually this clearly explains the problems with all three chinese chips.

Update,
I have received a legit Intel chip and the board do not show "nvram update.... failed" message anymore.

Mamba wrote on 2025-01-03, 09:40:

Update,
I have received a legit Intel chip and the board do not show "nvram update.... failed" message anymore.

Finally!
Enjoy your board.

Nexxen wrote on 2025-01-03, 10:11:

Mamba wrote on 2025-01-03, 09:40:

Update,
I have received a legit Intel chip and the board do not show "nvram update.... failed" message anymore.

Finally!
Enjoy your board.

Actually now comes the hard part:
Make the AGP works.
I was partially correct, the proper bios chip allows me to make the agp port work but only with a AGP 3.3V card apparently.
It is ok with a voodoo3000.

With GeForce2 GTS, GeForce fx5200 and fx 5700 I have the board stuck at “D3-14” post code.

Mamba wrote on 2025-01-03, 10:50:

Actually now comes the hard part: Make the AGP works. I was partially correct, the proper bios chip allows me to make the agp po […]
Show full quote

Nexxen wrote on 2025-01-03, 10:11:

Mamba wrote on 2025-01-03, 09:40:

Update,
I have received a legit Intel chip and the board do not show "nvram update.... failed" message anymore.

Finally!
Enjoy your board.

Actually now comes the hard part:
Make the AGP works.
I was partially correct, the proper bios chip allows me to make the agp port work but only with a AGP 3.3V card apparently.
It is ok with a voodoo3000.

With GeForce2 GTS, GeForce fx5200 and fx 5700 I have the board stuck at “D3-14” post code.

Mai una gioia. 🤣

Nexxen wrote on 2025-01-03, 10:51:

Mamba wrote on 2025-01-03, 10:50:

Actually now comes the hard part: Make the AGP works. I was partially correct, the proper bios chip allows me to make the agp po […]
Show full quote

Nexxen wrote on 2025-01-03, 10:11:

Finally!
Enjoy your board.

Actually now comes the hard part:
Make the AGP works.
I was partially correct, the proper bios chip allows me to make the agp port work but only with a AGP 3.3V card apparently.
It is ok with a voodoo3000.

With GeForce2 GTS, GeForce fx5200 and fx 5700 I have the board stuck at “D3-14” post code.

Mai una gioia. 🤣

🤣