VOGONS


Reply 20 of 39, by fsinan

User metadata
Rank Newbie
Rank
Newbie
CoffeeOne wrote on 2024-02-09, 16:49:
rasz_pl wrote on 2024-02-09, 15:28:

by Chkcpu analysis:
>Apart from this extra routine, these BIOSes are identical.
you wont gain anything by flashing official bios

That's what I wanted to say, too. Complete overkill to get a programmer and buy a new chip in this case.
Changing to the official BIOS does not hurt is true, but only if you have everything at hand.

But I will buy some real cache chips. I'm asking for that case.

System:1
Cyrix 5x86-100GP
U-Board ST1A with "0"KB Cache 😀
System:2
AMD K6-2-475(Changing frequently with Cyrix 6x86MX PR-233)
TMC MI5VP4
System:3
UMC U5S-40
486UL-P101

Reply 21 of 39, by fsinan

User metadata
Rank Newbie
Rank
Newbie
Chkcpu wrote on 2024-02-08, 19:09:
I agree, the official 2A4X5H03C BIOS should work fine and will indicate the actual L2 cache size. With the fake chips it will di […]
Show full quote
mkarcher wrote on 2024-02-08, 17:14:
fsinan wrote on 2024-02-08, 14:33:

WOW Great find, thank you. Mine is "2A4X5000C"...So it is fake...What happens if I change the bios? Does it boot? Later I'll change the chips.

Changing to the official BIOS does not hurt. Of course the message "256KB Cache" during boot will disappear as long as no cache is installed, but the official BIOS is fully prepared to work on a board without any working cache chips. You might get a BIOS complaint like "CACHE MEMORY BAD - DO NOT ENABLE CACHE", but IIRC that's a message on earlier BIOSes, and most likely AMI. If I remember wrong and you get that message with the original BIOS, just disable the L2 cache and you are fine.

I agree, the official 2A4X5H03C BIOS should work fine and will indicate the actual L2 cache size. With the fake chips it will display “None”. 😉

Because the 2A4X5000C only “fixes” the L2 cache size display to 256KB and doesn’t display any cache warning messages, I believe the 2A4X5H03C BIOS won’t complain about the fake L2 cache either.

Note that your BIOS is stored in a D27C010 128KB UV-EPROM chip and can’t be flashed.
Do you have an (E)EPROM programmer to program a new chip? It is usually easier to get a new 27C010 or equivalent (E)EPROM chip to program the official BIOS in, and keep the old BIOS chip as backup. It saves the hassle of erasing the old chip with UV-light first.

Cheers, Jan

Can I change D27C010 to a flashable bios chip?

System:1
Cyrix 5x86-100GP
U-Board ST1A with "0"KB Cache 😀
System:2
AMD K6-2-475(Changing frequently with Cyrix 6x86MX PR-233)
TMC MI5VP4
System:3
UMC U5S-40
486UL-P101

Reply 22 of 39, by mkarcher

User metadata
Rank l33t
Rank
l33t
fsinan wrote on 2024-02-09, 18:23:

Can I change D27C010 to a flashable bios chip?

Yes. Any usual 5V 128K x 8 flash chip will at least work like a ROM in that board. I don't know whether that board supports in-system flashing or you need to program the chip using a programmer or a different computer and doing a hot-swap flash.

Reply 23 of 39, by rasz_pl

User metadata
Rank l33t
Rank
l33t
fsinan wrote on 2024-02-09, 18:20:

But I will buy some real cache chips. I'm asking for that case.

yes, in that case "you wont gain anything by flashing official bios". Your fake bios will work fine with real cache.

Open Source AT&T Globalyst/NCR/FIC 486-GAC-2 proprietary Cache Module reproduction

Reply 25 of 39, by fsinan

User metadata
Rank Newbie
Rank
Newbie

Removed the fake chips. Still reporting 256K. Real ones on the way.

System:1
Cyrix 5x86-100GP
U-Board ST1A with "0"KB Cache 😀
System:2
AMD K6-2-475(Changing frequently with Cyrix 6x86MX PR-233)
TMC MI5VP4
System:3
UMC U5S-40
486UL-P101

Reply 26 of 39, by fsinan

User metadata
Rank Newbie
Rank
Newbie
rasz_pl wrote on 2024-02-10, 00:18:
fsinan wrote on 2024-02-09, 18:20:

But I will buy some real cache chips. I'm asking for that case.

yes, in that case "you wont gain anything by flashing official bios". Your fake bios will work fine with real cache.

Today I installed a set of srams total of 512. Set the jumper, reporting 512 but did not boot into DOS, stalls after bios info screen before dos boot. Changed the set of SRAMS again it does not boot. When disabling cache from bios it boots.

Either I installed two sets of bad srams or there is something wrong with this "fake cache" bios.

Should I change the bios? What can I do to understand that these chips are real with multimeter?

System:1
Cyrix 5x86-100GP
U-Board ST1A with "0"KB Cache 😀
System:2
AMD K6-2-475(Changing frequently with Cyrix 6x86MX PR-233)
TMC MI5VP4
System:3
UMC U5S-40
486UL-P101

Reply 28 of 39, by nuno14272

User metadata
Rank Member
Rank
Member

Are the sram chips purchased from a "legit" place ?

because theres an large amount of fake chacke modules being sold everywhere... fuuny when replacing fake cache with not working cache..

1| 386DX40
2| P200mmx, Voodoo 1
3| PIII-450, Voodoo 3 3000

Reply 29 of 39, by fsinan

User metadata
Rank Newbie
Rank
Newbie
weedeewee wrote on 2024-02-09, 17:02:
Chkcpu wrote on 2024-02-08, 19:09:
I agree, the official 2A4X5H03C BIOS should work fine and will indicate the actual L2 cache size. With the fake chips it will di […]
Show full quote
mkarcher wrote on 2024-02-08, 17:14:

Changing to the official BIOS does not hurt. Of course the message "256KB Cache" during boot will disappear as long as no cache is installed, but the official BIOS is fully prepared to work on a board without any working cache chips. You might get a BIOS complaint like "CACHE MEMORY BAD - DO NOT ENABLE CACHE", but IIRC that's a message on earlier BIOSes, and most likely AMI. If I remember wrong and you get that message with the original BIOS, just disable the L2 cache and you are fine.

I agree, the official 2A4X5H03C BIOS should work fine and will indicate the actual L2 cache size. With the fake chips it will display “None”. 😉

Because the 2A4X5000C only “fixes” the L2 cache size display to 256KB and doesn’t display any cache warning messages, I believe the 2A4X5H03C BIOS won’t complain about the fake L2 cache either.

Note that your BIOS is stored in a D27C010 128KB UV-EPROM chip and can’t be flashed.
Do you have an (E)EPROM programmer to program a new chip? It is usually easier to get a new 27C010 or equivalent (E)EPROM chip to program the official BIOS in, and keep the old BIOS chip as backup. It saves the hassle of erasing the old chip with UV-light first.

Cheers, Jan

Does the modified bios display the correct amount of cache when real cache is present, ie 64, 128, 256, 512K, or will it always display the fake 256K amount ?

Tried with 128K and 512K config, it shows the correct amount. But it hangs after "Starting MS-DOS"

System:1
Cyrix 5x86-100GP
U-Board ST1A with "0"KB Cache 😀
System:2
AMD K6-2-475(Changing frequently with Cyrix 6x86MX PR-233)
TMC MI5VP4
System:3
UMC U5S-40
486UL-P101

Reply 30 of 39, by CoffeeOne

User metadata
Rank Oldbie
Rank
Oldbie
fsinan wrote on 2024-03-21, 17:36:
weedeewee wrote on 2024-02-09, 17:02:
Chkcpu wrote on 2024-02-08, 19:09:
I agree, the official 2A4X5H03C BIOS should work fine and will indicate the actual L2 cache size. With the fake chips it will di […]
Show full quote

I agree, the official 2A4X5H03C BIOS should work fine and will indicate the actual L2 cache size. With the fake chips it will display “None”. 😉

Because the 2A4X5000C only “fixes” the L2 cache size display to 256KB and doesn’t display any cache warning messages, I believe the 2A4X5H03C BIOS won’t complain about the fake L2 cache either.

Note that your BIOS is stored in a D27C010 128KB UV-EPROM chip and can’t be flashed.
Do you have an (E)EPROM programmer to program a new chip? It is usually easier to get a new 27C010 or equivalent (E)EPROM chip to program the official BIOS in, and keep the old BIOS chip as backup. It saves the hassle of erasing the old chip with UV-light first.

Cheers, Jan

Does the modified bios display the correct amount of cache when real cache is present, ie 64, 128, 256, 512K, or will it always display the fake 256K amount ?

Tried with 128K and 512K config, it shows the correct amount. But it hangs after "Starting MS-DOS"

Did you buy ISSI61C1024 chips? One of those has 128kB. Tons of it are sold on Ebay, they are sold in 10ns and 15ns. 10ns is 100% re-labelled, because it never existed in this speed rating, only 12, 15, 20ns.
I lately bought 20 pieces of it from a Chinese seller, 8 of it were dead.
Do you have TL866II-Plus programmer? It is has a SRAM test functionality.
So when you test it with that and it fails, then you know 100% that the chip is not working.

Unfortunately when a SRAM chip passes the test, you don't know really, if it works, because it does not test the timing of the chip.
But still you can easily filter out the completely broken ones, so it is a great speedup in finding good chips. So I have now 12 chips and have to check, whether they are really good or not, .....

Reply 31 of 39, by fsinan

User metadata
Rank Newbie
Rank
Newbie
nuno14272 wrote on 2024-03-21, 17:21:

Are the sram chips purchased from a "legit" place ?

because theres an large amount of fake chacke modules being sold everywhere... fuuny when replacing fake cache with not working cache..

Bought them from ebay reseller, from china. Declared as new but these are not new pieces as I understand. Properly printed above and below and they seem to be legit.

Attachments

System:1
Cyrix 5x86-100GP
U-Board ST1A with "0"KB Cache 😀
System:2
AMD K6-2-475(Changing frequently with Cyrix 6x86MX PR-233)
TMC MI5VP4
System:3
UMC U5S-40
486UL-P101

Reply 32 of 39, by rasz_pl

User metadata
Rank l33t
Rank
l33t

Bios cache detection might be performed using slowest timings (the case in two bioses I disassembled), then bios applies your speed settings and its no longer stable.
You can try lowering FSB and setting slowest timings.

Open Source AT&T Globalyst/NCR/FIC 486-GAC-2 proprietary Cache Module reproduction

Reply 33 of 39, by CoffeeOne

User metadata
Rank Oldbie
Rank
Oldbie
fsinan wrote on 2024-03-21, 18:23:
nuno14272 wrote on 2024-03-21, 17:21:

Are the sram chips purchased from a "legit" place ?

because theres an large amount of fake chacke modules being sold everywhere... fuuny when replacing fake cache with not working cache..

Bought them from ebay reseller, from china. Declared as new but these are not new pieces as I understand. Properly printed above and below and they seem to be legit.

I don't know if it is really interesting, but:
Those were the 20chips, I bought:

issi61C1024.jpg
Filename
issi61C1024.jpg
File size
940.59 KiB
Views
236 views
File license
Fair use/fair dealing exception

The upper 12 passed the SRAM test, the other 8 failed it.
When you have one that fails this test, it will not work, not even on 20MHz FSB. If you have a mainboard that supports such a low FSB. But most likely your chips are good in principle, but the can't keep up with the tight timing.
You should try 3-2-2-2 read and 3T write as a first thing always.

Reply 34 of 39, by weedeewee

User metadata
Rank l33t
Rank
l33t
CoffeeOne wrote on 2024-03-21, 18:32:
I don't know if it is really interesting, but: Those were the 20chips, I bought: issi61C1024.jpg […]
Show full quote
fsinan wrote on 2024-03-21, 18:23:
nuno14272 wrote on 2024-03-21, 17:21:

Are the sram chips purchased from a "legit" place ?

because theres an large amount of fake chacke modules being sold everywhere... fuuny when replacing fake cache with not working cache..

Bought them from ebay reseller, from china. Declared as new but these are not new pieces as I understand. Properly printed above and below and they seem to be legit.

I don't know if it is really interesting, but:
Those were the 20chips, I bought: issi61C1024.jpg

The upper 12 passed the SRAM test, the other 8 failed it.
When you have one that fails this test, it will not work, not even on 20MHz FSB. If you have a mainboard that supports such a low FSB. But most likely your chips are good in principle, but the can't keep up with the tight timing.
You should try 3-2-2-2 read and 3T write as a first thing always.

Have you tried the cotton swab with nail polish remover(acetone) test on the printing of the IC ? best to try on the known broken ones first 😉

though maybe considering the amount of scratches on them, it's not necessary.

Right to repair is fundamental. You own it, you're allowed to fix it.
How To Ask Questions The Smart Way
Do not ask Why !
https://www.vogonswiki.com/index.php/Serial_port

Reply 35 of 39, by CoffeeOne

User metadata
Rank Oldbie
Rank
Oldbie
weedeewee wrote on 2024-03-21, 18:36:
CoffeeOne wrote on 2024-03-21, 18:32:
I don't know if it is really interesting, but: Those were the 20chips, I bought: issi61C1024.jpg […]
Show full quote
fsinan wrote on 2024-03-21, 18:23:

Bought them from ebay reseller, from china. Declared as new but these are not new pieces as I understand. Properly printed above and below and they seem to be legit.

I don't know if it is really interesting, but:
Those were the 20chips, I bought: issi61C1024.jpg

The upper 12 passed the SRAM test, the other 8 failed it.
When you have one that fails this test, it will not work, not even on 20MHz FSB. If you have a mainboard that supports such a low FSB. But most likely your chips are good in principle, but the can't keep up with the tight timing.
You should try 3-2-2-2 read and 3T write as a first thing always.

Have you tried the cotton swab with nail polish remover(acetone) test on the printing of the IC ? best to try on the known broken ones first 😉

though maybe considering the amount of scratches on them, it's not necessary.

No. But the broken ones are not interesting really. I am not interested, if they are fake or just not working. The seller refunded quickly the money for those. I wish they would (at least) test it with a TL866II by themselves, how hard it would it be?

Reply 37 of 39, by fsinan

User metadata
Rank Newbie
Rank
Newbie

I've tried my 10 512K chips not sstematically but tried so many changes.

And tried 128K umc chips config with 10 chips with many configs, loosened the timings ...

I think there is a problem with Mboard, thats why they HAD to use fake chips with this one. To be able to solve it without problem even if l2 cache is enabled in bios.

Thats my idea.

System:1
Cyrix 5x86-100GP
U-Board ST1A with "0"KB Cache 😀
System:2
AMD K6-2-475(Changing frequently with Cyrix 6x86MX PR-233)
TMC MI5VP4
System:3
UMC U5S-40
486UL-P101

Reply 38 of 39, by CoffeeOne

User metadata
Rank Oldbie
Rank
Oldbie
fsinan wrote on 2024-03-21, 19:42:
I've tried my 10 512K chips not sstematically but tried so many changes. […]
Show full quote

I've tried my 10 512K chips not sstematically but tried so many changes.

And tried 128K umc chips config with 10 chips with many configs, loosened the timings ...

I think there is a problem with Mboard, thats why they HAD to use fake chips with this one. To be able to solve it without problem even if l2 cache is enabled in bios.

Thats my idea.

What do you mean with 512K chips and 128K chips???

Reply 39 of 39, by fsinan

User metadata
Rank Newbie
Rank
Newbie
CoffeeOne wrote on 2024-03-21, 20:09:
fsinan wrote on 2024-03-21, 19:42:
I've tried my 10 512K chips not sstematically but tried so many changes. […]
Show full quote

I've tried my 10 512K chips not sstematically but tried so many changes.

And tried 128K umc chips config with 10 chips with many configs, loosened the timings ...

I think there is a problem with Mboard, thats why they HAD to use fake chips with this one. To be able to solve it without problem even if l2 cache is enabled in bios.

Thats my idea.

What do you mean with 512K chips and 128K chips???

I'm talking about configs.

4 of 1024 chips and a 256k controller is 512K chips config.

5 of 256K chips(one is cont) is 128K chips config.

Attachments

System:1
Cyrix 5x86-100GP
U-Board ST1A with "0"KB Cache 😀
System:2
AMD K6-2-475(Changing frequently with Cyrix 6x86MX PR-233)
TMC MI5VP4
System:3
UMC U5S-40
486UL-P101