weedeewee wrote on 2023-12-06, 18:26:

Nevermind the 24 (A11) then, I think it's only used for 16MB simms and the board doesn't support those, so it's likely there is […]
Show full quote

Nevermind the 24 (A11) then, I think it's only used for 16MB simms and the board doesn't support those, so it's likely there is no trace whatsoever.
On the backside photo of the board I see pin 19 (A10) is linked for four slots, but seperate and linked for the other four slots.

The datasheet for the 74f86n tells us pin 3 is an input output for an XOR or gate. It should be connected somewhere else as well.

Since 1MB simms are working, try to figure out where pin 18 (A9) from the simm slots goes to.
It's also possible it goes to multiple locations, so if you find one continuity, keep looking for a possible other location.

GigAHerZ, thanks for the input. Always good to have another pov.

edit: the other bank is connected to pin 6 of the 74f86n, which is also an output for an XOR or gate.
pin 1 & 2 are the inputs for pin3, pin 4 & 5 are the inputs for pin 6
It could be that the chip 74f86 is defective, or that a trace of its input is broken somewhere.

Pin 18 goes to U21 (74F244N), 5th pin from left and below. Didn't find any other traces....

butjer1010 wrote on 2023-12-06, 19:09:

weedeewee wrote on 2023-12-06, 18:26:

Nevermind the 24 (A11) then, I think it's only used for 16MB simms and the board doesn't support those, so it's likely there is […]
Show full quote

Nevermind the 24 (A11) then, I think it's only used for 16MB simms and the board doesn't support those, so it's likely there is no trace whatsoever.
On the backside photo of the board I see pin 19 (A10) is linked for four slots, but seperate and linked for the other four slots.

The datasheet for the 74f86n tells us pin 3 is an input output for an XOR or gate. It should be connected somewhere else as well.

Since 1MB simms are working, try to figure out where pin 18 (A9) from the simm slots goes to.
It's also possible it goes to multiple locations, so if you find one continuity, keep looking for a possible other location.

GigAHerZ, thanks for the input. Always good to have another pov.

edit: the other bank is connected to pin 6 of the 74f86n, which is also an output for an XOR or gate.
pin 1 & 2 are the inputs for pin3, pin 4 & 5 are the inputs for pin 6
It could be that the chip 74f86 is defective, or that a trace of its input is broken somewhere.

Pin 18 goes to U21 (74F244N), 5th pin from left and below. Didn't find any other traces....

pin 16 of U21 (74F244N) that's the output of pin 4 of U21 (74F244N). 74F244N is a buffer, for the A9 signal.
I was kinda hoping it would be near to A10.

weedeewee wrote on 2023-12-06, 19:00:

Really? from the photo you posted with the trace of A10 I would've sworn you have only 64Kilobyte of cache on the board. Sorry, […]
Show full quote

butjer1010 wrote on 2023-12-06, 18:57:

GigAHerZ wrote on 2023-12-06, 18:21:

Just 2 cents. […]
Show full quote

Just 2 cents.

I have the same board, i have MR-BIOS on it (THE BEST BIOS!!!) and i run it with 8x4MB SIMMs. The SIMM modules are 3-chip variants, so with higher density chips. Works perfectly.
I didn't have any memory issues with original bios either.

So just make sure that you don't have any damage between chipset and RAM sockets. Also, sometimes boards may behave weirdly, regarding the RAM, when the real issue is with cache. (Settings or damage or whatever) So make sure everything related to cache is also in good order.

I did add 128KB cache more than the original state! So it could be problem with cache? I will try to revert to 128kB and check! Thanks

Really? from the photo you posted with the trace of A10 I would've sworn you have only 64Kilobyte of cache on the board.
Sorry, that is photo from internet, not mine board. I will take a picture of mine, sorry
edit: oh well, maybe that's the other board you have. 😀

butjer1010 wrote on 2023-12-06, 19:48:

weedeewee wrote on 2023-12-06, 19:00:

butjer1010 wrote on 2023-12-06, 18:57:

I did add 128KB cache more than the original state! So it could be problem with cache? I will try to revert to 128kB and check! Thanks

Really? from the photo you posted with the trace of A10 I would've sworn you have only 64Kilobyte of cache on the board.
edit: oh well, maybe that's the other board you have. 😀

Sorry, that is photo from internet, not mine board. I will take a picture of mine, sorry

no problem, I just remembered I have a m321 board, though I think it's a rev 3, with 128k cache soldered on, and cpu in qfp soldered on, and 16? or 32M of ram.

weedeewee wrote on 2023-12-06, 20:04:

no problem, I just remembered I have a m321 board, though I think it's a rev 3, with 128k cache soldered on, and cpu in qfp soldered on, and 16? or 32M of ram.

Could it be the problem in Cache? Do i need U32 for 256KB?. I can remove chips, and leave only 4 chips, with IS61C64AH-20N in U17 (state before upgrade to 256), but can that be the fault with RAM?

butjer1010 wrote on 2023-12-07, 13:59:

weedeewee wrote on 2023-12-06, 20:04:

no problem, I just remembered I have a m321 board, though I think it's a rev 3, with 128k cache soldered on, and cpu in qfp soldered on, and 16? or 32M of ram.

Could it be the problem in Cache? Do i need U32 for 256KB?. I can remove chips, and leave only 4 chips, with IS61C64AH-20N in U17 (state before upgrade to 256), but can that be the fault with RAM?

My board also has no chip present in U32, so, I don't think that's the problem.
But... did you have to desolder old cache chips in order to upgrade to the 256k ?

weedeewee wrote on 2023-12-07, 16:26:

butjer1010 wrote on 2023-12-07, 13:59:

weedeewee wrote on 2023-12-06, 20:04:

no problem, I just remembered I have a m321 board, though I think it's a rev 3, with 128k cache soldered on, and cpu in qfp soldered on, and 16? or 32M of ram.

Could it be the problem in Cache? Do i need U32 for 256KB?. I can remove chips, and leave only 4 chips, with IS61C64AH-20N in U17 (state before upgrade to 256), but can that be the fault with RAM?

My board also has no chip present in U32, so, I don't think that's the problem.
But... did you have to desolder old cache chips in order to upgrade to the 256k ?

No, they were in sockets already, so i just swapped them. Did You see the picture i have sent You (mine board, and cache chips)?

butjer1010 wrote on 2023-12-07, 21:40:

weedeewee wrote on 2023-12-07, 16:26:

butjer1010 wrote on 2023-12-07, 13:59:

Could it be the problem in Cache? Do i need U32 for 256KB?. I can remove chips, and leave only 4 chips, with IS61C64AH-20N in U17 (state before upgrade to 256), but can that be the fault with RAM?

My board also has no chip present in U32, so, I don't think that's the problem.
But... did you have to desolder old cache chips in order to upgrade to the 256k ?

No, they were in sockets already, so i just swapped them.

Okay, I was just thinking if you had to desolder them a trace might've gotten damaged, but since they were/are in sockets, that line of thinking is out the door.

Did You see the picture i have sent You (mine board, and cache chips)?

Yes.

Do you have an oscilloscope?

weedeewee wrote on 2023-12-08, 00:03:

Okay, I was just thinking if you had to desolder them a trace might've gotten damaged, but since they were/are in sockets, that […]
Show full quote

butjer1010 wrote on 2023-12-07, 21:40:

weedeewee wrote on 2023-12-07, 16:26:

My board also has no chip present in U32, so, I don't think that's the problem.
But... did you have to desolder old cache chips in order to upgrade to the 256k ?

No, they were in sockets already, so i just swapped them.

Okay, I was just thinking if you had to desolder them a trace might've gotten damaged, but since they were/are in sockets, that line of thinking is out the door.

Did You see the picture i have sent You (mine board, and cache chips)?

Yes.

Do you have an oscilloscope?

No, i don't have one 🙁
Without it, i suppose i cannot make some serious repairs?

Oh, depends.
it just gives an image of a signal so easy to see if it's present, bad or not.

weedeewee wrote on 2023-12-08, 16:39:

Oh, depends.
it just gives an image of a signal so easy to see if it's present, bad or not.

I have another 386 computer, but it is soldered sx33 CPU on MBO, i can put 4x4MB on this motherboard, and everything works perfect, but i wanted to make "more powerful" 386 with this M321 board, with DX40 and 387-40, but i cannot add more than 8MB RAM on this computer, and that's way i wanted to try to fix it. This PC works fantastic, DOS 6.22 and Win 3.1, but i wanted to "look" mean-er with 16MB 😀
I don't know what else can i do to find why it wont recognize more than 1MB per slot? So sad....

Considering there are no schematics of the board , that I know of, available. You're only way to go at it is to verify if the signal is present, which means tracing it's path all the way back past the buffer chips and verifying the traces aren't interrupted or shorted in anyway.
You could try putting the multimeter in voltage mode on A10, boot the computer and see if the measurement changes while the computer is booting and detecting the memory. I'm guessing it's not doing that at the moment, which would imply a break in the trace or a bad buffer along the way.

weedeewee wrote on 2023-12-09, 09:39:

Considering there are no schematics of the board , that I know of, available. You're only way to go at it is to verify if the signal is present, which means tracing it's path all the way back past the buffer chips and verifying the traces aren't interrupted or shorted in anyway.
You could try putting the multimeter in voltage mode on A10, boot the computer and see if the measurement changes while the computer is booting and detecting the memory. I'm guessing it's not doing that at the moment, which would imply a break in the trace or a bad buffer along the way.

That's a good idea! Thanks a lot, i will try that and let You know.

weedeewee wrote on 2023-12-09, 09:39:

Considering there are no schematics of the board , that I know of, available. You're only way to go at it is to verify if the signal is present, which means tracing it's path all the way back past the buffer chips and verifying the traces aren't interrupted or shorted in anyway.
You could try putting the multimeter in voltage mode on A10, boot the computer and see if the measurement changes while the computer is booting and detecting the memory. I'm guessing it's not doing that at the moment, which would imply a break in the trace or a bad buffer along the way.

It goes around 3V-3,5V while booting, and then when it reads memory, it stops at 0,2V. Is this ok?

Seems similar to mine with 32M of RAM
just tested it to compare. 3v5 then the memory test starts and it's about 2v2 (keep in mind these are all averages due to the signal changing between 0 and 5v. )

I'll unscrew some parts to verify what board revision I have

Interestingly, the bios string 30-0100-001437-00101111-060692-PC CHIP-F links to the v2.x version on theretroweb.

edit: and the description on the board says rev 3.0

weedeewee wrote on 2023-12-09, 14:28:

Seems similar to mine with 32M of RAM just tested it to compare. 3v5 then the memory test starts and it's about 2v2 (keep in min […]
Show full quote

Seems similar to mine with 32M of RAM
just tested it to compare. 3v5 then the memory test starts and it's about 2v2 (keep in mind these are all averages due to the signal changing between 0 and 5v. )

I'll unscrew some parts to verify what board revision I have

Interestingly, the bios string 30-0100-001437-00101111-060692-PC CHIP-F links to the v2.x version on theretroweb.

edit: and the description on the board says rev 3.0

Yes, same voltage as on my board. Mine is rev 2.5. So it is not the problem in traces? i've changed U18 (74F86N), just to see if something changes, but nothing did. I could try to change U21 also (74F244N), where the address 9 goes, but does that matter?

butjer1010 wrote on 2023-12-09, 18:05:

weedeewee wrote on 2023-12-09, 14:28:

Seems similar to mine with 32M of RAM just tested it to compare. 3v5 then the memory test starts and it's about 2v2 (keep in min […]
Show full quote

Seems similar to mine with 32M of RAM
just tested it to compare. 3v5 then the memory test starts and it's about 2v2 (keep in mind these are all averages due to the signal changing between 0 and 5v. )

I'll unscrew some parts to verify what board revision I have

Interestingly, the bios string 30-0100-001437-00101111-060692-PC CHIP-F links to the v2.x version on theretroweb.

edit: and the description on the board says rev 3.0

Yes, same voltage as on my board. Mine is rev 2.5. So it is not the problem in traces? i've changed U18 (74F86N), just to see if something changes, but nothing did. I could try to change U21 also (74F244N), where the address 9 goes, but does that matter?

As far as I know A0-A9 are used for 1MB, A10 for 4MB ...
Which chips are on your 9 chips simms ? and do you have four 3 chip 4MB simms you could try ?

weedeewee wrote on 2023-12-09, 18:13:

butjer1010 wrote on 2023-12-09, 18:05:

weedeewee wrote on 2023-12-09, 14:28:

Seems similar to mine with 32M of RAM just tested it to compare. 3v5 then the memory test starts and it's about 2v2 (keep in min […]
Show full quote

Seems similar to mine with 32M of RAM
just tested it to compare. 3v5 then the memory test starts and it's about 2v2 (keep in mind these are all averages due to the signal changing between 0 and 5v. )

I'll unscrew some parts to verify what board revision I have

Interestingly, the bios string 30-0100-001437-00101111-060692-PC CHIP-F links to the v2.x version on theretroweb.

edit: and the description on the board says rev 3.0

Yes, same voltage as on my board. Mine is rev 2.5. So it is not the problem in traces? i've changed U18 (74F86N), just to see if something changes, but nothing did. I could try to change U21 also (74F244N), where the address 9 goes, but does that matter?

As far as I know A0-A9 are used for 1MB, A10 for 4MB ...
Which chips are on your 9 chips simms ? and do you have four 3 chip 4MB simms you could try ?

They are all 9 chips, 2 are OKI 514100A70SJ, and 2 are OKI 511000B-70J. I don't have any 3 chips 4MB simms 🙁
I could try to find them online, but i doubt it'll be normal price 😀

butjer1010 wrote on 2023-12-09, 22:46:

weedeewee wrote on 2023-12-09, 18:13:

butjer1010 wrote on 2023-12-09, 18:05:

Yes, same voltage as on my board. Mine is rev 2.5. So it is not the problem in traces? i've changed U18 (74F86N), just to see if something changes, but nothing did. I could try to change U21 also (74F244N), where the address 9 goes, but does that matter?

As far as I know A0-A9 are used for 1MB, A10 for 4MB ...
Which chips are on your 9 chips simms ? and do you have four 3 chip 4MB simms you could try ?

They are all 9 chips, 2 are OKI 514100A70SJ, and 2 are OKI 511000B-70J. I don't have any 3 chips 4MB simms 🙁
I could try to find them online, but i doubt it'll be normal price 😀

Yours are 3 chips simms?