8mb parity simms aren’t particularly uncommon

What device are these for?

Have you seen this list : AST Part Number Description 501159-003
Premmia, Premmia SE, and Premmia MTE 16MB Memory Expansion Kit (Two 8MB SIMMs (P/N 104000-185, -198, -203) 70ns/72 pin JEDEC 32-bit - Parity)

https://www.minuszerodegrees.net/manuals/AST/ … AST_TB_0899.TXT lots of equal approved types for a picky AST
These SIMMS are 16 main chips and 8 parity chips from searches..... Does that sound correct? most common 8MB have 16 plus 2 parity.

It's an old game development system, very uncommon hardware.

I tried several other 8MB 2x36 FPM parity SIMM modules, even extremely similar ones like the KMM5362003BG-7 and they work erratically, in all slots, whereas the KMM5362000B2G-7 work flawlessly.

Picky machine...

These KMM5362000B2G-7 SIMMS are 16 main chips and 8 parity chips from searches..... Does that sound correct? most common 8MB have 16 plus 2 parity.

Horun wrote on 2024-01-10, 02:22:

Have you seen this list : AST Part Number Description 501159-003 Premmia, Premmia SE, and Premmia MTE 16MB Memory Expans […]
Show full quote

Have you seen this list : AST Part Number Description 501159-003
Premmia, Premmia SE, and Premmia MTE 16MB Memory Expansion Kit (Two 8MB SIMMs (P/N 104000-185, -198, -203) 70ns/72 pin JEDEC 32-bit - Parity)

https://www.minuszerodegrees.net/manuals/AST/ … AST_TB_0899.TXT lots of equal approved types for a picky AST
These SIMMS are 16 main chips and 8 parity chips from searches..... Does that sound correct? most common 8MB have 16 plus 2 parity.

This is super useful, I just purchased four Kingston KTM 8000/M70 on eBay, we'll see if they do the trick...

Thanks!

Datasheets are here .. http://www.datasheet.hk/search.php?part=2m%20 … 0modul&stype=pd ...'-7' will be 70ns ... you should be able to find equivalents by a comparo of specifications.

edit: this is me saying there may be other modules out there, made after that time by other vendors... ie; SmartM

Can you post us the exact chip numbers on the working ones ? closest I have are Toshiba tms362020as (8main + 4parity per side)

Better would be both side photos of the working ones in case they've got weird presence detect pin coding.

Absolutely! These are the exact modules: https://ram-co-shop.de/8-MB-FPM-RAM-mit-Parit … M5362000B2G-7_1

From what I can see the RAM chips are: KM44C1000BJ-7
And the parity chips are: KM41C1000CG-7

This is the module's datasheet: https://www.datasheets360.com/pdf/396037558785180777

Thanks eh! 😀

I can take actual photos of the modules, if the pictures in the link above aren't decent enough, let me know if there is anything in particular you want to see.

Cheers!

That photo shows a SIMM with a presence detect "resistor" aka zero ohm link connected to the top pair of pads on the backside, the full trace routing is not visible but it appears it may be connected to Pin 69
https://ardent-tool.com/memory/Identification.html
Which seems to indicate 8MB/70ns correctly.

Given that by the time "mass market" demand shifted to 8MB SIMMs rather than demanding commercial applications alone, the speeds were tending to be 60ns, then there's a large number of 60ns 8MB around, which should work fine electrically, but if they don't have that PD coding, systems may reject them, due to not very well thought out mechanisms to enforce "better than 80ns"

So if you've got a pile of rejected SIMMs to look at, you should see if any have same number of chips and are coded to that 8MB/70ns setting, and confirming those really don't work, we'll have to figure out what else might be causing incompatibility. If you know which end of a soldering iron to hold, then you might like to try recoding some low value parts (If 70ns or faster and FPM) to that setting, solder bridging or short bits of wire are fine, to see if you can get those working.

However, this is all supposing that the installed SIMM is in fact identical, and not just very similar and doesn't have different setting of the PD resistors/bridges.

I can hold a soldering iron, but before I start tinkering, I am trying to understand the piece of the puzzle that makes the KMM5362003BG-7 not cooperate.
From a pure datasheet perspective, they are extraordinarily similar, although they do not use the same about of chips... why the exact number of chips matters?

Jollyroger wrote on 2024-01-10, 02:24:

It's an old game development system, very uncommon hardware.

Cool. Can you post more details, name/model/pictures or link to description?

Jollyroger wrote on 2024-01-10, 19:52:

I can hold a soldering iron, but before I start tinkering, I am trying to understand the piece of the puzzle that makes the KMM5362003BG-7 not cooperate.

Thats what the tinkering is supposed to answer 😀 its possible hardware was validated on one speed only, and other modules you tried advertise something unsupported/untested.
http://www.walshcomptech.com/ohlandl/config/mempresence.html
PRESENCE DETECT TRUTH TABLE https://www.ele.uri.edu/iced/protosys/hardwar … arpoint-8MB.pdf

rasz_pl wrote on 2024-01-10, 22:48:

Jollyroger wrote on 2024-01-10, 02:24:

It's an old game development system, very uncommon hardware.

Cool. Can you post more details, name/model/pictures or link to description?

Sure, it is a Sophia Systems SEGA Saturn Programming Box. https://segaretro.org/Sega_Saturn_Programming_Box

rasz_pl wrote on 2024-01-10, 22:48:

Thats what the tinkering is supposed to answer :) its possible hardware was validated on one speed only, and other modules you t […]
Show full quote

Jollyroger wrote on 2024-01-10, 19:52:

I can hold a soldering iron, but before I start tinkering, I am trying to understand the piece of the puzzle that makes the KMM5362003BG-7 not cooperate.

Thats what the tinkering is supposed to answer 😀 its possible hardware was validated on one speed only, and other modules you tried advertise something unsupported/untested.
http://www.walshcomptech.com/ohlandl/config/mempresence.html
PRESENCE DETECT TRUTH TABLE https://www.ele.uri.edu/iced/protosys/hardwar … arpoint-8MB.pdf

Thing is, when I install the KMM5362003BG-7 modules, they work erratically, at times. I wrote a memcheck test and with the recommended modules, all tests pass perfectly fine, whereas with the replacement modules the tests pass and fail randomly, but mostly fail. I tried different modules, and I had pre-tested those modules in an old PC to ensure they work fine.
From the datasheets, on paper the only difference between the 2000s and the 2003s is the maximum operating current, which is lower in the newer modules, and I wonder if the machine pushes them beyond spec...

Have you tried true parity 4 chip per 4MB modules? That is 4mb x 1 bit for parity. The rest is 4MB x 4bits to fill out the 32 bits for 4MB. That's 12 chips for 4MB and 24chips for 8MB parity.

Remember, I'm sure this PC requires presence detection SMD jumpers set too, just like IBM.

The other modules that uses quad cas 4MB x 4bit works differently from using four 4MB x 1 bits for parity.

Cheers,

Sorry, I am not sure I understand these statements.
1. Which kind of module should I try? Other modules that use eight 1Mx4bit DRAMs, like the KMM5362000B2G-7?
2. There are a number of jumpers in the area around the four SIMM slots, but without removing the board (non trivial, it's a daughterboard connected to a number of other boards and the chassis), it's hard to tell what the jumpers are connected to. Maybe I will have to take everything apart just to see where those jumpers are connected to...

Jollyroger wrote on 2024-01-11, 16:26:

Sure, it is a Sophia Systems SEGA Saturn Programming Box. https://segaretro.org/Sega_Saturn_Programming_Box

https://mcretro.net/sophia-systems-sega-satur … -final-repairs/
from I could spy in Shane McRetro video its a custom FPGA memory controller? most likely something bespoke instead of fully universal.
Normal PC memory controller detects inserted ram, its size, organization, number of RAS lines used etc, row size (refresh window)? Yours is most likely fixed to whatever they validated design with.

Jollyroger wrote on 2024-01-10, 19:52:

Thing is, when I install the KMM5362003BG-7 modules, they work erratically, at times. I wrote a memcheck test and with the recommended modules, all tests pass perfectly fine, whereas with the replacement modules the tests pass and fail randomly, but mostly fail. I tried different modules, and I had pre-tested those modules in an old PC to ensure they work fine.
From the datasheets, on paper the only difference between the 2000s and the 2003s is the maximum operating current, which is lower in the newer modules, and I wonder if the machine pushes them beyond spec...

KMM5362000BG build using KM44C1000B http://rfelektronik.se/manuals/Datasheets/KM4 … miconductor.pdf
https://datasheetspdf.com/datasheet/KMM5362000B2G.html two banks? 4 RAS lines.

KMM5362003BG build using KM44C1000BJ, cant locate datasheet
https://datasheetspdf.com/pdf/542921/SamsungE … s/KMM5362003G/1 one bank? two RAS lines (RAS0 and RAS2).

Different internal organization.

rasz_pl wrote on 2024-01-12, 02:54:

KMM5362000BG build using KM44C1000B http://rfelektronik.se/manuals/Datasheets/KM4 … miconductor.pdf https://datasheetspdf.com/da […]
Show full quote

KMM5362000BG build using KM44C1000B http://rfelektronik.se/manuals/Datasheets/KM4 … miconductor.pdf
https://datasheetspdf.com/datasheet/KMM5362000B2G.html two banks? 4 RAS lines.

KMM5362003BG build using KM44C1000BJ, cant locate datasheet
https://datasheetspdf.com/pdf/542921/SamsungE … s/KMM5362003G/1 one bank? two RAS lines (RAS0 and RAS2).

Different internal organization.

About the chip labelling: KM44C1000B is the generic name, KM44C1000BJ is the specific name for that very chip in a SOJ 26/20 package. There should be no difference in data memory chips. The main difference is that the 2003 module has a single merged parity chip KM44C1003B (no datasheet found) instead of four dedicated single-bit parity chips

Both modules are two banks: The second module uses /RAS2 and RAS3, not RAS0 and RAS2. PS/2 SIMMs have two /RAS lines per bank: /RAS0 and /RAS2 for the primary bank and /RAS1 and /RAS3 for the secondary bank. The reason for two /RAS line is mainly to split the load of driving up to 36 x1 chips. The older modules has 12 chips per bank (8 data chips (4 MBit: 1M x 4), 4 parity chips (1 MBit: 1M x 1), as already mentioned in the thread), whereas the newer module just has 9 chips per bank: 8 data chips in the same way, but only one 4MBit parity chip in a special "one /CAS per bit" configuration. This special parity configuration can cut costs and behaves the same as the 4 individual parity chips in usual circumstances.

A standard PS/2 SIMM memory controller always drives /RAS0 and /RAS2 to the same level, the same is true for /RAS1 and /RAS3. So as long as the drivers are not overloaded, the memory controller doesn't notice which RAM chips are connected to what /RAS line. Having only 9 chips per module allows a single /RAS line to be used without being afraid of overloading a standard PS/2 SIMM mainboard. The incompatibility of the newer SIMM may be rooted in different causes: Possibly this system has weaker drivers and can tolerate 6 chips per /RAS line, but not 9 chips per /RAS line. Another (remote?) possiblity is that this board (didn't look into the technology) tries to use the module as four banks of 16 bits by considering /RAS0 and /RAS2 as /RAS lines for dedicated halves of the first bank. This will work only if the routing of the /RAS lines to the memory chips is split exactly in the way the main board expects - and it will not work at all with the "merged parity chip" design of the later module, because splitting the bank into two x18 halves requires two parity chips connected to /RAS0 and the other two parity chips connected to /RAS2.

rasz_pl wrote on 2024-01-12, 02:54:

https://mcretro.net/sophia-systems-sega-satur … -final-repairs/ from I could spy in Shane McRetro video its a custom FPGA memory […]
Show full quote

Jollyroger wrote on 2024-01-11, 16:26:

Sure, it is a Sophia Systems SEGA Saturn Programming Box. https://segaretro.org/Sega_Saturn_Programming_Box

https://mcretro.net/sophia-systems-sega-satur … -final-repairs/
from I could spy in Shane McRetro video its a custom FPGA memory controller? most likely something bespoke instead of fully universal.
Normal PC memory controller detects inserted ram, its size, organization, number of RAS lines used etc, row size (refresh window)? Yours is most likely fixed to whatever they validated design with.

Jollyroger wrote on 2024-01-10, 19:52:

Thing is, when I install the KMM5362003BG-7 modules, they work erratically, at times. I wrote a memcheck test and with the recommended modules, all tests pass perfectly fine, whereas with the replacement modules the tests pass and fail randomly, but mostly fail. I tried different modules, and I had pre-tested those modules in an old PC to ensure they work fine.
From the datasheets, on paper the only difference between the 2000s and the 2003s is the maximum operating current, which is lower in the newer modules, and I wonder if the machine pushes them beyond spec...

KMM5362000BG build using KM44C1000B http://rfelektronik.se/manuals/Datasheets/KM4 … miconductor.pdf
https://datasheetspdf.com/datasheet/KMM5362000B2G.html two banks? 4 RAS lines.

KMM5362003BG build using KM44C1000BJ, cant locate datasheet
https://datasheetspdf.com/pdf/542921/SamsungE … s/KMM5362003G/1 one bank? two RAS lines (RAS0 and RAS2).

Different internal organization.

The daughterboard that includes the custom memory controller does indeed use an FPGA to implement it, and the both the circuitry and the bitstream have not been studied in any detail (yet).
It is entirely possible that the controller has been designed and validated only with a very specific module, which is literally the only one the user manual mentions to be used.