rasz_pl wrote on 2023-02-23, 01:09:

Fantastic! Would it be possible to also generate a boardview file? If you used Kicad this supposedly should work https://github. […]
Show full quote

Nelson68k wrote on 2022-11-02, 16:46:

I put the Gerberfiles into the Git reposotory.

Fantastic! Would it be possible to also generate a boardview file? If you used Kicad this supposedly should work https://github.com/whitequark/kicad-boardview
Boardview files can be opened with for example http://boardviewer.net https://github.com/OpenBoardView/OpenBoardView and help with troubleshooting, not to mention they are a great educational tool. Did you happen to also make a diagram?
Afaik this is the second new/old 386 motherboard that can be build today, with https://alexandrugroza.ro/microelectronics/sy … sbmc/index.html being the other one.

I'll have 1o of them, too with parts if anyone is interested. ENIG finish, too. Off topic, too. This is for M396F.

-William

nuno14272 wrote on 2023-02-22, 13:44:

the only disadvantage i see is not a cache model.

Cache is good. And a 486DX is faster. But I just wanted to create a 1:1 replacement for my defective board. Maybe someone else can use it as a replacement.

Greetings,

Anyone interested in building their own M396F from https://github.com/Marco-Both/M396F-Replika I have 10 PCB's. One for myself and then 9 others, with parts. I want to build one fully and test before I send out in the wild, but if you all can't wait, I can send out now with parts. I'm waiting on SIMM sockets at the moment. Give me a couple weeks. PCB's I'm selling at cost to break even but no profit. With parts, it will increase the kit quite a lot. Depending on if you want to find them all yourself or rely on what I have on hand. I may work on packaging up some kits. I'd like to stick in the US. Just thought it'd be awesome to at least have the PCB's available. Letting the community know I have some qty available.

Attaching the available PCB's as I test a fit of some dual rank SIMM sockets.

- William

rasz_pl wrote on 2023-02-23, 01:09:

If you used Kicad this supposedly should work https://github.com/whitequark/kicad-boardview

Hi, I do not use KiCad so far.

Interesting project. I may be interested in one of the PCBs, plus some components. I already have some appropriate stuff - CPUs (386SX, 486SLC etc), 30-pin SIMM sockets (also dual rank), EPROMs, chip sockets etc.
My SIMM sockets are in fact better than the ones pictured above, they have metal clips. 😀

stamasd wrote on 2023-03-23, 22:15:

Interesting project. I may be interested in one of the PCBs, plus some components. I already have some appropriate stuff - CPUs (386SX, 486SLC etc), 30-pin SIMM sockets (also dual rank), EPROMs, chip sockets etc.
My SIMM sockets are in fact better than the ones pictured above, they have metal clips. 😀

Hello ,

Metal clips, just couldn't tell from the angle.

-William

If you all want a M396F some let me know. The PCB's, IC's parts are all here, if you want one, let me know.

matching RAM 4MB each

a black beauty with 486SLC2 CPU.

overall-performance of SLC2-50 is lower than SLC-40 (only CPU power is higher )

Madao wrote on 2023-06-04, 15:55:

a black beauty with 486SLC2 CPU.
photo_2023-06-04_12-46-56.jpg
overall-performance of SLC2-50 is lower than SLC-40 (only CPU power is higher )

Interesting. How did you measure?

here.

Now, i have tweaked board with 486SLC2-50, a bit better performance.
It is not great jump from SLC33 to SLC2-50. I believe: Jump to SLC-40 is better.

Do any of you have the datasheet for the SARC chipset? Or any closely related one (such as the SARC RC2016A8 / Toshiba TC6154AF / PCCHIPS Chip 2 / CDCOM, all of which appear to be the same)? I've been looking for it, but can't find one.

It seems to be some sort of programmable ASIC. Any documentation that may have existed is probably lost to time.

I'm sorry. I don't have any datasheet. The chipset can be purchased from UTSource.

OK, thank you. For future people who might visit this thread, I have a sneaking suspicion that all of the chipsets mentioned by the OP and me above might just be Headland HT21 chips, for which datasheets are readily available. Some of the reason for believing this is the same revision of the motherboard in this thread is available where the chipset has a "PC CHIPS CHIP 2 / 6L70F2060" sticker on it and that can be tied back to HT21, at least indirectly. (See here for a photo of the exact same revision motherboard with the CHIP2 branding https://i.ebayimg.com/images/g/eDAAAOSwyeBkwCc-/s-l1600.jpg).

PC CHIPS CHIP 2 has previously (I think) been identified on some motherboards as an HT21. PC CHIPS CHIP 3 (note the 3) / 6L70F2060 certainly has also shown up on 286 boards identified as HT21: https://theretroweb.com/motherboards/s/pcchips-m216. Elsewhere PC CHIPS CHIP 2 has been shown to sometimes have Toshiba TC6154AF identifiers when labels are peeled...and that chip has seen substitutions in PC CHIPS motherboards with branding as SARC RC2016A8 and CDCOM. I have an M219 with CDCOM branding on it coming from that Ukraine motherboard seller. I plan to peel up the label when it arrives (I expect to see TC615AF identifiers) and will write a test program, trying to program the chipset registers to see if it behaves like a HT21.

...if all of this pans out we might have half a dozen different names and branding for the same chip. Ug!

Deksor Identifying PCChips 386 era chipsets has a good theory on SARC/CDCOM
UMC is the first Taiwanese semi foundry, they had whole technology park full of people working on designs/clones, they stopped with clones in early nineties after a string of lawsuits 😀
Taiwan's Other Semiconductor Giant - The Asianometry https://www.youtube.com/watch?v=q55lQUHg7Dg

I have also seen a few of these different names. I received two damaged PCBs from the same seller for the reproduction. Their IC was also labeled differently (SARC / PC Chip).

By simply measuring the GND and VCC pins, an electrical test can be made. And then simply solder on an HT21 and see if the board boots. In the meantime I have also received a few different BIOS files.

HT21/D datasheet https://theretroweb.com/chipsets/344#downloads

We've made some more research on that topic on theretroweb and we now think but can't really prove that this chipset is a Toshiba gate array programmed with an ALD design.
ALD being also known to relabel UMC and headland chipsets, which could explain the sarc 2015, but so far it's tough to find anything that ressembles the sarc 2016. It may actually be an original design for once.

But we can't be sure ...