Try to obtain an Adaptec 2740W or 2742W

Disruptor wrote on 2021-10-19, 14:40:

Try to obtain an Adaptec 2740W or 2742W

Cool, that was my thought process as well, from what I have read Adaptec boards seem to work (some need a tweak with the CFG file), I have a 2742 hanging around I was going to try, before I try putting something more exotic in the system, i.e. DPT

The Adaptec board may not work properly with the HiNT Caesar chipset. One of the things that was cut out was properly bus mastering support which the 2740W uses. The older AHA-1740 would work with CFG file changes, check old usenet posts.

NJRoadfan wrote on 2021-10-19, 22:37:

The Adaptec board may not work properly with the HiNT Caesar chipset. One of the things that was cut out was properly bus mastering support which the 2740W uses. The older AHA-1740 would work with CFG file changes, check old usenet posts.

Already all over that, that came up in my research and is referenced in my first post, just got to wait for the board to arrive, then I can have a real play to see what this chipset can and cannot do, all I can do at the moment is further research and order some odd bits. Was also going to make some sort of manual for this board and chipset (a bit further than jumper settings) as a lot of the information is dated or lost to time, also a lot of copy and pastes of those original posts for this chipset.

Feel like a child waiting for Christmas with a new unusual board on the way!

Well, I have started to write my own user manual for this board, still very much a rough draft/ work in progress and some horrendous grammar, still needs a lot of work, but like where it is going

https://docs.google.com/document/d/1WsVpnHlKR … of=true&sd=true

EDIT: Link stopped working after a major chop around of the document.

Needing to retire? That poor machine's lasted 30 years, if it was going to retire it would have been in the late 90s! 😉

simon_e_hall wrote on 2021-10-19, 14:28:

Well, I thought it was time to start a new build, as my current 486DX EISA build is beginning to show signs of needing to retire.

Still waiting for the motherboard, joys of shipping accross the pond 🙁
However, my parts are slowly being found and piling up in the corner of the house:

From my collection of parts:
1) Microsoft home mouse (with wheel) and keyboard
2) Cirrus Logic 5429 VLB card 1MB
3) Got mixed information on this chipset, regarding write back, it probably does not, but in my collection is a 486dx-66 with write back, so another experiment for this chipset, but I could also reuse the Am5x86 in my current EISA build.
4) 5.25" and 3.5" floppy drives (beige)
5) 50 pin SCSI CD drive (beige)
6) Got plenty of SCA SCSI drives and 50 pin adapters to choose from for the main drive, but just having the one instead of the multiple ones I usually do

Reuse from my old EISA build,
1) 3Com network card (if it will work with the chipset)
2) Weitek 4167 (used when I play with CAD software which supports it)
3) SB16 (bug free DSP version, but will now have a dreamblaster S2 from my collection attached to it)
4) Generic IO Card (or another fancier one from the collection)
5) DPT SCSI Card (not sure about this yet, again depends if it works with the chipset)

Purchased:
1) Beige case with 200w power supply
2) AHA-1740 (with floppy connector, as I know with a tweak in the config file this will work with the chipset and can at least setup a basic system to begin with)

And when I get the basic system up and running, apart from the the usual benchmarks:
1) Upgrade VLB card to its maximum 2MB

2) Apparently the motherboard can do 128MB, but again heard mixed reviews about the EISA bus having issues with large amounts of RAM, but it would be fun to have a 486 with 128MB, so if I can find it at a reasonable cost in the future 16MB x 9 60ns FPM SIMMs, but I might have some of those in my RAM collection, also intreseted to see how much is fitted to the board as it is coming with RAM.

3) Might try my hand at retro brit as the drives and case are at various stages of yellowing, what are peoples thoughts on that? I have never tried it before as I have always gone for black cases.

4) Beige monitor either CRT or modern flat type, as the one I am using is being donated to my eldest son, as I am building a all chinese parts PC for Christmas, it is complete overkill on a retro PC.

5) Go through the old EISA motherboard with a fine toothcomb to see if I can identify what its issue is.

Having much RAM on a 486 is nice, but you first have to ensure that it is covered by the cacheable area too.

... And if the ram is not covered by cache, you can put a ramdisk on the non-covered RAM area. 😉

If the chipset was made before the DX4 came out (early 1994), then you can probably forget about writeback L1...unless it's a Cyrix 486S. Those came out in 1993, and were the first 486s to support writeback L1 cache, but supposedly they used an implementation that is not compatible with the intel chips. Hell, I'm not even sure if it's compatible with Cyrix's later offerings either. Your board appears to made in mid-1992, which is really early for even a VLB board.
A jumper for the P24T could possibly indicate some kind of support for writeback L1, but again, POD83s aren't known to work all that well on boards made before 1995 and it's usually L1 cache related.

Anonymous Coward wrote on 2021-10-29, 14:03:

If the chipset was made before the DX4 came out (early 1994), then you can probably forget about writeback L1...unless it's a Cyrix 486S. Those came out in 1993, and were the first 486s to support writeback L1 cache, but supposedly they used an implementation that is not compatible with the intel chips. Hell, I'm not even sure if it's compatible with Cyrix's later offerings either. Your board appears to made in mid-1992, which is really early for even a VLB board.
A jumper for the P24T could possibly indicate some kind of support for writeback L1, but again, POD83s aren't known to work all that well on boards made before 1995 and it's usually L1 cache related.

That is what I thought, but while digging around for information on this board, came accross this on the forum:

Help ID-ing 386 EISA/VESA board - jumper settings?

which shows the same chipset, with what appears to be the same dated BIOS chip and further down a screen shot of the the 'Advanced Chipset Setup' with an option for 'Cache Scheme' with 'WR/Back' as an option. But that could be to do with the OPTI 495 on the board, cannot find much information on the SIS 85C206 on my board.

Pretty sure that write back cache option is for L2, not L1.

Those 206 chips just integrated a bunch of stuff that would normally be discrete logic, plus the 82284 clock generator, 82288 bus controller, 8254 system timer, dual 8259 interrupt controllers, dual 8237 DMA controllers, MC146818 clock chip etc. Pretty much all 486 boards have one. Many companies made their own, and they are relatively interchangeable, which is why on your board which has HiNT chipset it is mixed with an SiS part.

simon_e_hall wrote on 2021-10-29, 14:37:

That is what I thought, but while digging around for information on this board, came accross this on the forum: […]
Show full quote

Anonymous Coward wrote on 2021-10-29, 14:03:

If the chipset was made before the DX4 came out (early 1994), then you can probably forget about writeback L1...unless it's a Cyrix 486S. Those came out in 1993, and were the first 486s to support writeback L1 cache, but supposedly they used an implementation that is not compatible with the intel chips. Hell, I'm not even sure if it's compatible with Cyrix's later offerings either. Your board appears to made in mid-1992, which is really early for even a VLB board.
A jumper for the P24T could possibly indicate some kind of support for writeback L1, but again, POD83s aren't known to work all that well on boards made before 1995 and it's usually L1 cache related.

That is what I thought, but while digging around for information on this board, came accross this on the forum:

Help ID-ing 386 EISA/VESA board - jumper settings?

which shows the same chipset, with what appears to be the same dated BIOS chip and further down a screen shot of the the 'Advanced Chipset Setup' with an option for 'Cache Scheme' with 'WR/Back' as an option. But that could be to do with the OPTI 495 on the board, cannot find much information on the SIS 85C206 on my board.

Off topic, what up with this guy with "H" on the forehead?

Cheers,

pentiumspeed wrote on 2021-10-31, 20:31:

simon_e_hall wrote on 2021-10-29, 14:37:

That is what I thought, but while digging around for information on this board, came accross this on the forum: […]
Show full quote

Anonymous Coward wrote on 2021-10-29, 14:03:

If the chipset was made before the DX4 came out (early 1994), then you can probably forget about writeback L1...unless it's a Cyrix 486S. Those came out in 1993, and were the first 486s to support writeback L1 cache, but supposedly they used an implementation that is not compatible with the intel chips. Hell, I'm not even sure if it's compatible with Cyrix's later offerings either. Your board appears to made in mid-1992, which is really early for even a VLB board.
A jumper for the P24T could possibly indicate some kind of support for writeback L1, but again, POD83s aren't known to work all that well on boards made before 1995 and it's usually L1 cache related.

That is what I thought, but while digging around for information on this board, came accross this on the forum:

Help ID-ing 386 EISA/VESA board - jumper settings?

which shows the same chipset, with what appears to be the same dated BIOS chip and further down a screen shot of the the 'Advanced Chipset Setup' with an option for 'Cache Scheme' with 'WR/Back' as an option. But that could be to do with the OPTI 495 on the board, cannot find much information on the SIS 85C206 on my board.

Off topic, what up with this guy with "H" on the forehead?

Cheers,

It's from another one of my sci-fi television faves:
https://en.wikipedia.org/wiki/Red_Dwarf

Come on, man. You're Canadian and you've never seen Red Dwarf?

pentiumspeed wrote on 2021-10-31, 20:31:

simon_e_hall wrote on 2021-10-29, 14:37:

That is what I thought, but while digging around for information on this board, came accross this on the forum: […]
Show full quote

Anonymous Coward wrote on 2021-10-29, 14:03:

If the chipset was made before the DX4 came out (early 1994), then you can probably forget about writeback L1...unless it's a Cyrix 486S. Those came out in 1993, and were the first 486s to support writeback L1 cache, but supposedly they used an implementation that is not compatible with the intel chips. Hell, I'm not even sure if it's compatible with Cyrix's later offerings either. Your board appears to made in mid-1992, which is really early for even a VLB board.
A jumper for the P24T could possibly indicate some kind of support for writeback L1, but again, POD83s aren't known to work all that well on boards made before 1995 and it's usually L1 cache related.

That is what I thought, but while digging around for information on this board, came accross this on the forum:

Help ID-ing 386 EISA/VESA board - jumper settings?

which shows the same chipset, with what appears to be the same dated BIOS chip and further down a screen shot of the the 'Advanced Chipset Setup' with an option for 'Cache Scheme' with 'WR/Back' as an option. But that could be to do with the OPTI 495 on the board, cannot find much information on the SIS 85C206 on my board.

Off topic, what up with this guy with "H" on the forehead?

Cheers,

Yep, Rimmer off Red Dwarf, childhood nickname that has stuck forever....

Well tracking has informed me motherboard is in the UK! Now it has to weave its way West, old case and PSU completly torn down, cleaned and inspected over the weekend, no point trying to retro brite it in this weather, PSU looks good, keylock replaced with one with a key as well. Then hopefully figure out the good and bad bits about this chipset and get some answers about its capabilities.

Anonymous Coward wrote on 2021-10-29, 14:03:

If the chipset was made before the DX4 came out (early 1994), then you can probably forget about writeback L1...unless it's a Cyrix 486S. Those came out in 1993, and were the first 486s to support writeback L1 cache, but supposedly they used an implementation that is not compatible with the intel chips. Hell, I'm not even sure if it's compatible with Cyrix's later offerings either.
...

Indeed.
The old Cyrix DX and DX2 support writeback in their own layout but the new ones like the DX4 implement write back in an Intel compatible layout . They have printed "STANDARD LAYOUT" on the ceramics.

Can the old Cyrix DX2 use the 486S implementation?

Motherboard arrived last night, battery was just beginning to show signs of leaking, so one motherboard saved from Varta, the destroyer of worlds.

Board posts nicely. However, the BIOS ID String does little to resolve where this motherboard comes from, looks like another mangled ID, was at least expecting to see CAESAR at the end of the string, very odd.

40-0000-816276-00101111-031692-F

Nothing more that I am a 486, no major/ minor BIOS revisions, cannot find anything on that manufacture ID, last few digits at least seem right for the year 16th March 1992, keyboard controller revision F.

A little concerned about the corrupted character at the start of the VGA ID String on every boot, hope that is just a slight configuration tweak needed with timings.

The EISA Configuration Utility says it needs !HIT001.CFG so that is cool, as it is easily available on the Mr. Slug archive.

Identified most of the jumper settings now, just two I cannot identify, more testing to follow but on the late shift tonight and want to give the board a bit of a clean.