pshipkov wrote on 2024-09-26, 23:45:

What is the local storage ?
Did you consider using an extension IDE card like Promise Ultra-100/133, or other maybe SCSI one ?
I am not 100% about the EDO RAM issues with this board. How confident you are about not being a flaky memory module(s) but the board itself ?

I use a Western Digital 68AA with about 6,4 GB formatted capacity (IDE). From my experience, the board have problems using HDD of 8 GB (freezing at the 1-st POST-page), so the HDD must be smaller than 8 GB.
Because of no recognition or freezing if more than 2 PCI-cards at FSB 50 MHz are used, I would not use any external IDE-controller, as I need the place for other more useful cards (e.g. the Voodoo 2) .

All of my HOT-433 Rev.1 I have with northbridge-steppings BTA, BTS and BTC do have problems with EDO-RAM as described above.
I friend of mine has the same board-revision with a BCS-stepping and no problems with EDO-RAM (tested with my modules not working good on my boards), so I am pretty sure it's caused by the used northbridge only.
Here (3 (+3 more) retro battle stations) feipoa means, his BCA-stepping does work fine with EDO-RAM, too.
So, maybe, a stepping of BC... is needed for EDO on revision 1?

I have another question about the Intel 486DX4-100: does this CPU accept a multiplier of 2,5x in generally?
I tried to use the DX4-100 running until now @ 3x 50 MHZ at a higher FSB of 60 MHZ at 2,5x (using an additional CPU-interposer between the voltage-interposer and the mainboard).
Sadly, this CPU-interposer seems to have no effect regarding the multiplier (while its voltage is not used as there is the voltage-interposer) - the CPU runs always with the multiplier adjusted on the mainboard.

Am I wrong, or is the CPU-interposer maybe just broken?

Only thing I know with 2.5x multi available is the Pentium overdrive, there might have been some unicorn of unicorn 5x86 that did it.

BitWrangler wrote on 2024-09-29, 11:44:

Only thing I know with 2.5x multi available is the Pentium overdrive, there might have been some unicorn of unicorn 5x86 that did it.

I think, the POD does not need any CPU-interposer to work on a 486-board (even less an interposer containing a voltage-generator, I assume this one can do 3,3 V or 3,45 V only, as there is no jumper to adjust the voltage).

The multiplier of 2,5x is integrated in the POD by hardware (and it drops down to 1,0x if the original cooler is taken down or does not work - this was the very first "CPU-throttling-technology" worldwide on the market, if I remember correctly).

So in my eyes this multiplier is not meant for a POD.

I have some very vague ideas about a discussion about 2.5 on certain Cyrix DX4 100, with the reasoning that their DX4 wasn't great at hitting 120, and they weren't all that competitive with AMD and Intel DX4 and they wanted to offer the option for people upgrading from DX40 and DX2/80 machines who couldn't figure out how to change the bus speed. IDK if that was real info, or some BS discussion/speculation on a BBS, newsgroup or forum way back, between dudes who can barely get a SIMM module in a board, or what.

Things have gotten even more interesting.
With two other 486-CPU-interposers testet, again, no multiplier of 2,5x for an Intel-486 DX4 is possible.
Testet with one i486DX4-100WT and one i486DX4-100WB, just to be sure.

The question is: WHY?

What I know:
- no one AMD-486-CPU is able to accept a multi of 2,5x (according to the AMD-486-datasheet), so no sense to test it
- in the Intel-486-datasheet, there is no info about 2,5x, too (although here Is there a way to change the internal multiplier of 486DX/n processors? there is a THEORETICAL info about 2,5x - but I was not able to find it in the intel-papers)
- all other i486-CPUs different (or: older than) from the DX4 are hardwired at 1,0x or 2,0x, so they are not able to adjust their multis per jumpers on the board, so they do not benefit from such interposer at all (even the voltage of 3,3/3,45V is not needed for them, as they need 5 V, at least as desktop-CPUs), see also the screenshot from the i486-paper attached
- the POD (Pentium Overdrive) does not need this adaper, too, as it generates its own voltage and is internal hardwired to 2,5x (1,0x without original cooler)
- finally, I assume, this 2 interposers I have are not intended for Cyrix-CPUs

Does somebody have other ideas about this 2,5x-multiplier, and how to reach it?

Attached here is the comparison of Intel-486-CPUs from the Intel-486-datasheet regarding the ability to change the internal multipler - it can be seen, this can do only the DX4-CPU (this is the only CPU with Pin R17 " CLKMUL" internal activated)

Intel 486dx4 has only 2x and 3x multipliers controlled by r17 pin and thats it. Internally these are the only two options possible. You can not influence that from outside. The pin state can be low or high. 2x or 3x.

pshipkov wrote on 2024-09-30, 15:49:

Intel 486dx4 has only 2x and 3x multipliers controlled by r17 pin and thats it. Internally these are the only two options possible. You can not influence that from outside. The pin state can be low or high. 2x or 3x.

You are absolutely right about this, as the Intel-datasheeet shows the same.

But I still wonder why there is an option of 2,5x showed especially for the i486 on the bottom of the first interposer from yesterday - this is a really very "clear message" there...

BTW, adjusting 2,5x on all 3 interposers results in a real multi of 3,0x for the i486DX4.

It is impossible to say what the circumstances on the ground and within the ecosystem were at the time the adapter was made, so maybe designers had information about chip designs that never actually materialized, or simply exposed the multiplier which is a byproduct of the adapter's design.

This has been discussed a couple of times before. The 2.5x multiplier was indeed planned at some point, and made it into the datasheet, but it was not implemented in the final version. It would be interesting to see if anyone has DX4 prototypes that still have it.
Supposedly, the reason for the 2.5X multiplier existed was that at some point there was supposed to be a 486DX3, which I guess would have run at 63MHz (25x2.5) and 83MHz (33x2.5) like the POD83. It's not entirely clear why such a product would have been necessary since the DX/2-66 and DX/4-75 seemly had those bases covered. I guess for a short time those products were considered viable, but then the market conditions changed.
Probably whoever designed the CPU interposer was working with an earlier version of the DX/4 datasheet. It's not clear if the actual products hadn't been released yet or if the designer was just a doofus.
Of other note, I did find one use for the 2.5X jumper setting on my interposer. Some older 5V 486 motherboards won't boot with a Cyrix 5x86 in 2x or 3x mode, but will boot in 2x mode if the "2.5x" jumper is set.

Anonymous Coward wrote on 2024-10-06, 17:01:

Supposedly, the reason for the 2.5X multiplier existed was that at some point there was supposed to be a 486DX3, which I guess would have run at 63MHz (25x2.5) and 83MHz (33x2.5) like the POD83. It's not entirely clear why such a product would have been necessary since the DX/2-66 and DX/4-75 seemly had those bases covered. I guess for a short time those products were considered viable, but then the market conditions changed.

Well, a DX3-63 would've been a nice option on a badly designed board, only capable of running 25 or as an Overdrive for boards with fixed oscillator? Where as a DX3-83 would've been ridiculously faster than a DX4-75 for someone who would have that option. But probably yeah, too close performance points to make real sense ...

H3nrik V! wrote on 2024-10-07, 03:53:

Well, a DX3-63 would've been a nice option on a badly designed board, only capable of running 25

are there really boards like that in the wild?

H3nrik V! wrote on 2024-10-07, 03:53:

or as an Overdrive for boards with fixed oscillator?

oscillator and a socket is much cheaper

H3nrik V! wrote on 2024-10-07, 03:53:

Where as a DX3-83 would've been ridiculously faster than a DX4-75

afaik DX4-75 went only into laptops where 25MHz fsb was leveraged to extend battery life.

Expected something like that - spec for products that were never made.
Obviously haven’t seen any processors like that.

I acquired one of these boards earlier this year, and I am just now getting it together but having problems. I feel kinda dumb because I built my first computer back in 88'. My favorite system was a 386 and I got this board from someone locally on offerup.

At first I was getting nothing, my PSU wasn't even turning on, after I pulled the 387 it would spin the fan on the PSU but no beeps or anything on the screen. Then I ran into your post. I have a AMD 386DXL-40 in there, and I'm hoping just a different crystal will do it. No idea which crystal to get. Could you offer some advice?

pshipkov wrote on 2023-03-12, 05:27:

@BitWrangler It is surprising there is so much trouble with RAM. Usually these boards take 4x4Mb for maximum of 16Mb, but troubl […]
Show full quote

@BitWrangler
It is surprising there is so much trouble with RAM.
Usually these boards take 4x4Mb for maximum of 16Mb, but troubles with 4Mb sticks is certainly out of the ordinary.
At least you overcame the resistance.
Win98SE on this poor silicon and working ok ?
By ok you mean slideshow upon every user action, or actually ok ok ?

Last question - what will be your next steps with this guy ?

---

Several 386 motherboards based on Unichip U4800 passed through here, but none of them was worth talking about.
They were those compact PCBs with 128Kb level 2 cache only, clock generators limited to 80MHz (40MHz FSB) and very slow BIOS microcodes.
Finally managed to obtain a motherboard with the right characteristics.

Target Micro Typhoon 486-386/ISA based on Unichip U4800 chipset - a hybrid 386/486 assembly.

Had no idea what to expect from it, so curiosity was significant.
Unfortunately it came dead. Or so i thought.
The drama started with me inserting an 80MHz crystal and trusted CPU, RAM, VGA.
Hard no lights.
But i was desperate to get this guy up and running. Such boards with this chipset are extremely rare. Unlikely i will be able to find another.
Probed all over the place with multimeter and oscilloscope - everything came out clean and as expected.
Went over jumper settings countless times. Tried any other configurations, including random ones.
Desperation gradually ramped up. More and more invasive methods for reanimation were applied.
Endless swapping of components, reflowing joints, pins, pads. Bending.
Twice baked in the oven.
Heating with hot air gun sections of the PCB - blindly trying to fix eventual cold joints
At some point legs of the main IC started separating from the pads underneath. This resulted in even more soldering work around them.
4 level 2 cache sockets completely melted.
And so on.
Eventually i gave up and threw the motherboard in the pile of broken components.
...
2 weeks later i impulsively grabbed it, put it back on the test bench, inserted a 66MHz oscillator - it lit up right away and booted to DOS.
My jaw was on the floor.
Inserted an 80MHz oscillator - hard no lights.
No lights with 70, 80, 90, 100.
But worked right away with 110MHz one.
What the heck.
Over time, don't ask me how i arrived at it, i realized that with 70-100MHz crystals i must switch JP3 from 1-2 to 2-3 or back during POST.
Despite this nasty quirk i felt very guilty for all the abuse this motherboard went through for no reason.
The really surprising thing is that the PCB survived my Emergency Room, but hey, i take it.

With all the drama behind us (me and the board) i spent good time to heal it properly and finally moved onto testing it.

16Mb of RAM
256Kb level 2 cache in WriteBack mode
Cirrus Logic GD-5434, 2Mb
Standard IDE controller with CF card.

--- 386DX running at 55MHz, ISA at 18.33MHz

Any FPUs makes the system very unstable when FSB is in the range of 45-50MHz.
At the same time at 55MHz or higher FPUs don't get recognized.
Basically things work well at up to 33MHz FSB, but get weird in different ways past that - the POST issue outlined above which needs jumper swaps and now this FPU stuff.
Testing a 33MHz system is just not interesting, so decided to go with the 55MHz no FPU configuration.

All BIOS settings on max, except:
AT WAIT STATE = 1 WS (best is 0 WS)
AT BUS CLOCK = CLK2/8

Performance is actually pretty ok.
Above average performance in the DOS interactive graphics tests, but very slow in Windows accelerated GUI.

No Quake 1, PC Player Benchmark and offline graphics tests (no FPU).

--- TI 486SXL2

This didn't go well.
Motherboard does not light-up above 35MHz FSB (70MHz crystal).
Tried hard to figure this out, but without success.
Disappointing.
Decided not to waste time testing at such a low frequency (35MHz).

--- IBM BL3 running at 100MHz (2x50), ISA at 16.66MHz

Motherboard loved this processor, but the situation with FPUs still stands.
All BIOS settings on max, except:
AT WAIT STATE = 1 WS (best is 0 WS)

Average performance.

No Quake 1, PC Player Benchmark and offline graphics tests (no FPU).

--- IBM BL3 running at 110MHz (2x55), ISA at 18.33MHz

Not 100% stable, but feels very close to the threshold. Rare, sporadic hangs may occur from time to time.
Same FPU problems.

All BIOS settings on max, except:
AT WAIT STATE = 1 WS (best is 0 WS)
CACHE SCHEME = WRITE-TRHOUGH (required for stability in Windows)

Speedsys hangs.

Performance is pretty good.
Number 4 in Superscape and Doom.
Not great in Windows accelerated GUI.

No Quake 1, PC Player Benchmark and offline graphics tests (no FPU).

---

Looks like this motherboard was designed to handle 386 CPUs at up to 33MHz. The stability at 55Mhz feels like an accident than by design.
Obviously not a great option to build a retro computer around it, but at least it was enough to answer some performance questions about the Unichip U4800 chipset.

benchmark results

Hey joebroke,

Start with 66mhz oscillator. Match the jumpers config how it is shown on the photo.
Use trusted CPU, RAM (4x1Mb or 4x4Mb), and video card.
This should be enough to get something on screen if the board is still functional.
Let us know how it goes.
It will be good to have another one of these rare guys ticking agaiin.

Following pshipkov's lusty reporting on the SIIG i540 and Advansys ABP-5140 ISA SCSI controllers, I decided I would test them for myself. I'm using a PEM-4036YB motherboard with an 85 MHz crystal oscillator. The first thing I noticed was how my SIIG card came with a 40 MHz crystal osc, while the ABP-5140 has a 50 MHz crystal osc. That is strange. The photos provided by pshipkov show both cards having a 50 MHz crystal.

My ISA is at 10.6 MHz, unless specified otherwise.

I ran some Coretest baseline measurements:

AHA-1522B in coretest: 3870 kb/s
AHA-1542CP: 2899 kb/s (5 mb/s DMA)
AHA-1542CP: 3713 kb/s (6.7 mb/s DMA). 8mb/s hard hangup.

ABP-5140 (50M crystal):
DMA at 5 MB/s: 3209 kb/s
DMA at 8 MB/s: 4851 kb/s
DMA at 10 MB/s: 5872 kb/s
- increasing the ISA from 10.6 to 14.2 or 17 MHz didn't help much (10-20 kb/s only)

SIIG i540 (40M crystal):
DMA at 10 MB/s: 5163 kb/s
- about a 12% performance drop with the 40M crystal compared to the 50M crystal.

In general, the Advansys/SIIG numbers look promising, but only if you can run them at more than 5 MB/s DMA speeds.

The issues for me came when testing them in Windows 3.11. I used the win3.1 installer provided on the SIIG diskette. This basically adds c:\advscsi\advw32.386 to the system.ini file under the 386Enhanced section. I think it is supposed to be the 32-bit disk/file access driver, but looking at the 386 Enhanced section in the Control Panel, it still indicates 16-bit file/disk access. I decided to run Wintune 2.0, but left DMA at 5.0 MB/s for safe keeping. All other SCSI BIOS settings on default.

Wintune 2.0
SIIG i540
738 kb/s uncached. 364 kb/s cached. Yes, cached was slower!

Begin EDIT
Then re-ran the test with C:\windows\smartdrv.exe in autoexec.bat
612 kb/s uncached. 2668 kb/s cached 1st run, 5161 kb/s cached 2nd run.
End EDIT

Then ran the same tests with the AHA-1522B, which loads fastscsi.386 as its 32-bit driver.

Wintune 2.0
AHA-1522B
2287 kb/s uncached. 9051 kb/s cached.

Wintune 2.0
AHA-1542CP, DMA @ 5.0 MB/s
1903 kb/s uncached. 9108 kb/s cached.

pshipkov, are you able to provide some Wintune 2.0 results from within Windows 3.1x for your Advansys and SIIG cards? Something fishy going on - the SIIG i540 bombs in Windows.

The other perplexing questions is why my SIIG card came with a 40 MHz crystal oscillator and if I should upgrade it...

It sounds like upgrade to 50MHz oscillator is in order for the SIIG adapter but socket it, so if performance improves you can go with 60 or even faster crystal. That will be very interesting to see.

Numbers make sense.
These cards destroy Adaptec SCSIs.

The WinTune2 result is not looking good indeed.
Ok, will run a test.
It will be better if you can run the WT2 test with 386DX CPU instead.
I am not sure if my SXL2 chips here can be stable enough at 85MHz.
Also both adapters are currently in transition state for the buck gen.
This way i can deliver results faster, cannot remember where i left the stability at 80+ MHz.
I think i left a record somewhere but cannot find it in this thread, so i may remember incorrectly.

You could just run your PGA132 SXL2 at something like 2x30 (60 MHz) and I'll cross reference that. Or if you still have any custom SXL2 interposer that can do 5 V or any voltage, be it at 75 MHz or 80 Mhz, just run that. It's a quick test. My Wintune 2 numbers are so poor that any numbers you provide with an SXL will be helpful. If your numbers are also shitty, then something is up with Windows drivers. If your numbers are remarkable, then I've goofed up something. Of both are numbers are poor, then it is also possible the Advansys doesn't get along with an SXL in Win3.1 for whatever reason.

Last I recall, you had SXL2 at 75 MHz stable. 80 MHz wasn't quite stable, but I think that was sub-5V.

From my polling, SXL2 at 85 MHz has a good chance at 4.55 V (must us LP regulator, not MIC).

I also need to assemble another SXL for use with buck on the testbed.

Ran a quick test.
DTK PEM-4036YB
Default bios settings (F6)
66mhz crystal
Cyrix.exe -i1 -i2 -m, no clock doubling
Win3.1 no SIIG driver, dos 6.22

These most conservative parameters should be very easy to match on your side.

Interesting. What do your win3.11 cache settings look like? I wonder if my system is only applying cache when an Adaptec 152x is installed, as this is what it is expecting.

EDIT: Looks like I need to load smartdrv.exe from autoexec.bat in order to use non-32bit caching in Windows 3.11. I've only ever used the 32-bit built-in WIN31 caching. What does your autoexec.bat line starting with C:\WINDOWS\SMARTDRV.EXE look like?

What do your numbers look like when you use the 32-bit SIIG win31 driver?