Anonymous Coward wrote:

Ah. So you did.
oops.

I have an AHA-2842A SCSI board, with a WB jumper.

"If your system has write-back, and the system
behaves erratically or does not work, did you make
sure the jumper on jumper block J5 is installed?"

It seems the card is made to take into account the WriteBack.
I'll keep you informed !

Ok, I'v been able to work on this system.

I found a box with MHz LED (2 figures), Which I set on 99MHz turbo and 16MHz no turbo.
I mounted MB, CPU, RAM, SCSI and video. All works fine. Then added SCSI with 2GB HDD and 12X cdrom.
With Write Back, MSDOS won't install, so I fell back to write through. I'll go to 120MHz to make up for it 😁
MSDOS is installed, So next steps will be : AWE32+NEC385, MIDIMAN + MT-32 😀

Ok, this thing is driving me mad ...

After a long pause (moving into a new appartment, working on the P133 and the Athlon 700), I decided to get back to that 486.
It just doesn't want to work.

Currently, it will run several programs with no pb (himem, my packet driver, all from mtcp, all basic msdos commands) but won't run anything like pkunzip, speedsys or chkcpu without hanging/showing a message from EMM386 saying an error has been detected at some adress.

Here's what I tried, the only common denominators being the HOT419 DZ/R2 and the 5GB HDD :
- DX4 100 WB
- DX4 100 WT
- DX2 66 WT
(with correct, as far as i can tell, jumper hell settings)

- DC680-T caching ctrl (cache on/off)
- windbond stupid ctrl

- 64MB FPM
- 32MB FPM
- 16MB FPM

- 2 different VLB ET4000/W32p
- ISA cirrus logic 5429

- BIOS on fail safe
- BIOS on performance
- BIOS manually tuned

- any mix of the previous choices.

I just can't run chkcpu 🙁
Can a faulty HDD give that kind of problems ?

A bad hard drive is generally not going to silently corrupt data. It either reads your data, or it gives an error saying that it could not read the data. It is rare for a hard drive to return bad data.

What you describe sounds more like a RAM or chipset problem, caused either by parts that are running too fast or heat buildup. Even if you are not overclocking these are old parts now and they might have been in spec when new, but out of spec now. Run some memory exercisers and use a few different ones to see what you can shake out. It might take a few hours of runtime to spot a problem.

Are you using a DDO with your HDD ? If not, after a certain amounts of data written on it, it can cause a real mess... corrupted data, programs that would fail to start and such things. 😵

If you are, it's a bit extreme but...
I would begin to take the motherboard out of the case (like this I can take a look with a magnifying glass and search for damaged compoenents if any) then clean all the amovible
parts contacts with sheets of paper (SRAMS, BIOS, RAM's) and use a toothbrush with isopropyl alcohol for AT pins, SIMM slots, isa/vlb slots, bios and SRAM's sockets.
I will do the same for all the cards I'm planning to use for testing, even if they look clean.

I will even pull out the cmos battery, just for the time I will test it.
Finally, I will just use a clean boot floppy with the utilities you described instead of the hdd.

If it doesn't change anything, you can always try to add WS and see if it improves its stability.
If not, I really have bad statements for you 😢

sunaiac wrote:

Hi,

- Tseng Labs ET4000/W32p VLB (http://imageshack.us/photo/my-images/443/img1441.png/)
-

Didn't the /W32p indicate it was a PCI part? I had an ET4000/W32 VLB card back in the day, and it handled every DOS game I threw at it. I forget what the /W32i designator indicates...

And wasn't that one of the first Tseng Labs board to feature a precision RAMDAC? Or am I confusing it with the ET6000?

darksheer : DDO ?

I tried memtest, with only the tekram (cache off), a ISA GL5429, and one 16MB stick : reboot when memtest starts 🙁

edit : same with 1 4MB stick.
with EDO, there's no boot, so it's not because i put wrong memory.

So either I have a bad jumper setting, or my motherboard has a serious problem, given I tried completely different extension cards and CPUs and memories.
I also tried with or without cache (by bios settings) for L2.

I'll try the complete cleanup, but well ... 😖

DDO stands for Dynamic Drive Overlay, it's a sort of software that overcomes the 200/500 MB limit of ISA/VLB IDE HDD controllers. Without it, if you use a >200 or >500 MB HDD with an old HDD controller it won't see more space than its limit (basically, without it, even if you have a 2GB HDD and you're using fdisk under dos, you won't be able to create bigger partition than the limits). I know, it's a very dumb way of explaining it 🤣

THE POINT (if you're not using it) is that if you want to play it "genius" by creating a full HDD size partition on another PC that see its full capacity (a 2 GB FAT16 partition for example), and put HDD back into your old computer with the old controller... that will WORK fine UNTIL you hit the multi I/O limits and then... you're screwed 🤣 The HDD controller will progressively corrupts data on the drive (because it will re-writting on data < the limits) and you will experience programs and commands that will not work anymore (and other shitting things) until it will not be able to correctly boot to dos 😵

Poor motherboard, I really hope it's not dying all of a sudden... and it's just a bad contact issue 😳

Edit : Did you tried to add wait states or delay for the ram or cache under the BIOS, or with jumpers on the motherboard ? That will drop performance down a bit but greatly improve stability 🤣

Ok, there is no DDO 😀
It's a 5GB HDD, running on a tekram DC680T which can see all the capacity.
I did not activate LBA in the bios (I actually forgot) before installing MSDOS so I have a 500MB partition only.

Anyway, since I have the problem with only a floppy and memtest, I guess the HDD is not my primary concern right now 😁
My crashes being not random at all, I really think i'll have to check jumpers to while cleaning what can be cleaned ...

OK, funny thing : put a different CPU (DX2 66), different MB (486SLE chaintech), different memory (4x4MB 30pins), and different Ctrl (some goldstar VLB ctrl), and the thing still reboots on memtest86 and memtest86+ launch 😁

I'll try another PSU and another floppy, and then i'll have tried 2 completely different systems behaving exactly the same ...

😳 Have you tried different versions of memtest ? 😕 That would drive me completely mad 😵

tried memtest86 4.3 and memtest86+ 5.1 😁

Try older! I have the same issue on my 486 and older pcs!

Do you remember what you used ?

Memtest86+ 4.00 should work fine because keropi used it for an older motherboard here : How to enable L2 cache on a VI11 486 mobo?

darksheer wrote:

Memtest86+ 4.00 should work fine because keropi used it for an older motherboard here : How to enable L2 cache on a VI11 486 mobo?

Working ! (at least on that DX2-66 setup)
Thank you 😊
Now on to testing my DX4 100 setup 😀

Memtest fine on my HOT419 R3 with BIOS on failsafe and on optimal with a DX2 66.

I'll reinstall DOS after testing the whole hard drive with a low level format by the controller.
I still have no explanation as to why I was unable to start a whole bunch of programs when others work flawlessly ...

which configuration would you use, this my mother board doesn't clearly support write back DX4 : http://www.elhvb.com/supportbios.info/Archive … 19htm/419dz.htm

The Intel DX4 &E... column with just the writeback jumper set as in the P24D column ?

Found the problem !

After a complete reinstrall, most things were working and some other still not.
I'm on the DX2, so not a WT/WB problem.

Actually, the things not working are the ones coming from the network !!!
Whenever I get something from the ftp with mTCP, it's screwed on arrival.
Be it an exe that will hang, or a zip that becomes corrupt ...

my FTP is on a linux server, I'll try with a ftp from a windows machine, with and without the switch in-between.
Maybe something on my network doesn't like a 10Mbps half duplex network card.

Hint for others : type "image" in mtcp ftp client to avoid the ascii mode screwing everything ...
I tested sio many hardware setups just because of a configuration proclem ... 🙁

Well at least it's done 😀

Glad that you have resolved your issue, you must feel much better now 😁
That's why i'm a bit reticent to use lan on my 486 🤣 I'm using a cd-rw or I directly transfer on the cf for big files and use floppies with pkzip/pkunzip for small ones 😀