Memory gone dodo?

you mean system ram?

Yes, that always reminds me of the time when mine got bad and lost me data.

If it turns out not to be the memory, you could try swapping IDE cables 😉

I remember some buggy chipsets (mostly VLB) that had corruption issues. What IDE controller are you using?

the board is a DFI K6XV3+/66 (VIA MPV3), I'm using IDE round cables bought two months ago.

just bought 512MB SDRAM on ebay for 2,50€, I'll keep you informed 😀

hi

although memory is an obvious culprit here i wish
to add one comment which may not be relevant
but worth a look.

i have an MVP3 based board and it experienced
a fault with a PCI network card which
we may assume also uses DMA. to rectify that
i swapped out that NIC for another.

you may want to test this setup with all isa/pci
cards removed to determine if there is another
offending device on the bus. i don't have much
experience with MVP3 since the board i have
is quite a recent acquisition but i get the impression
that there are bugs with it's bus mastering
DMA engine support.

bye

following cards are installed:

- agp geforce4
- pci sb live
- pci voodoo1
- isa awe64
- isa gus pnp
- isa scc1

search for SBLIVE! + MVP3 + UDMA

this may be the problem

good luck

Do you use IDE cables with 80 wires for UDMA ?
And do you use the correct connector locations?
It is B--S-M.

sklawz wrote:

search for SBLIVE! + MVP3 + UDMA

this may be the problem

good luck

That sounds like the problem to me Crosstalk between the hard drive controller and the sound card so that whenever the sound card and hard drive are active at the same time the data gets corrupted. So would the solution be to reassign the sound card to a different IRQ or is it unfixable?

hi

i remember reading about those sblive! problems
years ago and thanks to the internet the posts
still exist for reference.

i also remember people talking about adjusting
the PCI latency timer, moving the card to a
different slot but have no personal experience
of this fault.

there are two variants of the sblive that i am
aware of though which may be different.
one is the older 5V PCI and other other the
3.3V+5V universal PCI. if the Sblive is found
to be the problem it may be worthwhile trying
the other variant.

bye

edit: typo

the board is a DFI K6XV3+/66 (VIA MPV3), I'm using IDE round cables bought two months ago.

While conflicts with the SB Live are certainly a possibility as well, I see two common culprits right here. Some versions of the Via chipset drivers had IDE issues, so you may want to try different drivers. Secondly, a great many of those rounded IDE cables are poorly constructed and prone to interference and intermittent connections... I've tried several of different types and from different sources over the years, and none of them have worked properly.

following cards are installed: […]
Show full quote

following cards are installed:

- agp geforce4
- pci sb live
- pci voodoo1
- isa awe64
- isa gus pnp
- isa scc1

One other possibility to try is changing the card placement order. In my Soyo VIA chipset builds with Sound Blaster Live!, I am generally obliged to put the SB Live! card at the opposite end of the bus from the AGP slot. So at the least, I would swap the Voodoo and SB Live! cards, or if you can, put it in the slot most distant from the AGP slot

the sblive is currently in PCI slot #2. In the PCI device list I can see that it shares the IRQ with the ACPI device manager. The voodoo1 is in slot #3, I think I will try to swap both cards.

okay putting the live in slot #3 did not change anything

It could also be a compatibility problem between chipset/mobo and your specific hard disk.

And you may want to try installing George Breese's PCI Latency patch for those VIA chipsets. Old VIA boards often have a shoddy BIOS that poorly sets PCI options.

swaaye wrote:

And you may want to try installing George Breese's PCI Latency patch for those VIA chipsets.

genius, this worked 😎

Nice.

I once had that patch make a SBLive go from just screetching nasty noises to perfect functionality on a VIA Apollo Pro Slot 1 board. Some of those VIA boards really have their PCI bus programmed very wrong.