zilog256 wrote on 2021-03-25, 10:21:

IDE/SATA converter on the PC

Not all IDE to SATA converters "play nicely" with all chipsets, how many different ones have you tried?

Have you seen A very persnickety PIII machine

Might be worth it to pickup an IDE native SSD. Performance is usually underwhelming and they cost more, but you might have better success if there is no adapter between the ide controller and the drive.

You also can boot from PCI SATA controller. And since TUV4X use 686B south bridge - native IDE ports are also connected to PCI bus, unlike on more modern motherboards. So speed should be the same or slightly better.

red-ray wrote on 2021-03-25, 10:43:

zilog256 wrote on 2021-03-25, 10:21:

IDE/SATA converter on the PC

Not all IDE to SATA converters "play nicely" with all chipsets, how many different ones have you tried?

Have you seen A very persnickety PIII machine

Owww! Thanks ! it's the obvious common hardware I didn't even bother to reconsider.....

It's an "el cheapo" Aliexpress model, I thought they were all the same.

If you have a link/reference to a particularly "good" one on Aliexpress, I'll take it 😀

I'll have a look at the link, thanks!

BinaryDemon wrote on 2021-03-25, 10:56:

Might be worth it to pickup an IDE native SSD. Performance is usually underwhelming and they cost more, but you might have better success if there is no adapter between the ide controller and the drive.

Hmm yes but they are outreageously expensive, and I have a bunch of SATA 2.5' and a few mSATA, not yet ready to settle for that 😀

The Serpent Rider wrote on 2021-03-25, 12:13:

You also can boot from PCI SATA controller. And since TUV4X use 686B south bridge - native IDE ports are also connected to PCI bus, unlike on more modern motherboards. So speed should be the same or slightly better.

I have 1 or 2 of these in a drawer I guess, thanks, worth a try!

Maybe I can also get my hands on a IDE/SATA adapters that came with my x58 Gigabyte motherboard "back in the days"

I'll get back to you all after following these leads.

Maybe the VIA chipset hates ssds/converters?
Better getting a SD-IDE adapter.
Fast enough for my C3 rig.

This might sound like an odd question, but are you using an 80-conductor ribbon cable with the SATA-to-IDE adapter. I've found that these adapters are pretty strict about using high-speed Ultra-DMA modes, and don't conform well to the older speed standards (i.e. not fully backward compatible). The TUV4X should be able to handle UDMA/100 with the proper cable.

EDIT: They also don't play well with a second device on the same cable (despite the Master/Slave jumper).

Here are some of the threads where I have posted about my experiences with SATA to IDE adapters . Maybe they will be helpful . (this is taken Verbatim from another thread discussing a possibly related issue : Windows 98 SE on SSD 128 GB )

Re: Master/Slave option for modern storage
Re: Planning to install SSD in my Windows 98 machine. Help, advice, suggestions, etc.?
Re: Gigabyte 6bxc with PCI SATA HBA

Thanks once again for all the answers/ideas, folks !

MKT_Gundam wrote on 2021-03-25, 12:44:

Maybe the VIA chipset hates ssds/converters?
Better getting a SD-IDE adapter.
Fast enough for my C3 rig.

I might end that way if nothing else works, CompactFlash is not that bad.

fosterwj03 wrote on 2021-03-25, 14:19:

This might sound like an odd question, but are you using an 80-conductor ribbon cable with the SATA-to-IDE adapter. I've found that these adapters are pretty strict about using high-speed Ultra-DMA modes, and don't conform well to the older speed standards (i.e. not fully backward compatible). The TUV4X should be able to handle UDMA/100 with the proper cable.

EDIT: They also don't play well with a second device on the same cable (despite the Master/Slave jumper).

I'm using a 80wire ribbon, yes, with only one device attached, the HDD/SSD 🙁

The cable is rather old and twisted, though,I'll try a new one, thought about it then forgot ; I might even use a 40wire ribbon as last resort

darry wrote on 2021-03-25, 14:40:

Here are some of the threads where I have posted about my experiences with SATA to IDE adapters . Maybe they will be helpful . ( […]
Show full quote

Here are some of the threads where I have posted about my experiences with SATA to IDE adapters . Maybe they will be helpful . (this is taken Verbatim from another thread discussing a possibly related issue : Windows 98 SE on SSD 128 GB )

Re: Master/Slave option for modern storage
Re: Planning to install SSD in my Windows 98 machine. Help, advice, suggestions, etc.?
Re: Gigabyte 6bxc with PCI SATA HBA

I'll have a look, thank you!

=> I tried the only PCI expansion card I have : it's a ...via chipset with 2 SATA and 1 UDMA100/133 : the "mass storage device" is seen by the BIOS init, but no boot at all (tried SSD via SATA and the original HDD on 80wire new cable). I then tried the same PCI expansion card on another rig (socket 939) => no boot whatsoever, but the card is seen and initialized under the Windows XP boot, and the connected drive is recognized etc ; so either these motherboards don't allow booting on PCI expansion cards, or the specific one I have has no ROM/option to do it 🙁

i thought I had a Sil3112 card somewhere, can't get my hands on it, but that should do the trick if can find it!

I'll go fiddle with the 80wire ribbons ; and probably try ordering a few other IDE/SATA or CF/IDE cards 🙁

EDIT:

=> tried a new IDE/SATA adapter (same model, spare one, just new/still packaged), a brand new 80wire ribbon = same thing ; even tried a 40wire ribbon, same thing (Windows XP reboots - W98SE boots flawlessly)

IDEA: I usually use SDI update to install the drivers..... maybe I should try the "famous" Hyperion 4 in 1 by hand....not the most reliable ones, but, hey...'nuff said...."Via"..... 🙁

EDIT2: considering to buy a Sil3114 PCI card for 10 euro, but it seems the BIOS won't recognize it as a bootable device either of P3/KT133 rigs (?) : Booting from a Sil3114 .... well, maybe with and older BIOS flash..... Silicon Image Sil3112 PCI SATA Controller - looking for old BIOS ; I can also try bios-mod the motherboard, integrating the bootROM for my current (lousy) VT6421a chip SATA PCI card.... https://www.wimsbios.com/forum/depth-high-tec … ios-t10679.html

So, I tried to inject my VT6421a chip SATA PCI card ROM into the TUVX4's BIOS => got my hands on CBROM32_140, VT6421a ROM, edited the file, found AFLASH, formatted a floppy, flashed the BIOS, no problem, but still no boot from the PCI card (either IDE HDD or SATA SSD) ; the boot options in the BIOS don't show anything special as for an external boot, except Network, and SCSI that I tried to no avail.

The ID/Vendor in the injected ROM file are the right ones 1106 / 3249

Either I did things wrong once again, or it can't work with that card. So I'll just invest in some other cards / converters at that point I guess.

Sorry for auto-replying again :

I tried something : plugged the 40 Gb Seagate as MASTER, and a Fujitsu 80 Gb as SLAVE on the UDMA100 controller, booted on an Acronis True Image CD, done a HDD clone with partition automatic resize.

Well....W98SE works on the clone, but the XP partition stays stuck on a small "Windows XP" logo, even when trying the Safe Mode via F5.... !?

EDIT: tried cloning on a 40 Gb (exact same size) WD HDD, same result, so size is not an issue, something goes wrong with cloning, no idea why.

=> a fresh install on 2 different HDDs is in order....

zilog256 wrote on 2021-03-27, 18:58:

Sorry for auto-replying again : […]
Show full quote

Sorry for auto-replying again :

I tried something : plugged the 40 Gb Seagate as MASTER, and a Fujitsu 80 Gb as SLAVE on the UDMA100 controller, booted on an Acronis True Image CD, done a HDD clone with partition automatic resize.

Well....W98SE works on the clone, but the XP partition stays stuck on a small "Windows XP" logo, even when trying the Safe Mode via F5.... !?

EDIT: tried cloning on a 40 Gb (exact same size) WD HDD, same result, so size is not an issue, something goes wrong with cloning, no idea why.

=> a fresh install on 2 different HDDs is in order....

Maybe try Clonezilla

Back after a few days of struggle on my spare time on this damn setup.

So it appears it's not really a problem with the SSD "per se". I'm now using a 60 Gb SSD with my previous adapters, dual boot 98SE (from previous clone) / XP (from scratch, SP3)

Sometimes XP boots flawlessly, sometimes it loads forever with a hardware black screen and reboot at the end, and sometimes a BSOD "STOP etc". (of course, Memtest86 OK, CPU@stock etc)

I've narrowed down the problem: it seems to be an IRQ conflict.

=> I initially had a SB128 on PCI1, an Ethernet card on PCI5, and a Nec USB 2.0 card on PCI6.

I've removed the cards, tried changing slots etc => no matter what I do, if there is something else than the SB128 on any slot, things may look OK at one boot, but bugs appear on the next reboot, cards are not seen / don't start the driver etc.

XP changes the IRQ of the cards randomly at each reboot, OK it's "normal" with ACPI and switching control to the OS for the IRQs instead of the BIOS ; but if only the settings were STABLE. Most of the time, it puts the SB128 and other cards on IRQ9.

So I tried it "the old way" (I had such machines "back in the day" but don't recall having so much problems, even with all PCI populated, SCSI cards etc) => "PnP OS = NO", and trying to set IRQ manually (classical 5 or 7 for the SB128 etc) => well, it either fails miserably with no boot/blackscreen/BSOD, or it boots but the IRQ under XP is NOT the one I've set in the BIOS ; I've also taken care of populating the PCI ports without shared IRQs (PCI1&5 shared, PCI2&6 shared etc), doesn't make a difference. I suppose there is no way to force IRQs under the BIOS, so I will try all the combinations on the 6 slots, the main problem being that what works during one boot may not the next time 🙁

As things are, I can declare myself beaten and keep only the SB128 on PCI1 and no other card, but the USB2 and Ethernet were quite handy....

Am I missing an obvious thing here?

since PCI has only four IRQ lines, any board with more PCI slots either has to use a bridge chip or suffer the consequences of shared interrupt lines, which shouldn't have been a problem but reality denied that idea.
Add to that the fact that the onboard pci devices, like IDE controllers, NICs, firewire, USB also tend to share the same interrupts.. and yeah... getting a lot of PCI cards to cooperate tends to be a struggle if the cards aren't well behaved. Add to that, that USB2 cards tend to use multiple interrupts, maybe PCI soundcards as well... yeah good luck ! 😀
and Good on you for narrowing down the cause of the problem to end up with a usable system.

weedeewee wrote on 2021-03-31, 20:19:

since PCI has only four IRQ lines, any board with more PCI slots either has to use a bridge chip or suffer the consequences of shared interrupt lines, which shouldn't have been a problem but reality denied that idea.
Add to that the fact that the onboard pci devices, like IDE controllers, NICs, firewire, USB also tend to share the same interrupts.. and yeah... getting a lot of PCI cards to cooperate tends to be a struggle if the cards aren't well behaved. Add to that, that USB2 cards tend to use multiple interrupts, maybe PCI soundcards as well... yeah good luck ! 😀
and Good on you for narrowing down the cause of the problem to end up with a usable system.

Thanks for the moral support, but I can't believe it's impossible to populate juste TWO PCI slots on such a motherboard 😁

zilog256 wrote on 2021-03-31, 20:22:

weedeewee wrote on 2021-03-31, 20:19:

since PCI has only four IRQ lines, any board with more PCI slots either has to use a bridge chip or suffer the consequences of shared interrupt lines, which shouldn't have been a problem but reality denied that idea.
Add to that the fact that the onboard pci devices, like IDE controllers, NICs, firewire, USB also tend to share the same interrupts.. and yeah... getting a lot of PCI cards to cooperate tends to be a struggle if the cards aren't well behaved. Add to that, that USB2 cards tend to use multiple interrupts, maybe PCI soundcards as well... yeah good luck ! 😀
and Good on you for narrowing down the cause of the problem to end up with a usable system.

Thanks for the moral support, but I can't believe it's impossible to populate juste TWO PCI slots on such a motherboard 😁

probably depends on the PCI cards & their respective drivers that you're using 😁

The drivers are directly included in the XP SP3 package, as far as I recall (didn't have to install them manually)

Have you tried Startech PATA to SATA adapters? This use marvell chipset bridge and is supposed to work properly regardless, for this reason I specifically purchased at little more expense. All others is junk jmicon chipset clone of clones with compatibility and reliability issues.

Secondly, do not use USB controller card till you are sure your computer working right first especially if you get the Startech adapters first.

Thirdly, I always reserve the interrupts for each PCI channel that each card needs in view of windows 98SE.

Fourth, quality motherboard is the one which make the difference. In all my experience in past if anything changes every time it is booted up always are either one of card is not working right or motherboard bios is buggy, update the bios to current version. Swap one of cards for different brand or different one at a time till the configuration stays same every time it is booted.

Fifth, I don't like VIA chipset and I had one with socket 462 (Athlon) on Asus board and the configuration did not change when I was running win 98SE but crashing and lock up was common back in the day, no matter what I try any 4 in 1 drivers and all the other drivers.

Cheers,

pentiumspeed wrote on 2021-03-31, 22:10:

Have you tried Startech PATA to SATA adapters? This use marvell chipset bridge and is supposed to work properly regardless, for this reason I specifically purchased at little more expense. All others is junk jmicon chipset clone of clones with compatibility and reliability issues.

Secondly, do not use USB controller card till you are sure your computer working right first especially if you get the Startech adapters first.

Cheers,

I agree that the Marvell one is likely the best and most compatible. The real Jmicron ones are good, IMHO, when they are compatible .

My experience so far is resumed in this post:

darry wrote on 2020-11-21, 17:16:

My experience with SIL3114 based cards has been good, both functionally and compatibility-wise, in 440BX boards (Asus P3B-F and […]
Show full quote

My experience with SIL3114 based cards has been good, both functionally and compatibility-wise, in 440BX boards (Asus P3B-F and Biostar M6TBD OEM) .

On a board of such vintage, because of LBA issues, I too prefer using either a PCI SATA controller or a PCI IDE controller that works well with IDE to SATA converters (Promise Ultra133 and JM20330 works well together, but not on my P3B-F).

IDE to SATA adapter compatibility varies widely .

The two best ones are, IMHO, those based on
- Marvell 88SA8052 (rather expensive and harder to find but reported as working on Intel ICH4 and Via 686B IDE controllers)
- Jmicron JM20330 (inexpensive and easy to find, but has issues on Intel ICH4 and Via 686B IDE controllers, works fine on Promise Ultra133 and can work fine on ICH2, depending on driver)

I have never tested either of those in DMA mode on a 440BX's 82371EB (PIIX4E). I have done limited testing with a JM20330 in such a setup, but only in PIO mode under DOS . It did seem to work in that scenario .

See mSATA to IDE Trouble

darry wrote on 2021-03-31, 22:18:

I agree that the Marvell one is likely the best and most compatible. The real Jmicron ones are good, IMHO, when they are compati […]
Show full quote

pentiumspeed wrote on 2021-03-31, 22:10:

Have you tried Startech PATA to SATA adapters? This use marvell chipset bridge and is supposed to work properly regardless, for this reason I specifically purchased at little more expense. All others is junk jmicon chipset clone of clones with compatibility and reliability issues.

Secondly, do not use USB controller card till you are sure your computer working right first especially if you get the Startech adapters first.

Cheers,

I agree that the Marvell one is likely the best and most compatible. The real Jmicron ones are good, IMHO, when they are compatible .

My experience so far is resumed in this post:

darry wrote on 2020-11-21, 17:16:

My experience with SIL3114 based cards has been good, both functionally and compatibility-wise, in 440BX boards (Asus P3B-F and […]
Show full quote

My experience with SIL3114 based cards has been good, both functionally and compatibility-wise, in 440BX boards (Asus P3B-F and Biostar M6TBD OEM) .

On a board of such vintage, because of LBA issues, I too prefer using either a PCI SATA controller or a PCI IDE controller that works well with IDE to SATA converters (Promise Ultra133 and JM20330 works well together, but not on my P3B-F).

IDE to SATA adapter compatibility varies widely .

The two best ones are, IMHO, those based on
- Marvell 88SA8052 (rather expensive and harder to find but reported as working on Intel ICH4 and Via 686B IDE controllers)
- Jmicron JM20330 (inexpensive and easy to find, but has issues on Intel ICH4 and Via 686B IDE controllers, works fine on Promise Ultra133 and can work fine on ICH2, depending on driver)

I have never tested either of those in DMA mode on a 440BX's 82371EB (PIIX4E). I have done limited testing with a JM20330 in such a setup, but only in PIO mode under DOS . It did seem to work in that scenario .

See mSATA to IDE Trouble

Jmicron chipset still had issues back then when I was watching this as well. Currently I have not found one that is using geniune chipset sold by other brands not chinese.

DOS does not support anything except the int13 calls hard ware motherboard compatibility is important means adapters and cards has to be designed correctly. Even the motherboard had detected as DMA mode or PIO. It comes down to the quality of adapter design than anything else.
Again, that is why I say try to get startech's marvell chipset based adapter first and try again.

Cheers,