VOGONS


First post, by justin1985

User metadata
Rank Member
Rank
Member

I've used 128Gb mSATA drives with the plastic 2.5" case type jMicron based generic 44-pin IDE adapters in several retro PCs over the last few years, and generally found them a really good solution. However, recently I've been having trouble with read/write corruption errors, and wonder if I've managed to cause a "chain reaction" of failures?

When I re-built a system using one of the 128Gb mSATA drives into a different case, I mounted the drive into the 3.5" bay (on an adapter tray) before connecting the IDE cable, and it was quite inaccessible and very difficult to line up and connect. I didn't notice that I'd connected the 44-pin IDE cable a row too low until I'd already powered on and seen no IDE drive connected. I powered off, and tried reconnecting it, and it seemed to work, although the Win98 installation was corrupted. I thought no big deal, formatted C: and reinstalled from the installer that was on a different partition - but then half way through Win98 installation I got a fatal error mentioning a .drv file at around 65%. I figured the files might have been corrupted when the drive was misconnected, so took the drive out, re-formatted all partitions via a USB adapter on a Win10 system, but then got a very similar error during Win98 installation (naming a different file).

So, perhaps the drive itself was damaged. I swapped the mSATA drive for a known good one, into the same adapter, re-formatted that, ran Win98 install again, but got a very similar error around the 65% stage of Win98 installation.

Perhaps it was the mSATA - IDE adapter that was broken - so I tried swapping that for a known good one - but with both the original and spare 128Gb drives in the new adapter - I got exactly the same errors! So now every combination of mSATA drive and adapter now has very similar errors during Win98 installation - Scandisk wasn't returning any errors during Win98 installation though. The motherboard (socketA with SiS chipset) now works fine with a CF card adapter connected to that IDE channel.

I clearly killed either the original mSATA drive or the original mSATA-IDE adapter (or both) by connecting the cable incorrectly - my bad. But could swapping drives and adapters have caused a 'chain reaction'? Could connecting the good mSATA drive to the potentially damaged adapter have killed it? And then connecting the bad drive to a new adapter have killed that too?

It feels like trying the mSATA drives in a passive mSATA-SATA adapter is the next logical step. But could that do any harm to anything too?

(or might it be possible that the SiS chipset seems to see the jMicron adapters OK, but actually some incompatibility leads to the errors?)

Reply 1 of 8, by Thermalwrong

User metadata
Rank Oldbie
Rank
Oldbie

Do your msata to IDE adapters have a voltage regulator on them or do they connect directly to the 5v pins?

Reply 2 of 8, by justin1985

User metadata
Rank Member
Rank
Member
Thermalwrong wrote on 2025-02-18, 03:28:

Do your msata to IDE adapters have a voltage regulator on them or do they connect directly to the 5v pins?

They certainly seem to have voltage regulators, and there's no obvious sign of any damage to them (I had smoked a similar adapter in the past by making a similar mistake and that had a visibly melted voltage regulator).

This type:

The attachment Screenshot_2025-02-18-07-56-08-86_3aea4af51f236e4932235fdada7d1643~2.jpg is no longer available

Reply 3 of 8, by darry

User metadata
Rank l33t++
Rank
l33t++

A failure/corruption at around 65% of installing, using multiple combinations of msata adapters and drives seems oddly specific and reproducible.

Why and especially how could the adapters all be apparently working for read and write operations until a specific point and then fail at the same moment during Windows installation/disk writing ?

I mean, practically everything is possible, but ths seems unlikely to be due to the adapters failing, IMHO.

Some unnoticed variable may have changed (BIOS setting, for example).

Reply 4 of 8, by Deunan

User metadata
Rank l33t
Rank
l33t

As I recall the Win9x installers are quite buggy. It was discovered when people started integrating various patches and driver packs directly into them, to make the whole installation process faster. As well as allow it at all on modern HW that was not recognized or supported. IIRC even the CAB file sizes sometimes matter - while it seems like just an archive file, the installer will crash and burn if you touch some of them. Then some other ones can be easily modified without ill effects.

TL;DR: It might be Win9x being dumb that is the problem. Especially if the install image is one of those custom patched ones. Try a different one, preferably pure MS version if you can, or if you need some extra patches or drivers try to build your own custom image with just the things you need. Maybe that will help. Oh and don't pay too much attention to what particular file or CAB is reported as damaged by the installer. It might be that, or it might be just some side effect of the running executable already being corrupted and doing weird things and failing.

Reply 5 of 8, by kotel

User metadata
Rank Oldbie
Rank
Oldbie

You could try to read smart data from the SSD's using e.g. crystal disk info or hdtune.

"All my efforts were in vain...
Let that be my disappointment."
-Kotel

Reply 6 of 8, by justin1985

User metadata
Rank Member
Rank
Member
Deunan wrote on 2025-02-18, 12:17:

As I recall the Win9x installers are quite buggy. It was discovered when people started integrating various patches and driver packs directly into them, to make the whole installation process faster. As well as allow it at all on modern HW that was not recognized or supported. IIRC even the CAB file sizes sometimes matter - while it seems like just an archive file, the installer will crash and burn if you touch some of them. Then some other ones can be easily modified without ill effects.

TL;DR: It might be Win9x being dumb that is the problem. Especially if the install image is one of those custom patched ones. Try a different one, preferably pure MS version if you can, or if you need some extra patches or drivers try to build your own custom image with just the things you need. Maybe that will help. Oh and don't pay too much attention to what particular file or CAB is reported as damaged by the installer. It might be that, or it might be just some side effect of the running executable already being corrupted and doing weird things and failing.

Interesting thoughts! It is an original Win98SE Full Retail installer downloaded from WinWorld, so no patches or anything pre-applied. The same installer directory copied to the CF card worked fine, so that seems unlikely unless the installer directory had become corrupted on my NAS in the meantime.

darry wrote on 2025-02-18, 11:01:
A failure/corruption at around 65% of installing, using multiple combinations of msata adapters and drives seems oddly specific […]
Show full quote

A failure/corruption at around 65% of installing, using multiple combinations of msata adapters and drives seems oddly specific and reproducible.

Why and especially how could the adapters all be apparently working for read and write operations until a specific point and then fail at the same moment during Windows installation/disk writing ?

I mean, practically everything is possible, but ths seems unlikely to be due to the adapters failing, IMHO.

Some unnoticed variable may have changed (BIOS setting, for example).

It's been both weirdly reproducible in terms of happening towards the end of the file copy stage of the installer, but also a different error message each time. One time it was an explicit DOS prompt type text "A serious disk IO error has occurred" message superimposed over the installer GUI, but other times its been the plain white type pop-up window with an error relating to a different specific driver each time (I seem to remember vga.drv one of the times).

I can't help wonder if it is something specific to this motherboard/chipset/BIOS and the 128Gb drives ... The BIOS was all reset to defaults, as I had stored it with battery out.

I have to confess to being a little scared of connecting these drives to other systems so far, for fear of propagating problems - but logically there isn't any way that should be able to happen, right?

kotel wrote on 2025-02-18, 12:26:

You could try to read smart data from the SSD's using e.g. crystal disk info or hdtune.

I'll have to give that a try, thanks for the reminder. Is there any DOS based SMART reporting tool?

I did wonder if it was a 'fake' SSD reporting 128Gb but actually being smaller, hence hitting an error only after a certain point, but the same has happened with both an AliExpress cheap brand mSATA, and a second hand very much genuine seeming Toshiba mSATA. More significantly, both mSATA drives had worked fine for months in different systems.

Reply 7 of 8, by kotel

User metadata
Rank Oldbie
Rank
Oldbie
justin1985 wrote on 2025-02-19, 12:00:
kotel wrote on 2025-02-18, 12:26:

You could try to read smart data from the SSD's using e.g. crystal disk info or hdtune.

I'll have to give that a try, thanks for the reminder. Is there any DOS based SMART reporting tool?

I did wonder if it was a 'fake' SSD reporting 128Gb but actually being smaller, hence hitting an error only after a certain point, but the same has happened with both an AliExpress cheap brand mSATA, and a second hand very much genuine seeming Toshiba mSATA. More significantly, both mSATA drives had worked fine for months in different systems.

Victoria, mhdd, hdat2, atapwd
Chose your pick
Note atapwd requires SATA in ide mode

"All my efforts were in vain...
Let that be my disappointment."
-Kotel

Reply 8 of 8, by Xanxi

User metadata
Rank Newbie
Rank
Newbie

I'm using mSata SSD with IDE adapters like this one or smaller one without boxes for years into my several Amiga 1200 computers with no problem, and even one in an older Windows XP laptop (works but i didn't care about the alignement and of course there is no TRIM, meanwhile the Amiga can't wear the SSD enough to worry about that).
However i use only reputable brands of SSD, like Transcend, and avoid the fake chinese brands which have been guilty before of faking the SSD size. I don't see if you mentionned the maker of yours but it is possibly failing at 65% because it is reporting a fake size.