Have you tried using a different IDE cable?

I haven't! That didn't occur to me, heh thank you.

When I get a new IDE cable though, is there anything I can run to test that the system is stable?

Not that I'm aware of... I don't know of any tests you could perform that would give you a definitive 'your controller is fine / faulty'.

The IDE controller is integrated into the motherboard's southbridge (Intel ICH2 in your case). I don't know of a way to isolate it from the rest of the system from a hardware standpoint. You could disable the IDE controller in the BIOS and use an expansion card. Or you could try installing an alternate chipset driver. And if those workarounds ended your problems then you could surmise that the IDE controller within the motherboard southbridge was faulty.

Take a known working blank drive, or one with nothing you care about, and install it in the machine in question. Then install some type of operating environment capable of networking. Transfer a large volume of files over the network to the machine in test, and then transfer them back and perform a binary compare to the original copies.

If this completes without error, I would have full confidence in the storage sub system

Thanks for the advice so far. I was worried that there wouldn't be any good diagnostic stuff, yeah. It also didn't occur to me that the chipset driver might be incorrect or faulty. I'm using the Intel (application accelerator?) driver from MSI's website, since I don't have a Medion recovery CD for this system.

maxtherabbit wrote on 2021-06-16, 02:19:

Take a known working blank drive, or one with nothing you care about, and install it in the machine in question. Then install some type of operating environment capable of networking. Transfer a large volume of files over the network to the machine in test, and then transfer them back and perform a binary compare to the original copies.

If this completes without error, I would have full confidence in the storage sub system

Unfortunately, that's exactly what I did do, making and transferring some ISO images, several times back and forth. Games like Quake 2 and Turok 1&2 were playing without issue. A week later, mystery corruption.

I guess all I can do is repeat the process and perform the network transfers a couple of times a day again and again until I'm confident. (Also since Win98 doesn't come with an MD5 hash utils, I've found this one to be useful.)

If data integrity of the IDE interface is suspected, I would say go in the BIOS and try to downgrade the performance level. Maybe you can drop it from ATA66 to ATA33, or DMA mode 2, or even PIO. Slow things down and see if it becomes stable. I had to do that for a certain HDD on a Pentium III once.

My quick and easy way to test stability on DOS machines (including CPU, memory, and disk) was to PKZIP a bunch of files and then unzip it again. If any data corruption occurred it would cause a CRC error during the unzip.

I'd suspect the SD-IDE adapter.

Have you tried using a real hard drive ?

weedeewee wrote on 2021-06-17, 06:34:

I'd suspect the SD-IDE adapter.

Have you tried using a real hard drive ?

I bought the SD adapter because of weird behaviour I was getting with the original hard drive. :)

MrD wrote on 2021-06-17, 07:11:

weedeewee wrote on 2021-06-17, 06:34:

I'd suspect the SD-IDE adapter.

Have you tried using a real hard drive ?

I bought the SD adapter because of weird behaviour I was getting with the original hard drive. 😀

how old is the psu and
has the board and psu been recapped ?

Both the PSU and the board are twenty years old and not recapped. I've just bought a replacement motherboard of the same type (with its own CPU and memory) and I'm having the same issues with this SD card adapter! I definitely recommend not buying Medion hardware! 😅

If data integrity of the IDE interface is suspected, I would say go in the BIOS and try to downgrade the performance level. Maybe you can drop it from ATA66 to ATA33, or DMA mode 2, or even PIO. Slow things down and see if it becomes stable. I had to do that for a certain HDD on a Pentium III once.

Yeah I've been trying that, but not with any sort of repeatable effect. 🙁

Oh, it's a medion... I once had a medion sixteen years ago. After a few years, I had to replace the psu, and a few more years later, it started hanging pretty regularly, probably needed recapping or some other component or connection on the board started flaking out. forgot about it for several more years, and eventually it got thrown out.
if I still had it laying around, I'd be tempted to recap it or use it to do some voltage ripple measurements.

Medion used MSI boards, and MSI was very badly hit with the capacitor plague. Buying a second motherboard of the same type won't affect that as the second one also potentially has bad caps and is 20 years old

If you're not sure yourself take a good high-res pic of the board and show it here. Flat caps are no guarantee of good operation, but bulging leaky ones guarantee bad stuff.

Same applies to the PSU, check for bad caps.

20 year old PSU.... yeah, it should be recapped at the very least.

Best is to replace it with a new ATX PSU that is at least 80+ Bronze rated from a respectable brand such as Seasonic or similar. You'll have infinitely cleaner power being supplied to everything and the PSU will run way cooler and quieter to boot. Plus you will be using way less electricity to power the same equipment.

The cleaner power also means that the motherboard voltage regulation and power filtering will run cooler and be under less stress as well.

cyclone3d wrote on 2021-08-06, 15:36:

20 year old PSU.... yeah, it should be recapped at the very least.

Best is to replace it with a new ATX PSU that is at least 80+ Bronze rated from a respectable brand such as Seasonic or similar. You'll have infinitely cleaner power being supplied to everything and the PSU will run way cooler and quieter to boot. Plus you will be using way less electricity to power the same equipment.

The cleaner power also means that the motherboard voltage regulation and power filtering will run cooler and be under less stress as well.

I agree, but that's a big spend for a motherboard (now a pair of motherboards) with IDE that I'm still trying to find a good testing utility for.

Well,
You are getting data corruption with multiple different drives so that leaves:
Motherboard
IDE cable
RAM - but you already tried different RAM
CPU
power supply

I will mention that I recently tested a bunch of old power supplies with my power supply tester and out of over 10 power supplies, 7 of them were so bad that they had out of range voltages. A couple of them were even missing or flaking on/off of certain voltages altogether. These power supplies were all pulled from "working" computers.

I really, really, really distrust almost all power supplies that are over 5-7 years old, especially if they are not high quality brands. I really will only use older power supplies if they test good, have good caps, and are the only thing that will physically fit in the case I am using.

What is the brand and model of the power supply you are using?

Define "out of range"

https://imgur.com/a/iX15XN8

This is the power supply it's using. Nothing fancy.

I've had a bad psu cause some problems before but it was related to general stability like rando reboots. data corruption in the way that is being described here isn't one that I've experienced do to a psu. The data corruption being described here sounds like a bad hard disk controller, or a bad hdd cable. It's happened to me before. Only suggestion I have for you is try get a PCI IDE or PCI sata card with proper 9x support and disable your onboard controller in n your bios and see if you still have same issues. If you don't than its your controller and thats how you test it properly.

Thanks for that idea Warlord. I hadn't considered that it would be possible to have an IDE controller card to compare it against.