I'm sorry, I can't answer your question - I mainly used Win95 in the pre-GB times.. 🙁
However, you can boot from the Win98SE CD-ROM and run a newer (?) ScanDisk from there.
Alternatively, Windows ME might be helpful in that special case. It always uses the Windows version of ScanDisk.

No, that's not normal. As long as ScanDisk does not find a bad sector, speed should be the same for every sectors it tests. A 4 GB drive should be scanned in 20 to 30 minutes max.

If it is getting slow even with no bad sectors found, I would check the IDE cable first. If it's loose, it may result in bus errors, slowing things down due to communication retries. If it's not the cable, replace the HDD.

Alright, just swapped out IDE cables, did fdisk, and doing a surface scan using a Win98SE CD ScanDisk tool. As it’s going along, it reads clusters at a consistent rate for the first few percentage, but once it hits around 6%, it bogs down and hangs for several seconds at a time. Took 30 mins to get to 9%, but still no bad clusters found. IDE controller is on the motherboard.

Took a Maxtor 6.4GB out of my Compaq and tried it with the original IDE cable in the Packard Bell. Sure enough, zero issues getting through a surface scan nice and quick. Replacing the old HDD then!

I think scandisk does retry each sector more than once if it finds something questionable. if it fails more than "x" times, it gets marked bad. So a disk may be weak but still "pass" the check. I usually see this on floppies, but I suppose a HDD is equally suseptible.

Cobra42898 wrote on 2020-01-13, 17:05:

I think scandisk does retry each sector more than once if it finds something questionable. if it fails more than "x" times, it gets marked bad. So a disk may be weak but still "pass" the check. I usually see this on floppies, but I suppose a HDD is equally suseptible.

I tink the same. If only that HDD wasn't so big.. 🙁
Sophisticated software for checking low-Level / physical stuff was made in DOS 3.x/DOS 4.x times (I'm thinking of Spinrite and CheckIt!),
so the newest that such special software was mader aware of are things like FAT16B paritions and drives up to 2GB.

Though in reality, that's a far stretch, even, perhaps. In its prime time, 2o to 120MB MFM/RLL and ESDI/IDE fixed disk were the more usual stuff.
Also, starting with ESDI/ATA, HDDs started to mask physical defects and physical drive geometry, too.
That's why DOS (chkdsk /f, scandisk, format etc) rarely finds sectors marked as "bad" these days, I guess.

Even though CHS was still the norm on the software level (LBA awareness started ca. with late 486 PCs),
the hardisk themselves started to get advanced capabilities, such as sector-caches, DMA/Multi-Word DMA and fake CHS.
(Originally, I suppose, non-PC platforms made more often use of the new features.)

Anyway, that being said, there must be a way to check the drive or truelly mark bad sectors.
Maybe though the use of some service software from the HDD manufacturer, not sure. 😀

I was lucky enough to replace the HDD with the same brand and size HDD. Passed ScanDisk in only 15 mins, so that old drive definitely has an issue. However, I’m gonna hang onto it for now in case someone has an idea as to what may be up or a more advanced way to check it, just for fun.

This is why one should use only enterprise-grade drives.
Consumer-grade drives retry and retry indefinitely in case of errors.
Enterprise-grade drives retry only very few times before reporting error.