Keep in mind that MHDD can not test more than one disk at a time.

A (maybe faster) way to do it would be:
- Make one or more custom PCs maxxed with hard disks (i.e.: connect all ports, 4 HDD per motherboard).
- Use a bootable linux with gsmartcontrol.
- Open gsmartcontrol and reject any disk that shows as failed.
- Open every disk and reject any disk that shows the "Attributes" tab in red (that means that some attributes show that the disk is going to fail).
- Launch the Short test. You can test more than one disk at the same time, that will save you much time. Reject any disk that fails the short test.
- Any disk that passed the short test is worthy of the heavier testing (i.e.: MHDD or the extended test). Remember that you can do SMART tests on multiple disks at a time, so testing 4 disks on a motherboard takes about the same time as testing the bigger one alone.

(About the "Attributes" tab... sometimes shows in red when the disk exceeded some temperature, maybe you can ignore that. But if it shows reallocated sectors, hardware errors or something like that in red, that disk is no good)

I would test in two passes.

First step : Read SMART infos (smartctl on Linux for instance. I don't know for Windows).
It just reads the status in the disk controller. It's instantaneous

• Bad health status -> Trash
• OK -> next step

This will eliminate most disks with problems.

Second step : Actual deep check (read/write etc...). I don't know tools for that (i never had to use them). Can take a while depending on the disk size because it reads/write the disk completely.

To avoid constant reboots :

• SATA : disks can be hotplugged. Just leave the case open and add and remove disks without rebooting.
• IDE : use a USB adapter so you can hotplug them

How fast? Try HDD Regenerator bootable ISO. It checks the entire disk for bad sectors and also fixes them if possible.

Link
https://archive.org/details/hdd-regenerator-2011-full

I forgot to mention these are all old IDE hard drives ranging from 40MB up to 80GB. When was SMART introduced?

dominusprog wrote on 2023-08-17, 17:36:

How fast? Try HDD Regenerator bootable ISO. It checks the entire disk for bad sectors and also fixes them if possible.

A few days max.. I don't reallyhave much free time lately.

Half-Saint wrote on 2023-08-18, 06:35:

I forgot to mention these are all old IDE hard drives ranging from 40MB up to 80GB. When was SMART introduced?

My bet was that this would be mostly SATA and a few late IDE.
My suggestion won't work with all quite old IDE drives. SMART will be available for the 80GB but not for the 40MB. And IDE<->USB sometimes don't report SMART properly.

But with such small disks, time will not be a problem to read/write them completely.
You can try several disks at once as @Zup suggested. And the following workflow:

- Step 1 : plug 1 disk and reboot
- Detected by the bios -> Next step
- Not detected -> trash

- Step 2 : plug 4 drives that were properly detected at step 1 and use a testing tool from a USB booted linux

First, before going deep with MHDD, set a new pc to readonly, and externally test them quick with Hdd Tune pro. If they go though on quick all green ,fast, set those drives aside ,they should be ok.
The ones not recognized, but spinning, use Different pc controllers, until they show up in bios/ mhdd/vitoria/hdat2. Trow away the ones that click with a different pcb(no/spinup-clicks even after hit or nospin with similar pcb).Manual spinup is available in mentioned programs.

L.e. old drives are iffy, try re-pluging them 30 times for proper spinup sequence. Don't trow away rare drives. Pre 20gb ones.

acl wrote on 2023-08-18, 06:55:

Half-Saint wrote on 2023-08-18, 06:35:

I forgot to mention these are all old IDE hard drives ranging from 40MB up to 80GB. When was SMART introduced?

My bet was that this would be mostly SATA and a few late IDE.
My suggestion won't work with all quite old IDE drives. SMART will be available for the 80GB but not for the 40MB. And IDE<->USB sometimes don't report SMART properly.

At least half of the time IDE was mainstream technology, actually - I have it on a mid-1996 drive 😀
Agreed the 40MB is most likely way too old for that…

As for USB converters, most reliable bet is a modern one with UAS in addition to MSC, which I'm actually not sure if it's meaningfully applicable to an IDE disk - failing that it's really up to the converter chip and the client software, as indeed there wasn't a real standard back then

acl wrote on 2023-08-18, 06:55:

... […]
Show full quote

Half-Saint wrote on 2023-08-18, 06:35:

I forgot to mention these are all old IDE hard drives ranging from 40MB up to 80GB. When was SMART introduced?

...

- Step 1 : plug 1 disk and reboot
- Detected by the bios -> Next step
- Not detected -> trash
...

Don't forget these drives use jumpers for configuration. If it's not detected in BIOS and the drive is spinning. Check the jumpers first.

For drives smaller then 8GB I use P-133 machine. I connect one drive to secondary channel and do full format in DOS. It doesn't take too long. For larger drives I use one Socket A machine with Windows XP and do the same. Sometimes I use Victoria HDD to do full scan with sector speed statistics.

When I had a large pile (though not this large), I used a Linux system, where the tool badblocks is great for this. I was able to hook up four drives at once (both IDE channels; OS was on SATA) and with badblock -nsv /dev/sda1 (and sda2 and sdb1 and sdb2) I could get through the drives in decent time.

Do make sure the drives are of similar size & performance. If you have a small drive vs a large drive, the small one will be finished long before the large one.

If you must mhdd all the way. Can boot from floppy, should work on P120 no problem.

kixs wrote on 2023-08-18, 22:50:

Sometimes I use Victoria HDD to do full scan with sector speed statistics.

Im a big fan of Victoria, cant be bothered to boot dos every time 😀.
I always to this with mechanical drives I actually want to use. Sector access times are really useful for spotting
- about to fail/unhealthy drives.
- remapped drives.
A ton of 2-10TB "refurbished" drives nowadays are being sold with erased SMART stats as "new", "old stock" or "almost unused", all a lie and blatant scam. That procedure not only hides running hours/max temp reached, but also moves G-List (growing list of weak/defective sectors) into P-list https://datarecovery.com/rd/what-are-p-lists-and-g-lists/ https://forum.hddguru.com/viewtopic.php?f=1&t=41352 hiding real drive surface health state.

Then I look at max temp reached during subsequent write cycle. Rotation sound is also important, noisy drive means tired/unhealthy bearing.

Personally I dont keep old HDDs. SSDs/flash cards all the way, there are converters for almost everything nowadays. I would make exception for MFM/RLL drives because I find the way they work cool, like big floppies 😀 I plan to try an implement my own st506 controller some day and experiment with different bitrates/encodings.

A friend of mine used a working scrap Core2 HP workstation that had 2 IDE channels and 6 SATA 2 ports onboard, along with a PCI SATA card to DBAN all the old hard drives getting scrapped at his work years ago, so he could do 4 IDE's at a time and 8 SATA drives. Had it running in a corner of their storage room for a couple days and he'd just check on it and swap drives/print out certificates as he went.
I ended up with 12 40gb thin 5400rpm Seagate drives out of that, + a mess of others.

Don't go quickly.

Take a time slot to test every five hard drives once a week or so till you tested these drives. You need to do a surface scan, the one that will log any reseeks is best, on a hard drive that does not support SMART. SMART feature built into hard drive's firmware logs any SMART attributes , was made standard in 1995, Try gsmartcontrol, available in 32bit and 64bit you can download portable version. Try this on hard drives that is made after 1998 or so.

Using SMART utility speeds up the checking the hard drive health. If you can query the SMART on a hard drive which takes seconds and look at attributes for any issues then, you don't need to perform a long, slow scan.

Cheers,