Although some drives are louder than others, it's dependent on the code in the BIOS what sound it makes. That one is Award BIOS. There may be a "boot up floppy seek" option you can turn back on in the Win98 machine BIOS, or, the floppy seek may have already been removed from it for compliance with Microsoft PC97 and so forth.

Stay away from LS120, and away from 2.88mb drives then.

Stick with NEC, Matsushita, or Fujitsu drives, though the last tend to be quieter, as I recall. (I AM getting old though).

In the latter 90s/early 2000s, as floppies were getting more rare, they were seen as a sign of the syatem being obsolete. (Imagine! Pht!) As such, it became more vogue for the few remaining makers of them to make them quieter, and less obtrusive.

Ordinary NEC and Matsushita drives are the clanky loud things we grew up with, And probably what you want. You want drives with mfg dates before about 1999, but dont quote me on that. It's a soft 'feels right' cutoff, not specific insider knowledge.

This is in addition to the 'track seek test' options mentioned above. I prefer to have the full seek, since it can help you spot a misjumpered old 5.25 drive. (They wont respond to the motor control line, but will spin, if misjumpered.)

The issue I cite is more 'they seek, but the seek is very quiet'

jakethompson1 wrote on 2025-10-14, 15:33:

Although some drives are louder than others, it's dependent on the code in the BIOS what sound it makes. That one is Award BIOS. There may be a "boot up floppy seek" option you can turn back on in the Win98 machine BIOS, or, the floppy seek may have already been removed from it for compliance with Microsoft PC97 and so forth.

Thanks. I'll take a look if I have that option.

wierd_w wrote on 2025-10-14, 15:39:

Stay away from LS120, and away from 2.88mb drives then. […]
Show full quote

Stay away from LS120, and away from 2.88mb drives then.

Stick with NEC, Matsushita, or Fujitsu drives, though the last tend to be quieter, as I recall. (I AM getting old though).

In the latter 90s/early 2000s, as floppies were getting more rare, they were seen as a sign of the syatem being obsolete. (Imagine! Pht!) As such, it became more vogue for the few remaining makers of them to make them quieter, and less obtrusive.

Ordinary NEC and Matsushita drives are the clanky loud things we grew up with, And probably what you want. You want drives with mfg dates before about 1999, but dont quote me on that. It's a soft 'feels right' cutoff, not specific insider knowledge.

This is in addition to the 'track seek test' options mentioned above. I prefer to have the full seek, since it can help you spot a misjumpered old 5.25 drive. (They wont respond to the motor control line, but will spin, if misjumpered.)

The issue I cite is more 'they seek, but the seek is very quiet'

Thanks for the advice. Interesting. Funnily enough, the drive in my test system is a Matsushita, I believe. No, scratch that, it's a Mitsumi?

Mitsumi was loud if it was from about 96 or older. Normally just fine in my opinion.

jakethompson1 wrote on 2025-10-14, 15:33:

Although some drives are louder than others, it's dependent on the code in the BIOS what sound it makes. That one is Award BIOS. There may be a "boot up floppy seek" option you can turn back on in the Win98 machine BIOS, or, the floppy seek may have already been removed from it for compliance with Microsoft PC97 and so forth.

The algorithm is actually an IBM algorithm, and it has been cloned by both AMI and Award. It turned out IBM patented this algorithm (which is included in the AT BIOS) and sued the cloners over it years later. Thus the kind of sound is not supposed to depend on the BIOS vendor. Also, as jakethompson1 correctly points out, it's not the drive that generates the seek motion that produces this sound, but the BIOS. Floppy drives are "stupid" and don't do things on their own (like making this noise). Furthermore, also the floppy controllers used PCs do not have such complex motion patterns integrated in their programming, so only software (the BIOS in this case) is able to produce the head motion that causes this sound.

Originally, IBM invented this algorithm to tell 1.2MB drives and 360KB drives apart. As 1.2MB drives rotate at 360rpm, but 360KB drives rotate at 300rpm, the BIOS needs to configure the floppy controller differently depending on the drive type. While usually the BIOS can read the drive type from the CMOS setup, there is an issue if the setup information is invalid or the user swapped the drive type: As the IBM AT does not have a setup tool in its ROM, it depends on being able to boot from floppy even if the CMOS contents are wrong or not present at all to boot the SETUP utility. Thus it was essential that the AT is able to identify the drive type automatically.

Indeed. The XT has/had a different seek test, and sounds different.

https://youtube.com/shorts/SQP7ycndat8?si=o28PgCWIu0hfvIyh

I still assert that the full seek test is useful for detecting bad cables, musjumpered drives, and 'sick' spindle motors/belts.

AMI's seek is distinctively "harsher" and faster than Award's, and Phoenix (4.0x) is notably more relaxed and slower.
Maybe it isn't the seek algorithm that is different, but the floppy timing tables?

jakethompson1 wrote on 2025-10-14, 15:33:

Although some drives are louder than others, it's dependent on the code in the BIOS what sound it makes. That one is Award BIOS. There may be a "boot up floppy seek" option you can turn back on in the Win98 machine BIOS, or, the floppy seek may have already been removed from it for compliance with Microsoft PC97 and so forth.

Yes, there's a boot up floppy seek in my BIOS. It was disabled, but I've enabled it now. Can't test it though until I get my replacement floppy drive. ^^;

The older you go, the louder they get (to a point). It was already mentioned that different BIOS also contribute to the sound. Teac drives were very common and I consider them the "loudest". They made them up until 2006, but those are very flimsily built, quiet and fast*.

I made a few quick videos seeking on a few drives. Older Teacs sound better.
TEAC and ALPS
MITSUMI and Panasonic
SAMSUNG and SONY
I also started my NEC, so you can hear that one. It's quite unique and I love it.
NEC seeking

Major Jackyl wrote on 2025-10-14, 23:38:

The older you go, the louder they get (to a point). It was already mentioned that different BIOS also contribute to the sound. T […]
Show full quote

The older you go, the louder they get (to a point). It was already mentioned that different BIOS also contribute to the sound. Teac drives were very common and I consider them the "loudest". They made them up until 2006, but those are very flimsily built, quiet and fast*.

I made a few quick videos seeking on a few drives. Older Teacs sound better.
TEAC and ALPS
MITSUMI and Panasonic
SAMSUNG and SONY
I also started my NEC, so you can hear that one. It's quite unique and I love it.
NEC seeking

Thanks for taking the time to record those. 😁

The one I bought off eBay is the Techmedia TFD-310 as it was only £13. It was rescued from an old PC and left in storage. No idea what those sound like. Worst case, I'll try tracking down a Mitsumi.

Major Jackyl wrote on 2025-10-14, 23:38:

The older you go, the louder they get ...

I've still got a Sugart SA-400 as far as I know the original/first 5.25" drive - full height, single-sided, 35-track.
I've also got a SA-400L - the "Long" version - 40 tracks!
and an SA-450 - double sided.

The exact timing of the PC boot-up seek is determined by BIOS, the raspyness/volume by the drive itself.

I've mostly used Panasonic drives in my systems over the years - didn't pay them a lot of attention (till I got to working on ImageDisk)
- they all sounder pretty "normal" to me.

Major Jackyl wrote on 2025-10-14, 23:38:

I also started my NEC, so you can hear that one. It's quite unique and I love it.
NEC seeking

Yeah, that is Phoenix, and while we're at it here is someone's AMI: https://www.youtube.com/watch?v=8vut7LWaOmU

jakethompson1 wrote on 2025-10-14, 17:44:

AMI's seek is distinctively "harsher" and faster than Award's, and Phoenix (4.0x) is notably more relaxed and slower.
Maybe it isn't the seek algorithm that is different, but the floppy timing tables?

The algorithm has one primary purpose: Move past track 44 (or so) to make 40-track drives hit the inner bumper and misstep. This is "intentional headbanging" on those drives. So the implementation is likely very similar as well (although not every BIOS might use the exact value of "track 50" as target). Using a different step rate is indeed a simple explanation for a more harsh or a more relaxed "floppy seek". Also, the delay between the three phases (recalibrate, step 50 tracks inwards, step 40 tracks outwards, single step outwards until the track 0 sensor activates) may be different. Also, in the last phase, the step timing may depend on processor steep, while in the first two phases, the step timing is generated by the FDC, most likely based on the value in the floppy parameter table.

I never notices the differnce between vendors, so thanks for pointing it out. During my PC career, I first had a classic AMI 286, then a Hi-Flex AMI 486, then an Award 4.51 Pentium. All of those did the floppy seek stuff, and I don't remember being surprised about the way the Pentium system sounded when it was new.

I got my replacement floppy drive today. And yeah, bootup floppy seek does the trick. 😁 For this particular floppy drive, it's quite soft sounding.