VOGONS


Storage options for Windows 98SE?

Topic actions

Reply 20 of 46, by gdjacobs

User metadata
Rank l33t++
Rank
l33t++

Seatools can be used to set a maximum drive geometry which works with your bios limits. WD drives can be manipulated by creating an hpa area -- not as slick, but will work in many circumstances. Couple them with an ATA to SATA bridge and away you go.

All hail the Great Capacitor Brand Finder

Reply 21 of 46, by soggi

User metadata
Rank Member
Rank
Member
momaka wrote on 2024-09-24, 21:38:

The last thing I don't like about SCSI HDDs is that they tend to be very loud. And at least in my area, the choice of drives always seems to be either very old 7200 RPM crap with ball bearings that sounds like a tank running on your driveway, or newer 10k RPM HDDs that sound like a jet engine / airplane at take off. Not that old IDE HDDs are that much more quiet... but overall I find it easier to find models that use FBD and aren't as loud as SCSI.

Then you definitely got the wrong drives and had very bad luck! I own some 10k and even 15k UW160/320 SCSI/SCA drives (68/80 pin) which you hardly could hear in a room w/o any noise. These are quieter than the very most IDE drives I own, I just recall a 120 GB Maxtor DiamondMax Plus 9 drive which has approximately the same (very low) noise level.

I also know those types which sound like a circular saw...f.e. a Micropolis 4 GB one I own, the loudest HDD I ever heard.

BTW I also think that some UW160/320 SCSI is much more better than any Flash/SSD based storage, f.e. because you don't have to bother with TRIM, wear level, write cycles and such stuff.

kind regards
soggi

Vintage BIOSes, firmware, drivers, tools, manuals and (3dfx) game patches -> soggi's BIOS & Firmware Page

soggi.org on Twitter - inactive at the moment

Reply 22 of 46, by stamasd

User metadata
Rank l33t
Rank
l33t
momaka wrote on 2024-09-24, 21:38:

The last thing I don't like about SCSI HDDs is that they tend to be very loud. And at least in my area, the choice of drives always seems to be either very old 7200 RPM crap with ball bearings that sounds like a tank running on your driveway, or newer 10k RPM HDDs that sound like a jet engine / airplane at take off. Not that old IDE HDDs are that much more quiet... but overall I find it easier to find models that use FBD and aren't as loud as SCSI.

The IBM 73GB SCA80 drives I have are some of the quietest drives I know.
But if sound is a concern, you'll be happy to know that with modern devices like BlueSCSI you'll be spared any sound at all. 😜

I/O, I/O,
It's off to disk I go,
With a bit and a byte
And a read and a write,
I/O, I/O

Reply 23 of 46, by Mondodimotori

User metadata
Rank Member
Rank
Member
momaka wrote on 2024-09-24, 21:38:

Ouch, that's a bit pricey for my tastes. Bu I get it, at least it's tested and has warranty.

Here, there's a local flea market every weekend and lately I've been getting my IDE drives from there. Untested, of course, and have gotten quite a few non-working ones. But then I often get them at $1-2 per drive, so it's worth the gamble. So far my luck has been around 50-ish %. Alternatively, I can often get a whole old desktop for about $7, and these regularly come with the HDDs inside. Right now, the average one is usually a Core2Duo system with a SATA HDD of some sort... but also not uncommon to see Athlon XP PCs with IDE HDDs (in fact, it's the 2nd most common system type here.)

Of course, of course. It's a long way from China, but it's a professional seller with decently high feedbacks and return policy + ebay warranty. Of course I know that, if I dig a little deeper and with less expectations, I can easily find used HDD for cheap. But the fact is that I have bad experiences with older HDDs. They get unbearably slow after 7-8 years of constant use, even when they are 7200 rpm. So I decided to just shill those money for one or two NOS that are gonna last me more than a decade in these older sistems. It's not like I'm gonna daily run them.
It all depends on your use case.

Reply 24 of 46, by momaka

User metadata
Rank Oldbie
Rank
Oldbie
soggi wrote on 2024-09-25, 03:23:

Then you definitely got the wrong drives and had very bad luck! I own some 10k and even 15k UW160/320 SCSI/SCA drives (68/80 pin) which you hardly could hear in a room w/o any noise.

Maxtor Atlas U320 72 GB and Seagate Cheetah 10k 300 GB (ST3300007LC).
You're right, they are not terribly loud drives. In fact, the Seagate Cheetah is a relatively modern FDB drive and thus doesn't make that much noise.
But I suppose what bothers me about 10k and 15k drives is the high-pitched airplane turbine whine/noise.

soggi wrote on 2024-09-25, 03:23:

I also know those types which sound like a circular saw...

Haha, me too.
DEC RZ40-AA (Quantum Atlas II) and DEC RZ40-AS (Seagate Medalist or similarly-aged drive, I think).
I can't remember which one was louder. I think it was one of the Quantum drives above (I have multiple of these.) You can hear it in the other room with doors closed everywhere.

Mondodimotori wrote on 2024-09-25, 13:28:

But the fact is that I have bad experiences with older HDDs. They get unbearably slow after 7-8 years of constant use, even when they are 7200 rpm.

??
All of mine are still fine.
I suppose it really depends on the particular drive brand and model, though. I know a few specific ones that do develop bad sectors and get very slow over time. My Seagate Barracuda ATA IV and 7200.7 drives have all been rock solid so far, though. The one from my parents' desktop that they bought back in 2005 is just a year shy of 20 years old.

Reply 25 of 46, by soggi

User metadata
Rank Member
Rank
Member
momaka wrote on 2024-09-29, 23:46:
?? All of mine are still fine. I suppose it really depends on the particular drive brand and model, though. I know a few specifi […]
Show full quote
Mondodimotori wrote on 2024-09-25, 13:28:

But the fact is that I have bad experiences with older HDDs. They get unbearably slow after 7-8 years of constant use, even when they are 7200 rpm.

??
All of mine are still fine.
I suppose it really depends on the particular drive brand and model, though. I know a few specific ones that do develop bad sectors and get very slow over time. My Seagate Barracuda ATA IV and 7200.7 drives have all been rock solid so far, though. The one from my parents' desktop that they bought back in 2005 is just a year shy of 20 years old.

I also had these question marks... How could HDDs get slow itself? That's quite impossible. Sure, there can be defects or the operating system gets bloated over time and therefor the system will be slower...but the HDD itself? I've never heard that before (about HDDs, not SSDs). If the system got bloated then it could be time make a fresh install!?

kind regards
soggi

Vintage BIOSes, firmware, drivers, tools, manuals and (3dfx) game patches -> soggi's BIOS & Firmware Page

soggi.org on Twitter - inactive at the moment

Reply 26 of 46, by momaka

User metadata
Rank Oldbie
Rank
Oldbie
soggi wrote on 2024-09-30, 01:28:

If the system got bloated then it could be time make a fresh install!?

Yup, I too suspect this is what people mean when they say their HDDs (not SSDs) start going slow - it's the OS install getting bloated.
Only time an HDD can go slow is if the read heads become weak or there are too many reallocated sectors all over the drive, causing the HDD to be "trashing". But these two cases are rather rare... or at least when they happen, typically the drive fails to work at all or times out the second any attempt at reading from its surface is made.

Now, SSDs on the other hand, I have seen get slower and slower over time. The 120 GB Inland SSD I got from MicroCenter and put into the laptop of my former workplace did exactly that in the course of a year. It was a Windows 7 install with update disabled and all software "set in stone", so to speak (i.e. I didn't change much past the initial configuration.) At first, the laptop would boot in less than 15 seconds. Over the course of a year, this went down to maybe almost a minute... which is still not too terrible, but certainly noticeable. Indeed it's true that SSDs' memory cells may get weak over time and become hard to read. So no surprise there.

Reply 27 of 46, by soggi

User metadata
Rank Member
Rank
Member

It's about TRIM, garbage collection and such things on SSDs and they have limited write cycles - I still don't trust SSDs on storing important data, I don't trust them completely at all. Additionally I'm mostly using WinXP and it doesn't have good support for SSDs, not to speak of earlier Windows versions.

kind regards
soggi

Vintage BIOSes, firmware, drivers, tools, manuals and (3dfx) game patches -> soggi's BIOS & Firmware Page

soggi.org on Twitter - inactive at the moment

Reply 28 of 46, by y2k se

User metadata
Rank Member
Rank
Member

You shouldn't trust any single storage device to store important data, regardless of the medium.

Tualatin Celeron 1.4 + Powerleap PL-IP3/T, ASUS P2B, 256 MB RAM, GeForce 4 Ti 4200, Voodoo2 SLI, AWE64, 32GB IDE SSD, Dell 2001FP
P233MMX, Intel LT430TX, 64 MB RAM, Sierra Screamin' 3D, AWE64 Gold, 32GB IDE SSD, Viewsonic A75f

Reply 29 of 46, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++

Though on a spindle drive you typically only get 2/3 to a half the data transfer speed on inner cylinders than outer cylinders so it does get subjectively slower as it gets more full... also because there's more tracks with data on to seek between.

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 30 of 46, by soggi

User metadata
Rank Member
Rank
Member
y2k se wrote on 2024-09-30, 02:22:

You shouldn't trust any single storage device to store important data, regardless of the medium.

That's right but another story...

I was talking about inadequacies of SSDs in general and specifically on WinXP and earlier. For example HDDs have unlimited write cycles in theory, SSDs not (as said above) and you don't have to bother with TRIM and garbage collection. So SSDs are great as boot disk with no important data (be careful where your programs store their data) on systems with full support for them, there they have their (big) advantages and can be replaced, if necessary.

BitWrangler wrote on 2024-09-30, 02:37:

Though on a spindle drive you typically only get 2/3 to a half the data transfer speed on inner cylinders than outer cylinders so it does get subjectively slower as it gets more full... also because there's more tracks with data on to seek between.

You can measure it, but you won't have the feeling in real environment and say "Hey, this is quite slow now, this only can be the inner part of the platter!". And the latter (seeking for data) is OS stuff.

kind regards
soggi

Last edited by soggi on 2024-09-30, 03:07. Edited 2 times in total.

Vintage BIOSes, firmware, drivers, tools, manuals and (3dfx) game patches -> soggi's BIOS & Firmware Page

soggi.org on Twitter - inactive at the moment

Reply 31 of 46, by Joseph_Joestar

User metadata
Rank l33t++
Rank
l33t++
soggi wrote on 2024-09-30, 02:08:

It's about TRIM, garbage collection and such things on SSDs and they have limited write cycles - I still don't trust SSDs on storing important data, I don't trust them completely at all. Additionally I'm mostly using WinXP and it doesn't have good support for SSDs, not to speak of earlier Windows versions.

Sure, for important data and everyday work, using an SSD on an OS without TRIM would be bad. But for a retro gaming rig? Where nothing of importance is stored, except maybe some savegames? I don't see an issue there.

Personally, I've been using smaller (120 GB) SSDs on all of my Win9x systems for a couple of years now and haven't had any problems. And if you're worried about garbage collection, just leave the system idling at the BIOS overnight once per month, and it will do that on its own.

PC#1: Pentium MMX 166 / Soyo SY-5BT / S3 Trio64V+ / Voodoo1 / YMF719 / AWE64 Gold / SC-155
PC#2: AthlonXP 2100+ / ECS K7VTA3 / Voodoo3 / Audigy2 / Vortex2
PC#3: Core 2 Duo E8600 / Foxconn P35AX-S / X800 / Audigy2 ZS
PC#4: i5-3570K / MSI Z77A-G43 / GTX 970 / X-Fi

Reply 32 of 46, by soggi

User metadata
Rank Member
Rank
Member

I don't say you can't use SSDs on such systems, I just say I wouldn't and I also wouldn't recommend to do so. If someone is fine with the risks the use of a SSD has and he/she wants to use it despite of that, it's OK.

kind regards
soggi

Vintage BIOSes, firmware, drivers, tools, manuals and (3dfx) game patches -> soggi's BIOS & Firmware Page

soggi.org on Twitter - inactive at the moment

Reply 33 of 46, by Mondodimotori

User metadata
Rank Member
Rank
Member
momaka wrote on 2024-09-29, 23:46:

??
All of mine are still fine.
I suppose it really depends on the particular drive brand and model, though. I know a few specific ones that do develop bad sectors and get very slow over time. My Seagate Barracuda ATA IV and 7200.7 drives have all been rock solid so far, though. The one from my parents' desktop that they bought back in 2005 is just a year shy of 20 years old.

One was a Seagate 750GB 5200rpm, another one a 500Gb from a brand I can't remember. They were both 2.5' drives in my last two laptops and, even with regular formatting of the OS, they used to go slow after several years of use. And for several years I mean 5+ of constant daily use. Even a clean windows installation would go slow after a couple of days of use.
Only by installing new drives (either SSD or new HDD) would the systems go back to their usual speed. I did replace that 750GB Seagate drive with an SSD, after several full formats that didn't bring any speed back, and my laptop (Asus N56VZ with i7-3630qm, bought 11 years ago) it runs like it's brand new again.

What I'm getting at is: I wouldn't trust any mechanical HDD with 7+ years of daily use even if it wasn't slow: They are mechanical, and it's inevitable that the mechanical part will fail. Daily running an HDD of 10+ year is akin to an extreme sport IMHO. Especially if you do that for several hours a day and have important data stored on it. Those moving parts will break down sooner or later.

Probably by not daily running them you can get away with a longer useful lifespan. But since I cannot reliably verify that before purchasing an used drive (you either ask the seller to provide a cristal disk reading or trust them blindly), I prefer to pay a little more and get NOS.

Reply 34 of 46, by soggi

User metadata
Rank Member
Rank
Member

This still sounds weird to me, can't remember that I saw such a behavior by HDDs over the decades - I saw it, but with SSDs.

Nothing lasts forever, everything will break sooner or later, no matter if mechanical or electrical, even the latter will have defects of the material after (excessive) use and as said HDDs have unlimited write cycles in theory, especially consumer SSDs have been far from that.

HDDs could last very long, if you have an eye on the temperature, it should be relatively stable on not to high ranges.

I have no problems with HDDs running more than ten years nearly every day, the HDD (TOSHIBA MK8034GSX) in this IBM/Lenovo T60 works for ~18 years now, one external has even more hours (see screens below) - I guess a SSD would have already died for several reasons.

kind regards
soggi

Last edited by soggi on 2024-10-01, 03:01. Edited 1 time in total.

Vintage BIOSes, firmware, drivers, tools, manuals and (3dfx) game patches -> soggi's BIOS & Firmware Page

soggi.org on Twitter - inactive at the moment

Reply 35 of 46, by y2k se

User metadata
Rank Member
Rank
Member

The Samsung 970 Evo in my main PC is 4.25 years old and shows 39.2 TBW. Samsung rates its endurance has 600 TBW. So, at that rate, it would be good for 65 years. It has no moving parts, whereas the spinning drives do, so not only can the actual magnetic recording surface wear out, so can the read/write heads come out of alignment.

Based on Backblaze data, SSD failure rates are lower than HDD failure rates, around 1% AFR for SSDs versus 1.5% AFR for HDDs.
https://www.backblaze.com/blog/ssd-edition-20 … e-stats-review/
https://www.backblaze.com/blog/backblaze-driv … ts-for-q2-2024/

The other difference is the failure mode. HDDs failure progression tends to take a little time from first error to unusable drive, whereas SSD tends to quickly go from first error to unusual drive (though firmwares have gotten better in terms of just locking the drive to a read only state instead of full bricking).

Tualatin Celeron 1.4 + Powerleap PL-IP3/T, ASUS P2B, 256 MB RAM, GeForce 4 Ti 4200, Voodoo2 SLI, AWE64, 32GB IDE SSD, Dell 2001FP
P233MMX, Intel LT430TX, 64 MB RAM, Sierra Screamin' 3D, AWE64 Gold, 32GB IDE SSD, Viewsonic A75f

Reply 36 of 46, by Mondodimotori

User metadata
Rank Member
Rank
Member
soggi wrote on 2024-10-01, 02:01:
This still sounds weird to me, can't remember that I saw such a behavior by HDDs over the decades - I saw it, but with SSDs. […]
Show full quote

This still sounds weird to me, can't remember that I saw such a behavior by HDDs over the decades - I saw it, but with SSDs.

Nothing lasts forever, everything will break sooner or later, no matter if mechanical or electrical, even the latter will have defects of the material after (excessive) use and as said HDDs have unlimited write cycles in theory, especially consumer SSDs have been far from that.

HDDs could last very long, if you have an eye on the temperature, it should be relatively stable on not to high ranges.

I have no problems with HDDs running more than ten years nearly every day, the HDD (TOSHIBA MK8034GSX) in this IBM/Lenovo T60 works for ~18 years now, one external has even more hours (see screens below) - I guess a SSD would have already died for several reasons.

kind regards
soggi

Well, I didn't get rid of that 750GB seagate. It's sitting on my shelf, I can always put it back in the sistem and show you the readings, including the power on time. But I confirm what I said: It was slow even after a clean windows install. So slow that it made the OS unstable and barely usable. Let alone surfing the net. And just by replacing it (I put a sata ssd becaue thei've gotten dirty cheap) the sistem got back to be fully functional, like when it was new.
Maybe it's a difference between 2.5 and 3.5 inch drives, maybe being in a laptop it ran hotter (but the Asus N56VZ had the HDD in a separated tray from the rest of the sistem), maybe 10+ years of bumps and being moved around for thousands of chilometers took a tool on the drive.

But I've also noticed that the HDD (a 5200 rpm maxtor with 40GB of storage) i have in my Socket A, manufactured in 2004, is still pretty snappy. But, after all, it sat unused between at least 2010 up until last summer, so it actually has a little over 6 years of actual use, and at least a couple of those were not daily.
I still have the idea to replace it with a bigger drive, but for now it's not a priority.

Reply 37 of 46, by douglar

User metadata
Rank l33t
Rank
l33t
y2k se wrote on 2024-10-01, 02:24:
The Samsung 970 Evo in my main PC is 4.25 years old and shows 39.2 TBW. Samsung rates its endurance has 600 TBW. So, at that r […]
Show full quote

The Samsung 970 Evo in my main PC is 4.25 years old and shows 39.2 TBW. Samsung rates its endurance has 600 TBW. So, at that rate, it would be good for 65 years. It has no moving parts, whereas the spinning drives do, so not only can the actual magnetic recording surface wear out, so can the read/write heads come out of alignment.

Based on Backblaze data, SSD failure rates are lower than HDD failure rates, around 1% AFR for SSDs versus 1.5% AFR for HDDs.
https://www.backblaze.com/blog/ssd-edition-20 … e-stats-review/
https://www.backblaze.com/blog/backblaze-driv … ts-for-q2-2024/

The other difference is the failure mode. HDDs failure progression tends to take a little time from first error to unusable drive, whereas SSD tends to quickly go from first error to unusual drive (though firmwares have gotten better in terms of just locking the drive to a read only state instead of full bricking).

Very good summary.

If you chose a spinning disk because you like the aesthetic, because it is more compatible with your hardware, or because you have one available? All three of those choices make perfect sense to me.

But the decision that some people make to pass on an SSD because its life span can be measured, in favor of a spinning disk that has a lifespan that can't be measured but will in all likelihood will fail more than 5x sooner? That risk-reward analysis doesn't make sense to me.

Reply 38 of 46, by momaka

User metadata
Rank Oldbie
Rank
Oldbie
Mondodimotori wrote on 2024-09-30, 13:31:

One was a Seagate 750GB 5200rpm, another one a 500Gb from a brand I can't remember. They were both 2.5' drives in my last two laptops and, even with regular formatting of the OS, they used to go slow after several years of use. And for several years I mean 5+ of constant daily use. Even a clean windows installation would go slow after a couple of days of use.

Well, to make a proper apples to apples comparison, and since you say you still have those HDDs, perhaps (if you have the time and interest to do so) you can hook them up to a working system and do a benchmark with HD Tune. Also check the SMART logs prior the benchmark and after it. This will often tell if there's any "funny business" going on with the HDD. To get more accurate results, make sure that either the HDD is formatted (blank) or that File Indexing is turned off in Windows for that particular HDD (I personally turn off File Indexing for ALL of my HDDs in Windows, because I find the indexing service always tries to index stuff in the worst possible time and interrupts my work/gaming/PC activities.)

Also, there's a good chance those 750 GB and 500 GB Seagate HDDs you have are from the 7200.10/11/12 line (but in 2.5" form), and these IME tend to be a lot more problematic than the 7200.9 and 7200.7 series.

Mondodimotori wrote on 2024-09-30, 13:31:

What I'm getting at is: I wouldn't trust any mechanical HDD with 7+ years of daily use even if it wasn't slow: They are mechanical, and it's inevitable that the mechanical part will fail.

Yes. Though I shall note that in my observance, it all depends on the particular brand and model in question.
IME, Western Digital tend to do better when you keep them powered On all the time or for extended periods of time and don't power cycle them much. With Seagate, I find the opposite to be true: their spindle FDB are not as good as WD's, it seems, so they are more likely to be affected by long running hours. On the other hand, it seems their heads a little tougher and/or better engineered than WD's, because Seagate HDDs can take on a lot of power cycles and head load/unload cycles.

soggi wrote on 2024-10-01, 02:01:

HDDs could last very long, if you have an eye on the temperature, it should be relatively stable on not to high ranges.

Temperature, power cycles, running hours, air pressure, and noise on the power line are all factors that can affect their life.
I've seen quite a few bad HDDs originating from systems with cheap crap PSUs, and in my experience with cheap PSUs, I've also had a few HDDs run hotter overall due to dirty power from the PSU.

soggi wrote on 2024-10-01, 02:01:

I have no problems with HDDs running more than ten years nearly every day, the HDD (TOSHIBA MK8034GSX) in this IBM/Lenovo T60 works for ~18 years now, one external has even more hours (see screens below)

I see those drives at 20k and 30k hours, respectively... which isn't a whole lot, but it's not that little either. Actually, it's a pretty average figure for a home desktop PC that's been ran for a few hours every day on average. And one can tell it was a home desktop PC because of the high number of power cycles. Work-related desktops rarely get power-cycled and typically have twice those hours with half the power cycles or less. Most of my used HDDs are around that range. Now, my personal HDDs that I bought new, have a lot less hours and even more power cycles, as I tend to standby/sleep the PC when I anticipate I'll be away from it for more than an hour... especially in the summer, when I don't want to add more heat in the computer room. At winter, I tend to be a lot more lenient and sometimes keep my system running all day so that it keeps the computer room a little warmer. 😀

Speaking of high power-on hours, my highest is from a 100GB Hitachi Deskstar 3.5" desktop IDE HDD - it has over 100k hours at this point. No bad sectors so far, though. And before anyone thinks to say that perhaps this HDD counts the power-on time in minutes rather than hours (yes, I have a few HDDs like that), I've checked and this HDD properly counts the power-on time in hours. That's over 11 years of constant (24/7) uptime. Doesn't surprise me, though. I believe I bought it back around 2013 or later and got it with close to 100k hours at that point. So I've put maybe 2000-3000 hours on it and it rolled over 100k.

douglar wrote on 2024-10-01, 15:14:

The other difference is the failure mode. HDDs failure progression tends to take a little time from first error to unusable drive, whereas SSD tends to quickly go from first error to unusual drive (though firmwares have gotten better in terms of just locking the drive to a read only state instead of full bricking).

That's one reason I still prefer HDDs over SSDs. If you pay a little attention to them from time to time, you may just get alerted when they are about to fail. Of course, that's not always the case either. Some HDDs will fail completely out of the blue. Use it and all be fine one day... turn it on the next day and it's *click-click-meeep----click-click--meep" ad-infinitum (I'm looking at you Western Digitial WD800JD series!)

Also, another reason why I like HDD is that they could have sat for 10+ years in storage, but the data off them will still be readable if the drive is healthy. With SSDs, we will have to see about that. I imagine the new SSDs that have really tiny memory retention cells may become a lot more likely to give data corruption if left unpowered for many years. I already have flash drives that have done this to me (though if I have to be honest, it's only been the really cheap no-name flash drives with capacitices 8 GB and over.) That's not to say flash memory can't be trustworthy, though. I have some 4 GB drives from 2009/2010 that still retain all of their original data error-free. Nothing can top out my 512 MB SanDisk Cruizer "mini" though - that one my mother gave me way back in 2005 when she got a bunch from her workplace, and this flash drive is still retaining data from back then without errors.

But all in all, I suppose we will see how reliable or unreliable flash memory technology is in SSDs as time goes by. Personally, the only reason why I don't trust modern SSD for long term reliability is because manufacturers are always trying to cram more and more memory cells per given space, thus making it more likely for data corruption to occur. And it seems these days that we're still accelerating more towards becoming a consumer society, no matter where we live around the world. If something breaks, just throw it away and get a new one for cheap... and do the same when that breaks too. With that kind of expectation, manufacturers often design things to last "just long enough". So we can forget about long-lasting stuff to ever be made again like they have in the past. The life cycle of every product is carefully calculated.

Reply 39 of 46, by soggi

User metadata
Rank Member
Rank
Member
Joseph_Joestar wrote on 2024-09-30, 03:06:

And if you're worried about garbage collection, just leave the system idling at the BIOS overnight once per month, and it will do that on its own.

How does it work? What's the difference being in a AMI/(Phoenix-)Award BIOS and being in DOS or Win9x in that case?

y2k se wrote on 2024-10-01, 02:24:
The Samsung 970 Evo in my main PC is 4.25 years old and shows 39.2 TBW. Samsung rates its endurance has 600 TBW. So, at that r […]
Show full quote

The Samsung 970 Evo in my main PC is 4.25 years old and shows 39.2 TBW. Samsung rates its endurance has 600 TBW. So, at that rate, it would be good for 65 years. It has no moving parts, whereas the spinning drives do, so not only can the actual magnetic recording surface wear out, so can the read/write heads come out of alignment.

Based on Backblaze data, SSD failure rates are lower than HDD failure rates, around 1% AFR for SSDs versus 1.5% AFR for HDDs.
https://www.backblaze.com/blog/ssd-edition-20 … e-stats-review/
https://www.backblaze.com/blog/backblaze-driv … ts-for-q2-2024/

The other difference is the failure mode. HDDs failure progression tends to take a little time from first error to unusable drive, whereas SSD tends to quickly go from first error to unusual drive (though firmwares have gotten better in terms of just locking the drive to a read only state instead of full bricking).

Unfortunately there's no TBW value for my HDDs on the screenshots, so I can't really classify this. On first sight it seems 39.2 TBW is a bit low for a main PC in 4.25 years...but it's just my thought when I summarize my DLs, copying data and stuff (not to speak of TMP files, logs and virtual memory).

Nice statistics (I really looked at them), but you forgot this thread is about Windows 98 SE storage options - it's called "Storage options for Windows 98SE?". I said this more than once (f.e. "I was talking about inadequacies of SSDs in general and specifically on WinXP and earlier. "). Yes, right, I additionally was talking about WinXP, but it shares with Win98SE that it basically also doesn't support SSDs and additionally both don't fully support the HDDs/SSDs from the stats you linked to - I know there are some (older SSDs) with manufacturer tools supporting XP that way. SSDs will age faster on such systems. And there's another problem - the more the SSD is filled, the more the rest of the cells are strained. I have very often the case that a partition or the complete HDD is filled ~90% or more and the rest gets a lot of writes.

Mondodimotori wrote on 2024-10-01, 11:42:
Well, I didn't get rid of that 750GB seagate. It's sitting on my shelf, I can always put it back in the sistem and show you the […]
Show full quote

Well, I didn't get rid of that 750GB seagate. It's sitting on my shelf, I can always put it back in the sistem and show you the readings, including the power on time. But I confirm what I said: It was slow even after a clean windows install. So slow that it made the OS unstable and barely usable. Let alone surfing the net. And just by replacing it (I put a sata ssd becaue thei've gotten dirty cheap) the sistem got back to be fully functional, like when it was new.
Maybe it's a difference between 2.5 and 3.5 inch drives, maybe being in a laptop it ran hotter (but the Asus N56VZ had the HDD in a separated tray from the rest of the sistem), maybe 10+ years of bumps and being moved around for thousands of chilometers took a tool on the drive.

But I've also noticed that the HDD (a 5200 rpm maxtor with 40GB of storage) i have in my Socket A, manufactured in 2004, is still pretty snappy. But, after all, it sat unused between at least 2010 up until last summer, so it actually has a little over 6 years of actual use, and at least a couple of those were not daily.
I still have the idea to replace it with a bigger drive, but for now it's not a priority.

I believe, you're telling your true experience...but it's still weird because I didn't have such a behavior with any HDD and I can't see what's the reason for that. BTW the HDDs (see the screenshots) are also 2.5" and one is placed in the T60 and the other is an external in a case...they also made many thousands of kilometers via car, flights and train.

douglar wrote on 2024-10-01, 15:14:

Very good summary.

If you chose a spinning disk because you like the aesthetic, because it is more compatible with your hardware, or because you have one available? All three of those choices make perfect sense to me.

But the decision that some people make to pass on an SSD because its life span can be measured, in favor of a spinning disk that has a lifespan that can't be measured but will in all likelihood will fail more than 5x sooner? That risk-reward analysis doesn't make sense to me.

As said before...we are talking about Win98SE...that's another risk as with modern OS.

momaka wrote on 2024-10-01, 17:03:

Well, to make a proper apples to apples comparison, and since you say you still have those HDDs, perhaps (if you have the time and interest to do so) you can hook them up to a working system and do a benchmark with HD Tune. Also check the SMART logs prior the benchmark and after it. This will often tell if there's any "funny business" going on with the HDD. To get more accurate results, make sure that either the HDD is formatted (blank) or that File Indexing is turned off in Windows for that particular HDD (I personally turn off File Indexing for ALL of my HDDs in Windows, because I find the indexing service always tries to index stuff in the worst possible time and interrupts my work/gaming/PC activities.)

Also, there's a good chance those 750 GB and 500 GB Seagate HDDs you have are from the 7200.10/11/12 line (but in 2.5" form), and these IME tend to be a lot more problematic than the 7200.9 and 7200.7 series.

I would do it the same way and yes maybe it's a series with problems.

momaka wrote on 2024-10-01, 17:03:

IME, Western Digital tend to do better when you keep them powered On all the time or for extended periods of time and don't power cycle them much.

This reminds me of my first external drive (3.5" by WD). It was extremely annoying that it sends it self to sleep after some few minutes and I couldn't change it...so over the day it had to spin up many times. You had to wait some seconds every time and at least it wasn't good for the HDD, it had soon had a failure. Searching the WWW I found out this "feature" was to save energy (but to kill the HDD). Since then I only use 2.5 external HDDs which don't go to sleep mode.

momaka wrote on 2024-10-01, 17:03:
soggi wrote on 2024-10-01, 02:01:

HDDs could last very long, if you have an eye on the temperature, it should be relatively stable on not to high ranges.

Temperature, power cycles, running hours, air pressure, and noise on the power line are all factors that can affect their life.
I've seen quite a few bad HDDs originating from systems with cheap crap PSUs, and in my experience with cheap PSUs, I've also had a few HDDs run hotter overall due to dirty power from the PSU.

There was a study by Google back then (~15-20 years?) which showed that it's mostly about temperature. Sure, bad voltage/power supply could kill a HDD instantly or within a short time (esp. it's logic board), but let's assume we have a good PSU installed. BTW I guess a chinese firecracker PSU killed my IBM 80 GB HDD in 2004...it was a really crappy 420W PSU and I replaced it with a Tagan 380W which was then very good.

momaka wrote on 2024-10-01, 17:03:
soggi wrote on 2024-10-01, 02:01:

I have no problems with HDDs running more than ten years nearly every day, the HDD (TOSHIBA MK8034GSX) in this IBM/Lenovo T60 works for ~18 years now, one external has even more hours (see screens below)

I see those drives at 20k and 30k hours, respectively... which isn't a whole lot, but it's not that little either. Actually, it's a pretty average figure for a home desktop PC that's been ran for a few hours every day on average. And one can tell it was a home desktop PC because of the high number of power cycles. Work-related desktops rarely get power-cycled and typically have twice those hours with half the power cycles or less. Most of my used HDDs are around that range. Now, my personal HDDs that I bought new, have a lot less hours and even more power cycles, as I tend to standby/sleep the PC when I anticipate I'll be away from it for more than an hour... especially in the summer, when I don't want to add more heat in the computer room. At winter, I tend to be a lot more lenient and sometimes keep my system running all day so that it keeps the computer room a little warmer. 😀

Speaking of high power-on hours, my highest is from a 100GB Hitachi Deskstar 3.5" desktop IDE HDD - it has over 100k hours at this point. No bad sectors so far, though. And before anyone thinks to say that perhaps this HDD counts the power-on time in minutes rather than hours (yes, I have a few HDDs like that), I've checked and this HDD properly counts the power-on time in hours. That's over 11 years of constant (24/7) uptime. Doesn't surprise me, though. I believe I bought it back around 2013 or later and got it with close to 100k hours at that point. So I've put maybe 2000-3000 hours on it and it rolled over 100k.

The two HDDs on the screens are (as said) the T60 internal and the oldest of my 2.5 external ones (counts up to ~ten years)...I also have other HDDs in other computers, which I used earlier (and longer) but can't get a screen of so fast...I think one, two or three have more hours...so yes, the values aren't extremely high, but the highest I had a hand on.

momaka wrote on 2024-10-01, 17:03:

That's one reason I still prefer HDDs over SSDs. If you pay a little attention to them from time to time, you may just get alerted when they are about to fail. Of course, that's not always the case either. Some HDDs will fail completely out of the blue. Use it and all be fine one day... turn it on the next day and it's *click-click-meeep----click-click--meep" ad-infinitum (I'm looking at you Western Digitial WD800JD series!)

Yes, so I could save the very most of the data from the above mentioned IBM 80 GB HDD which was killed by the chinese firecracker PSU. A SSD doesn't give you a sign with it's sound.

momaka wrote on 2024-10-01, 17:03:

Also, another reason why I like HDD is that they could have sat for 10+ years in storage, but the data off them will still be readable if the drive is healthy. With SSDs, we will have to see about that. I imagine the new SSDs that have really tiny memory retention cells may become a lot more likely to give data corruption if left unpowered for many years. I already have flash drives that have done this to me (though if I have to be honest, it's only been the really cheap no-name flash drives with capacitices 8 GB and over.) That's not to say flash memory can't be trustworthy, though. I have some 4 GB drives from 2009/2010 that still retain all of their original data error-free. Nothing can top out my 512 MB SanDisk Cruizer "mini" though - that one my mother gave me way back in 2005 when she got a bunch from her workplace, and this flash drive is still retaining data from back then without errors.

Yes, another important point...

momaka wrote on 2024-10-01, 17:03:

But all in all, I suppose we will see how reliable or unreliable flash memory technology is in SSDs as time goes by. Personally, the only reason why I don't trust modern SSD for long term reliability is because manufacturers are always trying to cram more and more memory cells per given space, thus making it more likely for data corruption to occur. And it seems these days that we're still accelerating more towards becoming a consumer society, no matter where we live around the world. If something breaks, just throw it away and get a new one for cheap... and do the same when that breaks too. With that kind of expectation, manufacturers often design things to last "just long enough". So we can forget about long-lasting stuff to ever be made again like they have in the past. The life cycle of every product is carefully calculated.

The so-called "planned obsolescence". You are from Bulgaria, I was born in the GDR - I think we both know that our countries couldn't afford it to build things that won't last very long back than in socialistic times, the resources weren't there. We're still using the mixer RG28 (-> https://de.wikipedia.org/wiki/AKA_electric_RG28), a wedding present for my parents in 1978 - there's even a documentary (-> https://de.wikipedia.org/wiki/Kommen_R%C3%BCh … n_den_Himmel%3F) about it (sorry, all only in German). Today a new mixer won't last that long for sure.

kind regards
soggi

Last edited by soggi on 2024-10-02, 02:51. Edited 1 time in total.

Vintage BIOSes, firmware, drivers, tools, manuals and (3dfx) game patches -> soggi's BIOS & Firmware Page

soggi.org on Twitter - inactive at the moment