Old HDD discussion

[Shocker]

The title is pretty self-explanatory. I define "old" for HDDs as a model released at least five years ago. I'll do three of my own first:

Samsung SV1533D
Basic specifications:

• 10.2GB/platter, 3 heads
• 5400RPM (ball bearing)
• Ultra DMA mode 4 (66.7MB/s)
• Media transfer rate: 291Mb/s
• Buffer size: 512KiB (0.5MiB)

The motor driver is an L6269 from ST. I don't know if these have the same problem as SMOOTH. Like the ST3320620A that I have (mentioned elsewhere), there is no discrete RAM chip on this one.

ST3120026A
Basic specifications:

• 80GB/platter, 3 heads
• 7200RPM (fluid-dynamic bearing)
• Ultra DMA mode 5 (100MB/s)
• Media transfer rate: 683Mb/s
• Buffer size: 8MiB

You're gonna love this - there is NOTHING made by ST on this one.

The PC I got it from had one of those Antec PSUs of doom, but (to my memory) none of the caps were popped. I haven't tested anything else from that system.

So far this drive has 34,136 hours, 1,232 cycles and 0 bad sectors. It's currently the largest drive I have with longitudinal recording.

ST380817AS
Basic specifications:

• 80GB/platter, 2 heads
• 7200RPM (fluid-dynamic bearing)
• Serial ATA 150MB/s with NCQ
• Media transfer rate: 687Mb/s
• Buffer size: 8MiB

This one vibrates more than the other 7200.7s of mine, but still less than the 7200.10s.

No, the media transfer rate being different from the ST3120026A isn't a typo. They actually changed the zone layout (older revision, Alpine, did up to 57.3MB/s; this one, Puma, does 58.5MB/s) when they added NCQ to the 7200.7. I'm assuming they did so at the same time they released the 7200.8, because there's no 200GB 7200.7 with NCQ.

(On a side note: I don't think Seagate's claim of native SATA was true for Alpine. As I remember there is a chip made by LSI right next to - and connected to - the SATA connector. I know that Puma is native SATA, and SATA only. There's also a SATA/300 version - Puma 2 - which presumably also changed the geometry again. But in the grand scheme of things, this whole SATA implementation thing doesn't really matter, because none of those drives are fast enough to saturate PATA anyway.)

Was it worth the trouble to redesign to get another 1.2MB/s??? I don't think so. But only Seagate engineers would know for sure.

[Wester547]

Re: Old HDD discussion

ST3120026A
Basic specifications:

• 80GB/platter, 3 heads
• 7200RPM (fluid-dynamic bearing)
• Ultra DMA mode 5 (100MB/s)
• Media transfer rate: 683Mb/s
• Buffer size: 8MiB

You're gonna love this - there is NOTHING made by ST on this one.

The PC I got it from had one of those Antec PSUs of doom, but (to my memory) none of the caps were popped. I haven't tested anything else from that system.

So far this drive has 34,136 hours, 1,232 cycles and 0 bad sectors. It's currently the largest drive I have with longitudinal recording.

I don't want to spoil anything but I also have a ST3120026A (though it only has 2,000 power on hours and 1,500 power cycles on it, but no bad sectors), and it does have a small little ST chip on it.... it's the BUX series, the small protection chip we spoke about a few months ago. Hopefully yours doesn't have that as it is somewhat notorious for breaking down with smoke to boot (though it can be removed and the drive can still function without it but I believe it's there so as to minimize the damage a power spike or surge could cause to a PCB). You might need to stick a flashlight onto the PCB to see it.

The largest longitudinal drive I have is the ST3250824A I've talked about before, it has 21,000 power on hours and 3,500 power cycles, but sadly has one offline uncorrectable sector, one pending sector, 8 reallocated sectors and 9 errors in the S.M.A.R.T. log (the last one for which I'm at fault, for reasons mentioned in the next sentence). It can't complete the long benchmark of HD Tach so the damage must be greater than it seems since the drive seems to function fine otherwise (and can still do its full ~67MB/s of read transfer rate in HD Tune Pro, along with its burst rate of almost 100MB/s in HD Tach). I also didn't know some 7200.7 drives vibrate more than others.

[mariushm]

Re: Old HDD discussion

The oldest I have in my system is this one :

ID	Attribute DescriptionThreshold	Value	Worst	 Data	Status
03	Spinup Time		21	181	176	5916	OK: Value is normal
04	Start/Stop Count	0	100	100	938	OK: Always passes
09	Power-On Time Count	0	49	49	37430	OK: Always passes
0C	Power Cycle Count	0	100	100	892	OK: Always passes
C2	Temperature		0	121	102	29	OK: Always passes

ATA Device Properties	
Model ID	WDC WD4000AAKS-00YGA0
Serial Number	WD-WCAS84337477
Revision	12.01C02
World Wide Name	5-0014EE-1AB49824C
Device Type	SATA-II
Parameters	775221 cylinders, 16 heads, 63 sectors per track, 512 bytes per sector
LBA Sectors	781422768
Buffer	16 MB
Multiple Sectors	16
ECC Bytes	50
Max. PIO Transfer Mode	PIO 4
Max. MWDMA Transfer Mode	MWDMA 2
Max. UDMA Transfer Mode	UDMA 6
Active UDMA Transfer Mode	UDMA 6
Unformatted Capacity	381554 MB
ATA Standard	ATA8-ACS
	

ATA Device Physical Info	
Manufacturer	Western Digital
Hard Disk Family	Caviar SE16
Form Factor	3.5"
Formatted Capacity	400 GB
Average Rotational Latency	4.2 ms
Rotational Speed	7200 RPM
Interface	SATA-II
Buffer-to-Host Data Rate	300 MB/s
Buffer Size	16 MB
Spin-Up Time	13 sec
	
ATA Device Manufacturer	
Company Name	Western Digital Corporation
Product Information	http://www.wdc.com/en

No bad sectors, nothing.
Had a Seagate 640 GB and an even older Seagate 250 GB sata drive, the 640GB got sold, the 250 GB is in my sister's computer.

Outside my system... I have a 4.3 GB ide drive somewhere from my first computer i ever owned, a 40GB ide ... probably wd in my workbench computer, a hitachi 2.5" 34 GB ide in my laptop.. all working fine. I should also still have some 2.1 GB Seagate IDE drives around probably with some bad sectors.

The oldest drives I own.. i posted picture with them here already and on another forum (hence the text at the bottom of the picture) but I can post it again...



The one on the left is plain ide 6.4 GB, works fine, just slow and no longer have use for it.

The one on the right is a 500-550MB fujitsu scsi drive partitioned and formatted to 350 MB to get around some bad sectors. Still worked fine last time i used it about 3-4 years ago, with an Adaptec ISA scsi controller on a 486 system I owned. Had a seriously fast speed of 5-6 MB/s.

[larrymoencurly]

Re: Old HDD discussion

Here's a 100MB Toshiba that still works (louder than an air raid siren) and is so old it was actually made in Japan. The ceramic daughter board has a Toshiba chip and apparently some 4066 CMOS analog switches. I was not able to use any of its 3 relays to fix my garage door opener (contact ratings too low).

[Shocker]

Re: Old HDD discussion

I don't want to spoil anything but I also have a ST3120026A (though it only has 2,000 power on hours and 1,500 power cycles on it, but no bad sectors), and it does have a small little ST chip on it.... it's the BUX series, the small protection chip we spoke about a few months ago. Hopefully yours doesn't have that as it is somewhat notorious for breaking down with smoke to boot (though it can be removed and the drive can still function without it but I believe it's there so as to minimize the damage a power spike or surge could cause to a PCB). You might need to stick a flashlight onto the PCB to see it.

I checked those too. I wasn't kidding when I said no ST on the board. If you get as lucky as that, tell me about it.

This drive doesn't have a separate flash ROM, by the way. The best I can remember, this is the case for all PATA 7200.7s of mine, though the ST380817AS I mentioned does have a separate chip. The only thing that I can tell is made by ST on that drive is the 5V TVS.

I also didn't know some 7200.7 drives vibrate more than others.

That's technically the case for all HDDs - well, anything with rotating parts. The only general statement that can be made is that the average vibration level differs between models. From what I've heard, Samsung drives are notorious for high vibration, which I suspect is causing the reliability problems some people experience with them. But that SV1533D I mentioned has low vibration. And 0 bad sectors (but does report 1 read error). Funny how that works.

EDIT: It's about as heavy as the old Seagate drives, like the ST3120026A.

EDIT 2: I don't really care where the HDDs are assembled, as long as it isn't Thailand .

[TELVM]

Re: Old HDD discussion

[kc8adu]

Re: Old HDD discussion

these are all new drives.
i need to dig out some of mine.
how about a 5 MEGABYTE 5 1/4" full height.
for the younger crowd that takes up the space of 2 cdrom drives.

[eccerr0r]

Re: Old HDD discussion

I've never had an ST506, only an ST412. However it's long been out of service.
My Maxtor 120GB w/2M cache however is still in service.

Model Family: Maxtor DiamondMax Plus 9
Device Model: Maxtor 6Y120L0
ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE
3 Spin_Up_Time 0x0027 204 200 063 Pre-fail Always - 17110
4 Start_Stop_Count 0x0032 253 253 000 Old_age Always - 1735
5 Reallocated_Sector_Ct 0x0033 253 253 063 Pre-fail Always - 1
8 Seek_Time_Performance 0x0027 250 242 187 Pre-fail Always - 45182
9 Power_On_Minutes 0x0032 046 046 000 Old_age Always - 745h+00m
12 Power_Cycle_Count 0x0032 252 252 000 Old_age Always - 398
196 Reallocated_Event_Count 0x0008 252 252 000 Old_age Offline - 1
197 Current_Pending_Sector 0x0008 253 253 000 Old_age Offline - 0
(Many parameters removed for brevity)
Note that 745 hours is not very long. The problem with SMART on this drive is that it wraps around a lot. This particular field wraps around every couple of days. The other problem with SMART on this drive is that it also uses an approximation for seconds - every time 64 (power of two) seconds pass it counts as one minute... so every hour it thinks has passed is actually quite a bit longer than an hour. (Trust me on this, this was the definitive reason why the SMART timestamps seemed not to coincide with real time)

The reason why I bring up hours is that I do SMART tests every so often. The hour count wrapped around too, and now is at 1721 hours. This wrapped around every 65536, so the real hour count is 67257. But NO, due to the 64/60 second anomaly the real hour count is 71740 hours which is over 8 years of power on hours...

Time to get a new drive... someday...

(I also note that field #9 in the SMART log has been going down and down... It's only 46 now. It used to be 100 or 253 or something, I forget... I just wonder about the apocalypse when it hits 0...)

[Shocker]

Re: Old HDD discussion

I have a Samsung SV0322A with 0 bad sectors, but can't get it to work quite right regardless. I'm guessing the electronics are the problem.

Also, someone before me busted pin 28, so I can't use cable select without replacing the connector. Curiously, the firmware chip is socketed, so I could conceivably do a board-swap with an identical drive without soldering. This drive has an EtronTech RAM chip, in case that's the problem.

I wonder if current Western Digital PATA drives are actually SATA designs with a bridge added...

[Shocker]

Re: Old HDD discussion

I think I've found evidence that Seagate originally intended on making the entire 7200.9 product line-up with 166GB platters. I found these model numbers, which I've never seen elsewhere, in the Desk Reference, under the 7200.9 entry in the index:

ST380210A
ST380210AS
ST380810A
ST380810AS
ST3120210A
ST3120210AS
ST3120810A
ST3120810AS
ST3160210A
ST3160210AS
ST3160810A
ST3160810AS
ST3500632A
ST3500632AS
ST3500832A
ST3500832AS

Based on the model numbers, the release models include:

40GB, 1 platter
80GB, 1 platter
120GB, 1 platter
160GB, 1 platter
200GB, 2 platters
250GB, 2 platters
300GB, 2 platters
320GB, 3 platters
400GB, 3 platters
500GB, 4 platters

Another thing about the 7200.9 - StorageReview says the 500GB model has a substantially lower transfer rate than the 7200.8.

(Note that they tested the 8MiB version of the 7200.8, so I figure the 16MiB one might keep up with the 7200.9 in some of the real-world tests. And maybe there's a different firmware version without such poor random write performance.)

Is there really a point in making separate 300GB and 320GB models??? I don't get it.

The 7200.10 line isn't much better, but at least they specify the head counts.

I got a PC with an ST380013AS and, although it has the controller and flash ROM made by ST, as well as an EtronTech RAM chip, it works with 10,152 hours, 1,258 cycles and 0 bad sectors.

Not only are the SATA lines connected to a bridge chip, and multiple lines from the bridge to the controller, but also this bit of weirdness: It won't spin up unless you actually connect it to a computer, unlike most other drives (including the PATA models) which will start as soon as they have power.

Maybe some of those ST controllers are bad, but not others...IIRC the bad ones were manufactured in Taiwan. The one on this drive was made in Malaysia.

The SV0322A I mentioned makes some unsettling noises (probably the bearings) so I don't know if I should even try fixing it.

[Wester547]

Re: Old HDD discussion

Not sure if location of manufacturer matters much. You'd think the quality is equally bad whether it's from China, Malaysia, Singapore, Taiwan, Thailand, or even Morocco, etc (I think the reason why HDD warranties, for an example, tend to be longer when made in Thailand than Malaysia, at least observing that with Western Digital, is more factories being in Thailand, hence why the floods of last year weighed so heavy upon HDD production and availability). All integrated chips have a varying quality or tolerance when such a huge volume is produced and some will simply do better than others. I've seen SMOOTH chips fail a few thousand hours after 10,000 hours of use and many more power cycles, so I think 10,000 hours is too premature to bode longevity. I know, SMOOTH chips are undoubtedly bad, but still.

I don't even know if EtronTech RAM chips are bad, I've just seen complaints about them on other forums, but that was from people who were trying to overclock their GPU RAM successfully, so...

Really? SATA drives don't run on their power connector alone? I guess that's their way of trying to be "green" by comparison to molex connectors.

EDIT: I know, you meant the microcontroller.

[Shocker]

Re: Old HDD discussion

I've seen SMOOTH chips fail a few thousand hours after 10,000 hours of use and many more power cycles, so I think 10,000 hours is too premature to bode longevity.

I get that, but I wasn't talking about the SMOOTH chips. I was talking about the big 176-pin chip that is the largest on the board.

Really? SATA drives don't run on their power connector alone? I guess that's their way of trying to be "green" by comparison to molex connectors.

It's just a quirk of the particular drive mentioned. Newer Seagate SATA drives (including the ST380817AS) don't exhibit this behaviour. I'm pretty sure it's caused by the bridged design on the ST380013AS.

[Shocker]

Re: Old HDD discussion

That last post doesn't tell me anything. Tell us what type of drive it is.

[Shocker]

Re: Old HDD discussion

The newest Fujitsu drive I have is an MPE3084AE:

• 8.6GB/platter, 2 heads
• 5400RPM (ball bearing)
• Ultra DMA mode 4 (66.7MB/s)
• Media transfer rate: Unspecified
• Buffer size: 512KiB (0.5MiB)

Fujitsu says the sustained transfer rate is 20.3-34.4MB/s. What a load of crap...

• Measured: 9.3-22.2MB/s

Maxtor did the same thing. The 90340D2:

• Specified: Up to 18.6MB/s
• Measured: 8.6-12.8MB/s

ST3120026A for comparison:

• Specified: 32-58MB/s
• Measured: 28.9-57.3MB/s

This time, only a small difference that pales compared to the utter LIES from Maxtor and Fujitsu. The MPA3026AT doesn't manage much more than half of the specification:

• Specified: 8-15MB/s
• Measured: 4.5-8.7MB/s

Does anyone know of any other HDD manufacturers that lie(d) about transfer rate???

EDIT: A Seagate Medalist 4312 I have doesn't actually support UDMA 4 (not that it matters) and doesn't report the raw value of the runtime AT ALL.

[Pentium4]

Re: Old HDD discussion

I love how reliable old IDE drives are. I have an old Western Digital from 1998 and it has 66,000 running hours on it from a PSU with horrible ripple most of its life.

[eccerr0r]

Re: Old HDD discussion

I've had old and new IDE disks die, haven't really seen much of either way. Perhaps except those stepper motor hard drives. Those have never died on me...

But then I'll die of old age first waiting for one to seek...

[Shocker]

Re: Old HDD discussion

[nothing]

[Shocker]

Re: Old HDD discussion

I used the ST3120026A mentioned to store a copy of some DVD rips (not all of them as the total is >240GB). I also got a Samsung SP4002H:

• 40GB/platter, 2 heads
• 7200RPM (ball bearing)
• Ultra DMA mode 5 (100MB/s)
• Media transfer rate: 557Mb/s
• Buffer size: 2MiB

to store a copy of CD rips (all of them this time).

Another thing involving the Maxtor 90340D2 is this confusing statement in the manual:

The DiamondMax™ 3400 drive design allows greater shock tolerance than that afforded by larger, heavier drives.

The only 5.25" HDD still in production at that time was the Quantum Bigfoot. Why would anyone who cares about performance get one of those???

And as I stated elsewhere my experience is that heavier drives are more reliable. So what am I to make of that statement???

[Steve.B]

Re: Old HDD discussion

Talking of older hard drives, did any of you spot Dave Jone's little tear down a couple of weeks back?

https://www.youtube.com/watch?v=CBjoWMA5d84

Now that's what you call a hard drive!

[Shocker]

Re: Old HDD discussion

I just fried the 90340D2.

What happened was that I had the power connector backwards when about to plug it into a PSU, and the connector wasn't adequately shrouded - so the pins touched. The shroud on this one is clearly shorter than those on the other drives mentioned. So there's another problem with Maxtors.

The fried chip is a Texas Instruments PTLS2270A-K. Nothing else is visibly damaged.

More about the SP4002H: It appears to be heavier than 1-platter Barracuda ATA IV/V/7200.7 (but lighter than the 2-platter versions), but it's hard to compare the two holding them in my hands because of different shapes (SP4002H has flat base, Barracuda ATA IV and later have flat lids). I think it's relatively quiet for a ball-bearing drive, but still way louder than anything with FDB.