http://www.tweaktown.com/document.php?dType=article&dId=180
Lots of speculation here but interesting nonetheless....
Clocker
Lots of speculation here but interesting nonetheless....
Clocker
75/60 GXP PRoblems Exposed ???
- Thread starter Clocker
- Start date
Tea
Storage? I am Storage!
Groltz, you have returned to form in a big way my man! That's your best shonky link in quite a while. 
Oh. Woops. It's not Groltz the famous tongue-in-cheek Linker to Appalling Reviews Extrordinaire, it's ... er ... Clocker. Are you serious Big C? Or have you started on a new career in the fishing industry? Boat fishing, I mean. The sort where you hang your line out the back and motor along waiting for some innocent fish to swallow your line?
I'll just keep quiet. Bet you Tannin falls for it!
Oh. Woops. It's not Groltz the famous tongue-in-cheek Linker to Appalling Reviews Extrordinaire, it's ... er ... Clocker. Are you serious Big C? Or have you started on a new career in the fishing industry? Boat fishing, I mean. The sort where you hang your line out the back and motor along waiting for some innocent fish to swallow your line?
I'll just keep quiet. Bet you Tannin falls for it!
James
Storage is cool
They didn't really "prove" anything, did they?
GMR has nothing to do with the failures, the technology is used by everyone and has been around for ages.
I think the only useful bit of information is that apparently the failure occurs somewhere on the circuit board - but from there to it being a heat problem with the controller chip? Hmmm.... what does everyone else think?
GMR has nothing to do with the failures, the technology is used by everyone and has been around for ages.
I think the only useful bit of information is that apparently the failure occurs somewhere on the circuit board - but from there to it being a heat problem with the controller chip? Hmmm.... what does everyone else think?
Tannin
Storage? I am Storage!
Tea is seldom wrong about that sort of thing.Tea said:
Let's examine the merit of the TweakTown article.
1: IBM drves have high failure rates. Do they? I have heard so many things about the 45GB Deskstar GXP that I find it hard to credit that there was not a major problem with that particular model, and retain an open mind as to the merits of the rest of the 75GXP line. My own very limited experience with them, a mere six or eight drives in the 20GB and 30GB sizes, has been 100% OK. No failures at all - not that less than 50 to 100 drives can tell you much. And I know how much hysteria can develop about a particular drive once ill-informed people start to gossip about it. I've seen it many times over the years: there soon develops so much smoke and ill-founded rumour that it is impossible to see if there is any fire at the base of it or not.
The 45GB Deskstar 75GXP can probably be considered guilty. The other 75GXP drives are under strong suspicion, no more than that. But if there is any particular evidence that the 60GXP is another horror drive, I've yet to hear of it. TweakTown certainly don't bother with putting forward any.
2: GMR heads are the problem. What a completely ridiculous assertion. All hard drives use GMR heads! IBM introduced MR head technology back in the early Nineties, the improved MRX head in 1997, and GMR heads in 1998. By 1999 all the majors were using GMR as standard, and they still are. GMR is the only competitive hard drive head technology on the market right now. Every single hard drive you can buy uses GMR.
They quote AnandTech: drives "have chips that account for the expansion of platters" which make errors in GXPs but not in other drives. Oh? The whole idea of modern embedded-servo drives - a technology that came to high-end drives in the late 1980's and into general use in about 1992 - is that the positioning information is encoded into the platter itself. Seek speed aside, this is the great advantage of voice-coil drives over the dreadful old stepper motor drives of 5.25 inch 20MB days - they are more-or-less immune to the thermal expansion problems that plagued even the best of stepper drives.
In any case, the argument makes no sense because, as we shall see later, IBM drives use ceramic platters which (if I remember my high-school science correctly) expand far less with heat than do the aluminium platters used by other drive makers.
Our expert reviewer says "I believe IBM is already using this glass platter technology in some of their drives". Is that right? Well, fancy that! And here I was selling 1.08GB TravelStar 3LPs back in 1997 and thinking that they had glass platters.
So much for page one of the article. Shall I continue with page two? Or would that be tedious?
There is a lot of fluff and distraction in the article. However, if you can repair the75GXP problem with a circuit board replacement then that is signifigant knowledge. That means that data can be restored. That also means that the problem is limited to the circuit board and not the interior temp, or the heads, or the platters ...
Bozo
Storage? I am Storage!
IBM was touting their glass platter technology long before the 75GXP. One would have to presume that the glas platters are used in the Deskstar family of drives. I might be wrong, but I seem to remember reading that the 75GXPs used glass platters on the IBM site.
The heat issue could be a valid one. My 75 GXPs are 2 years old and working fine. All three of them are in DataStor removable carriers with no fans. The carriers have vent slots in them for cooling. My case has 6 fans in it. Two blowing air into the case at the motherboard and 4 sucking air out of the case. The idea was to create negative pressure in the case and drawing air in through the drive carriers. Also, I left the top cover off the drive carriers. The room the computer is in is in the basement and very seldom gets above 65F.
I also installed 75GXPs in computers I built at work. These computers run 24/7 in an industrial enviorment. They are in an air conditioned cabinet that is at 70F. Been running for 9 months without a problem. (I do check them once a week though)
So maybe there is a heat problem with the IBMs.
Bozo
The heat issue could be a valid one. My 75 GXPs are 2 years old and working fine. All three of them are in DataStor removable carriers with no fans. The carriers have vent slots in them for cooling. My case has 6 fans in it. Two blowing air into the case at the motherboard and 4 sucking air out of the case. The idea was to create negative pressure in the case and drawing air in through the drive carriers. Also, I left the top cover off the drive carriers. The room the computer is in is in the basement and very seldom gets above 65F.
I also installed 75GXPs in computers I built at work. These computers run 24/7 in an industrial enviorment. They are in an air conditioned cabinet that is at 70F. Been running for 9 months without a problem. (I do check them once a week though)
So maybe there is a heat problem with the IBMs.
Bozo
Vlad The Impaler
Learning Storage Performance
30 GB GXPs are bad as well. We sold over 100 of the blighters, and we have had a 20% thus far. Always the same, they sound like a cement mixer! Bearing in mind that the odd basic server has one or two of these in, this is a right pain. I do not think that it is board related though.
On a related note, has anyone had much experience with Quantum 8.4GB LCT drives? They have a Philips chip on their boards that gets red hot. I have 4 at work that I have retrieved from PCs, all with this chip literally melted. The fifth one had a head fault, but the Philips chip gets way to hot to touch after a few minutes.
On a related note, has anyone had much experience with Quantum 8.4GB LCT drives? They have a Philips chip on their boards that gets red hot. I have 4 at work that I have retrieved from PCs, all with this chip literally melted. The fifth one had a head fault, but the Philips chip gets way to hot to touch after a few minutes.
Vlad The Impaler
Learning Storage Performance
Sorry, I meant 20% failure rate. Doh!
I've had both a 15GB 75GXP and a 41GB 60GXP fail with ye old grinding noise. My first reaction was #@%^ing IBM drives. This happened about the time the story broke.
Since then I've had time to think about what happened. This is what I've learned.
At about the time both crashes happened, I was using both drives for benchmarking purposes and moving them from machine to machine. Sometimes I was bumping them about (not hard!, but they did get bumped - or the machine did), whilst they were on.
Sonn after they developed bad sectors and grinding noises.
Running IBM's DFT software and low-level formatting them restored them to good graces and they have been running fine ever since. I have also been careful not to move the machine whilst they are running. Zero problems.
A couple of things to think about.
Are the IBM's more sensitive to operating shock than their competitors? I certainly treated WD and Quantum drives the same way without drama - though none had the platter sensity of the IBM's, except the WD which I know I have never treated the same way.
Would users admit to bumping drives around whilst they were on? I can imagine how the RMA or warranty claims would go.
How many end users perform a low level format when returning the defective drive? I know a few that have worked fine for six to nine months after a low-level format. Then again, is this something that the end user should have to deal with?
There may be problems with some IBM drives, but I don't think it is as wide spread as the FUD Meisters would have you believe.
My experience anyway - user error ...
Since then I've had time to think about what happened. This is what I've learned.
At about the time both crashes happened, I was using both drives for benchmarking purposes and moving them from machine to machine. Sometimes I was bumping them about (not hard!, but they did get bumped - or the machine did), whilst they were on.
Sonn after they developed bad sectors and grinding noises.
Running IBM's DFT software and low-level formatting them restored them to good graces and they have been running fine ever since. I have also been careful not to move the machine whilst they are running. Zero problems.
A couple of things to think about.
Are the IBM's more sensitive to operating shock than their competitors? I certainly treated WD and Quantum drives the same way without drama - though none had the platter sensity of the IBM's, except the WD which I know I have never treated the same way.
Would users admit to bumping drives around whilst they were on? I can imagine how the RMA or warranty claims would go.
How many end users perform a low level format when returning the defective drive? I know a few that have worked fine for six to nine months after a low-level format. Then again, is this something that the end user should have to deal with?
There may be problems with some IBM drives, but I don't think it is as wide spread as the FUD Meisters would have you believe.
My experience anyway - user error
...
Tannin
Storage? I am Storage!
Reposted from last night: complete with a spelling correction this time!
3: Heat is the problem. Both the 60GXP and the 75GXP "run so damn hot", we are told. Once again there is no evidence raised to support this claim. But having made an unsubstianted allegation, the writer then cheerfully accepts it as gospel truth and makes it his springboard for a whole new line of investigation, wondering just how the heat destroys GXPs, and getting into some rather silly detail. "I believe not only the platters are operating at a temperature unsustainable, but also the chips", he says, meanwhile having entirely failed to demonstrate (or even claim!) that either one is any hotter than the platters or the electronics of similar drives from other manufacturers. Is a GXP hotter-running than a D740 or a Caviar? Hell, I don't know. But I do know that I'd want some actual evidence before I made such sweeping claims.
It doesn't stop our boy though. IBM drives are particulary prone to overheating, he has decided, and he trumpets out his grand conclusion that "IBM must address this problem immediately before more victims are claimed." Who knows? Maybe it's true. But despite his grand claim to "have proved" that cooling is the issue, the artiticle hasn't managed to present any evidence at all. Let's read on.
4: Replacing circuit boards cures GXPs. Now we get to the really good bit. "We have proven that replacing the circuit board of a few IBM 40GB 75GXP drives (thus replacing the controller chip) ... with the same circuit board from ... another working 75GXP drive fixes the problem." That's right! Replacing circuit boards can fix drive problems! That's why we used to swap circuit boards on Miniscribe 3650s back in the 1980s. It was routine stuff then and it is still routine today, though nowhere near as common as it used to be because the relative balance between replacement cost and per-hour labour cost has changed, because modern drives are smaller and more complex, because now that drives have three year warranties it is usually pointless outside dedicated data recovery operations, and because the pace of change is faster now and an out-of-warranty drive is usually worth next to nothing anyway. Observing that swapping a circuit board fixes some faulty hard drives is akin to observing that some days the sun shines and some days it rains: it tells us nothing that we couldn't have guessed already.
Same as the rest of the article, actually.
-------------------------
My rating:
Presentation: 8/10
Imagination: 7/10
Language skills: 6/10
Logical deduction: 3/10
Empirical methods: 1/10
3: Heat is the problem. Both the 60GXP and the 75GXP "run so damn hot", we are told. Once again there is no evidence raised to support this claim. But having made an unsubstianted allegation, the writer then cheerfully accepts it as gospel truth and makes it his springboard for a whole new line of investigation, wondering just how the heat destroys GXPs, and getting into some rather silly detail. "I believe not only the platters are operating at a temperature unsustainable, but also the chips", he says, meanwhile having entirely failed to demonstrate (or even claim!) that either one is any hotter than the platters or the electronics of similar drives from other manufacturers. Is a GXP hotter-running than a D740 or a Caviar? Hell, I don't know. But I do know that I'd want some actual evidence before I made such sweeping claims.
It doesn't stop our boy though. IBM drives are particulary prone to overheating, he has decided, and he trumpets out his grand conclusion that "IBM must address this problem immediately before more victims are claimed." Who knows? Maybe it's true. But despite his grand claim to "have proved" that cooling is the issue, the artiticle hasn't managed to present any evidence at all. Let's read on.
4: Replacing circuit boards cures GXPs. Now we get to the really good bit. "We have proven that replacing the circuit board of a few IBM 40GB 75GXP drives (thus replacing the controller chip) ... with the same circuit board from ... another working 75GXP drive fixes the problem." That's right! Replacing circuit boards can fix drive problems! That's why we used to swap circuit boards on Miniscribe 3650s back in the 1980s. It was routine stuff then and it is still routine today, though nowhere near as common as it used to be because the relative balance between replacement cost and per-hour labour cost has changed, because modern drives are smaller and more complex, because now that drives have three year warranties it is usually pointless outside dedicated data recovery operations, and because the pace of change is faster now and an out-of-warranty drive is usually worth next to nothing anyway. Observing that swapping a circuit board fixes some faulty hard drives is akin to observing that some days the sun shines and some days it rains: it tells us nothing that we couldn't have guessed already.
Same as the rest of the article, actually.
-------------------------
My rating:
Presentation: 8/10
Imagination: 7/10
Language skills: 6/10
Logical deduction: 3/10
Empirical methods: 1/10
timwhit
Hairy Aussie
Excellent review of that article Tannin. Maybe you should go into business of reviewing web articles. Private consultants can make some good money... ask Mercutio...
-Tim
-Tim
Mercutio
Fatwah on Western Digital
Allow me to correct that: Private consultants used to make good money. Now we post on web forums and whine a lot.
Heat: Two of the 75GB monsters are sitting in one of my cases, one on top of the other (and a floppy on top of that)... in a RAID0 array, with minimal additional cooling (there's an intake fan below the lowest drive). And the drives are fine. Nary a peep. They get warm, but not as warm as the first-gen Barracuda ATAs did.
Heat: Two of the 75GB monsters are sitting in one of my cases, one on top of the other (and a floppy on top of that)... in a RAID0 array, with minimal additional cooling (there's an intake fan below the lowest drive). And the drives are fine. Nary a peep. They get warm, but not as warm as the first-gen Barracuda ATAs did.
timwhit
Hairy Aussie
Mercutio said:
Thats why its funny...
Mercutio
Fatwah on Western Digital
tim, you forgot your
...grumble damn aussie judge grumble grumble.
...grumble damn aussie judge grumble grumble.
Tannin
Storage? I am Storage!
Hatchet job: yup, that's as good a description as any. Been far too long since I took the gloves off and laid into a review properly, as in the old days. Good for the soul. :mrgrn:
In fact, I'm in the mood for another one. Any more handy targets kicking round? There is nothing quite like the taste of the froth at the back of one's mouth as one prepares to go over the top and stick the Doc Martins in.
In fact, I'm in the mood for another one. Any more handy targets kicking round? There is nothing quite like the taste of the froth at the back of one's mouth as one prepares to go over the top and stick the Doc Martins in.
Tannin
Storage? I am Storage!
Actually, I don't have a VCR. I lent it to a friend a while ago and when she asked what time I wanted it back I said, "er ... don't bother about it". That was a couple of years ago. I turn my TV on now and then though. I've quite possibly used it once already this year.
But as for Clockwork Orange, yeah. Good pick-up LiamC. No prizes for guessing that I've never seen it, but I did read the book, maybe 25 or 30 years ago. Must have made quite an impression: I can't remember any of it!
But as for Clockwork Orange, yeah. Good pick-up LiamC. No prizes for guessing that I've never seen it, but I did read the book, maybe 25 or 30 years ago. Must have made quite an impression: I can't remember any of it!
Vlad The Impaler
Learning Storage Performance
What we were after now was the old surprise visit. That was a real kick and good for laughs and lashings of the old ultraviolent.
Tannin
Storage? I am Storage!
Bump to make this easy to find.
Tannin
Storage? I am Storage!
I was speaking to a gentleman who owns one earlier today, LiamC, mentioned this thread; thought it might be easier to find now. As for the ultraviolence, well, that depends on just who you have in mind. :eekers:
UV - that's simple. The person who let the 75GXP's out the door Ditto the same person who let Fujitsu's little numbers proliferate.
Doesn't this fill anyone with dread? Fujitsu has bought IBM's Deathstar business, and F's recent past with denying anything wrong with one series of hard drives in the face of overwhelming evidence. Who'd take a chance on customer "service" like that?
Ditto the same person who let Fujitsu's little numbers proliferate.Doesn't this fill anyone with dread? Fujitsu has bought IBM's Deathstar business, and F's recent past with denying anything wrong with one series of hard drives in the face of overwhelming evidence. Who'd take a chance on customer "service" like that?
Tannin
Storage? I am Storage!
Nope. Hitachi. Fujitsu just got out of the desktop storage business. I don't think they'd take the Deathstar Division if you gave it to them. Hell, they had a desktop manufacturing divison of their own up until six months or so ago - and they threw it away.
Just dropped by to verify my forums login (as Doug suggested), and what do I see? A 75GXP thread, right st the top! I just had to have a peek! :wink: Too bad I forgot to check the posts dates until I almost finished reading the thread...
Anyho, bumped into those pearls on the way :
Apparently only bad things remain constant in the computing world (e.g. replacing drive PCBs - i.e. drives failures), the good things (e.g 3.y. warranties) tend to come and then rapidly disappear... But then again, ain't the universe itself - not just computing - a lot like that? ;-)
So, Tony, what was it about them hard drive warranties that you said? :diablo:
Anyho, bumped into those pearls on the way :
Tannin said:
Apparently only bad things remain constant in the computing world (e.g. replacing drive PCBs - i.e. drives failures), the good things (e.g 3.y. warranties) tend to come and then rapidly disappear... But then again, ain't the universe itself - not just computing - a lot like that? ;-)
So, Tony, what was it about them hard drive warranties that you said? :diablo: