Problem with big disks

[drsolly]

I've found a major problem that affects IBM 37 gb EIDE drives for sure, and might affect other drives that exceed the 33.8 gb barrier.

I'll be happy to explain it here, plus the
rather unsatisfactory work-around, but I've
noticed a lot of complaints that info in the
magazine is just repeating info that was
on the message boards.

So, should I explain the problem here, or would people prefer to wait a month or so to read it in the magazine?



------------------

[Charles]

As has been said many times by Bob Crabtree
Editor of Computer Video he repeats selected
items on this board in the mag for the benefit of readers not connected,
I'm sure this info would fall in that category.
Charles

------------------

[bcrabtree]

Alan,

DO please post here!

Bob C

[David]

Having already followed your link from the "Hard Drives and Controllers" section - I think there should be more than a "shall I - shalln't I" comment here.

However, as a marketing ploy - you could probably treble the sales of the next issue by only revealing the problem and workaround in the mag http://www.safety-online-internet.com/host/computervideo/ubb/smile.gif But it's OK - Bob has already said you can reveal all!

I wait wioth baited browser!

D

[drsolly]

IBM are shipping 37 gb EIDE drives, so I bought a few. Then I tried to install them in an Intel motherboard. No luck. So I put in a Promise card, the usual solution to putting
more or bigger drives on a PC. Again, failure.

I spent quite a long time trying to make this work, I was sure that an IBM drive in an Intel motherboard had to work. What I got, was a 4 gb drive; 10% of the space I'd paid for. Some of the ways I tried to install it, I didn't even get that.

Eventually, in desperation, I went to the IBM web site to see if the specs of the drive might give me a clue to what was wrong. Buried deeply (under "Jumpers and configuration" of 37 gb drives), I found this:

http://www.storage.ibm.com/techsup/hddtech/bios338gb.htm

This explains that with the Award 4.5 Bios, and with the Promise card, there's a 33.8 gb limitation. Somewhere in the LBA design, someone put a two-byte word to count something. Well, I can tell you, it isn't
just the Award 4.5 bios. It's also in the Bios that Intel were shipping on their motherboards last year. IBM's workaround is a jumper that limits the drive to being 33.8 gb. There's the same fix for
34 gb drives.

The fact that this jumper exists, shows that IBM knew about this problem when they designed the drive, and when they shipped it. If only someone in IBM had had the intelligence to include a piece of paper with the drive that explains this, I'd have been saved a lot of time (several days) and trouble. Even better, put it on a sticky label on the drive.

I guess there's a similar problem with any drive greater than 33.8 gb.


------------------

[PD HEXUS]

Alan,

I found this problem with the the 34GXP's on our Sharptooth DC1000 review system.

At that time I raised the issue with IBM's brilliant R&D team.

They informed me that Promise had just submitted them a Beta BIOS for the FastTrack66 that apparently resolved the issue.

They did not mention anything to me about it being a combined Award 4.51PG/Promise issue.

If I'm not mistaken the BIOS on your Intel OEM mainboard is a hardwired Phoenix so if there is an issue it does not seem to be restricted to just Award 4.51PG and Promise FastTrack66 BIOS variants.

As large Ultra-ATA drives are becoming so commonplace hopefully a fix to this LBA limit issue will be forthcoming sooner rather than later, otherwise I'm sure you won't be the last one to experience lost time and trouble.


Paul Dutton
Multimedia Direct - Focus on the Action


[This message has been edited by PD (edited 26 September 1999).]

[drsolly]

IBM plainly knew about this issue, even when they were designing the drive. There will *always* be people who have older Promise cards and motherboard Bioses. This problem
is going to affect a *lot* of people.

My main beef is that IBM don't tell you about it. There should be a slip of paper in with the drive, or even better, a label stuck on the drive case, that explains it. Some way of warning you before you purchase would be the most honest thing.

The next bunch of drives I buy, will not be IBM. Great technology - pity I can't make use of it because you forgot to tell me some important technical information. What will you forget to tell me next time?

Oh, and hey, you guys designing Bioses for
interfacing with drives. I remember the 32 mb limit, the 528 mb limit, the 8.4gb limit, and now the 33.6 gb limit. Are you going to be astonished when drive sizes go over 64 gb? 128gb? 256 gb? Or are you going to have the wit to realise that this is likely to happen, so the controller and Bios for the drive should be able to cope.


------------------

[Tunni]

Hi Doc

My brain starts to bleed when things get reeeally technical, however does this limitation apply to Macs?

Regards
Tunni

[Nigel Longman]

Don't I remember reading about Bob C having some trouble with IBM drives some while back, and how patronising they were to him, telling him to leave installation to 'a professional'? Put me off buying IBM for good.

Nigel

[bcrabtree]

Tunni,

We've got a 37GB IBM, so I'll try whacking it into a B&W G3 to find out.

Hopefully, I'll have a similar sized Maxtor arriving in the next few days, so will try that, too, and also see how or whether Maxtor has got around the problem for Windows PCs.

And, Nigel, yes, you did remember correctly. However, my publishing that comment did actually make IBM a lot more attentive to our (the mag's) needs!

Bob C

[PD HEXUS]

Hi Alan,

You're right that there will always be people who have older device controllers and mainboard BIOS revisions and that some of them may be affected by attaching mass storage devices for which the controller was not designed to support.

Was the drive you acquired an OEM or retail unit?

Most OEM drives are shipped "bare" without fancy packaging, documentation, screws or additional block jumpers which along with bulk purchasing allows OEM system ntegrators to achieve a price advantage over a retail boxed unit that should ship with all of the above including a comprehensive installation guide that also provides manufacturer contact details for support of issues that were not determined when the installation guide was compiled.

When I found the LBA addressing limitation with the Promise FastTrack66 to which this thread refers IBM provided the resolution immediately, since that time (when the drive was not available to purchase) they have provided that information on their Web site. I would say that that is a manufacturer that you could have confidence in. The fact that they have made provision for legacy BIOS code by providing jumper settings that will allow an integrator to clip the capacity of the drive to suit is equally commendable.

I also think that it is worth mentioning that this is not an IBM specific issue. It is just that they were one of the first to market with Ultra-ATA/66 drives of this capacity. I believe several other manufacturers have subsequently released
Ultra-ATA HDD of competitive capacity which when integrated into systems with the same device controller BIOS limitation's are likely to have the same "issue".

In the positions that I have been previously engaged, as in this one, I can select whichever HDD to offer to customers.

Having seen the compatibility issues, comparative performance figures and return levels of various HDD from production lines
yielding up to 3000 systems per month, I continue to implement systems based upon IBM hard drives because in my experience they offer my customers the best price/performance/relaibility ratio.

Regards,

Paul Dutton
Multimedia Direct - Focus on the Action

[This message has been edited by PD (edited 29 September 1999).]

[drsolly]

<<My brain starts to bleed when things get reeeally technical, however does this limitation apply to Macs?>>

I haven't the foggiest idea. Macs are really really easy to use, so I guess you plug it in and it works fine. Or it doesn't. Ask Apple,
they should know.




------------------

[drsolly]

<<Most OEM drives are shipped "bare" without fancy packaging, documentation, screws or
additional block jumpers which along with bulk purchasing allows OEM system ntegrators to achieve a price advantage over a retail boxed unit that should ship with all of the above including a comprehensive installation guide that also provides manufacturer contact
details for support of issues that were not determined when the installation guide was
compiled.>>

I probably got an OEM deal, not a retail deal. I rather doubt if you can get these drives retail. But I seriously question if the additional one penny cost of a piece of paper is that great as to jeopardise competitiveness.

I already mentioned that I don't think it's
just an IBM issue.

My complaint is the lack of that piece of
paper.

It isn't IBM drives I'm compaining about. It's the dolt in IBM who didn't have the wit to realise that a lot of people will have a lot of hassle and call IBM causing them a lot of technical support costs. And that a single piece of paper costing less than a penny would prevent all that.

And the issue *was* known when the drive was launched. That's why the jumper exists!

And if IBM's view is that their drives can only be installed by professionals, I'd say that A) I *am* a professional, and that B) anyone who isn't, is being told by IBM to buy their competitors products, so whoever said that to the editor of Computer Video is another dolt.

However, IBM will be please to hear thatI'm planning to order two more of their drives, in the hope that there aren't any other gotchas that they didn't tell me about.



------------------