Samsung HD not working on old machine...

[irlandes]

I am not going to tell my final Linux goal, because that is posted on another thread with lots of good help. My problem is, I need to do some things to a Samsung HM100JC 100MB laptop HD, and the only machine I can do it on is an old Compaq 5441, 400MHZ desktop machine.

That machine works good albeit slowly with an old Maxtor 32049H2 20GB HD with three distros and Win 98 on it.

I have several laptop to desktop adaptors. In the past, one was used in this 5441 with a Hitachi DK227A-41 4GB HD. It worked until the Hitachi got damaged, now the Hitachi sounds like a blender with bolts in it.

I have tried the Samsung every possible way, including no jumpers, middle plug and other HD removed, plus other combos. An old HD from a same age Compaq worked with no adaptor of course.

Yet, this Samsung will not work at all on one of the adaptors.

To me, all these HD seem to be the same. ATA 100 interface.

Yes, power plugged in; yes, red wire on pin #1 next to jumper pins. Yes, adaptor plugged in correctly. Yes, I have tried all three adaptors.

Am I missing something installing this new HD?

(It works in the laptop where it will end up. Note I am not telling why I am doing this because it will violate rule on multiple posting, but it really would help if I could look at this HD on the old desktop. Then, I would go back to original problem.)

Thanks for any tips.

[saikee]

Ever thought of putting it into an hard disk enclosure and hook it up as a USB hard disk? If it can be read by the computer externally then the problem is half-solved.

100Gb 2.5" laptop should be downward compatible to work in lower UDMA modes but old laptop Bios may have a limit to cope with large capacity hard disks.

I believe laptop Bios are not as standardized as those in desktops.

[banzaikai]

Two things to check (off the top of my head)...

First, the 44-pin IDE adaptors I have are hard-wired to make the drive master. Normally, if you need to transfer stuff from one to another, you just wire them up as pri/sec master (single) on each port. NO CDs, NO slaves of any sort. Not a real problem, except when you deal with Compaq, as they love the "Cable Select" method of doing things. I use dedicated single-drive ATA-100 cables for each one of my adaptors, to ensure I don't have this problem. So, that's the first thing I'd check. If the cable has the labels "Master/Primary/Drive0" and "Slave/Secondary/Drive1" by the connectors, you've got a "CS" cable. I always replace these with "normal" ones anytime I find one.

Second, you may have (as saikee mentioned) a BIOS limitation on the Cpq5441. I took a gander at the specs here (http://h10025.www1.hp.com/ewfrf/wc/document?lc=en&lang=en&cc=us&docname=c00031041&dlc=en&product=93344&printable=yes&), but could find no BIOS update files or max HD specs. My guess is that it originally had the 4.3GB drive, and a maximum BIOS size limit of 32GB, which is why it sees the 20GB just fine, but the 100GB has a hissy fit. You don't have a "compatibility mode" jumper on this drive, so you need an updated BIOS (couldn't find one at Compaq's site), or another computer without the limitation.

{FYI: The first BIOS limit users ran into was when HDs hit the 32MB range. This was due to how the drive's sectors were mapped, and how many bits were used to map them. After playing with cluster sizes, geometry tables, and bit densities (FAT-12/16/32), they then expanded the limits to: 540MB, 2GB, 4GB, 8GB, 32GB, 132GB, and 4TB(?). You can usually tell where your mobo fits by seeing which drive is correctly ID'ed and which is not. In our case, the 20GB shows up, but the 100GB doesn't = 32GB limit.}

You may want to look into downloading the "Ultimate Boot CD (UBCD (http://www.ultimatebootcd.com/download.html))". It has many utils to check the BIOS limits, drives (most models are supported), and darn near everything else.

banzai "bigger isn't always better" kai

[irlandes]

Once again JL supplies answers! Thanks so much.

I am aware of the external enclosures and some other options. I am not a bankrupt old retiree, but I do tend to budget my computer expenses very carefully to keep within established constraints to meet financial goals. (We are building a house in rural Puebla.)

So, when I have a task, my first goal is to try to do it with existing hardware. In this case, it seemed as if that old Compaq was the answer.

What you both say about BIOS and/or c/s problems makes sense. At this point, I am betting on BIOS.

I have spent enough time on it to make it reasonable to try some more hardware if cost is not terrible.


I have to baby sit a grandson this evening. So, I think I will stop and run out to Wal-mart and look at external HD. Based on another thread, it seems I can clone to the HD, then clone back to my new one (using live CD) and when it hits the end it stops.

If not, I will evaluate the cost as to whether to return it. (I don't like that but a significant percentage of women's clothes are routinely returned.)

I had considered an external HD, but my machine only has USB 1.0. I assumed that would not work. My son-in-law thinks it will work just to slowly it is not considered usable. I can let it run all week. :D

The external HD is probably more cost effective for the long term than an enclosure, since the HD can be used at all times. And, on future machines.

I will return, but maybe not for a day or two. Thanks a million!!!

[irlandes]

Reminder. Thanks to Saikee I have plenty of info on cloning, so I do not wish to prompt further discussion here on that aspect of the problem. Thanks.

[irlandes]

Back from the store. I bought a Maxtor (Seagate?) 160GB Personal Storage 7200 RPM device, $74.


I decided it is the sort of thing I can use long-term anyway. (Thereby causing the rapid deployment of some new tecnology which will obsolete it instantly, sigh!)

In the worst possible case, if I can't clone using it, it is still a good thing to have. In that case, I will simply build up the new HD app at a time, and keep the old HD in case I need Windows for one of the government URL's we have to use.

[IsaacKuo]

If it's a BIOS limitation, one way to get around it is to go into the BIOS and manually set the hard drive statistics. On my older Pentium II and Pentium III machines, the BIOS freezes up when trying to detect any 48bit LBA hard drive (any hard drive bigger than 100gigs, and almost every modern hard drive even if smaller than that). However, they work just fine (with Linux) when I manually set the hard drive stats.

The manually stats I enter do NOT match the hard drive's true stats. Actually, I just connect up an old 20gig drive and use the stats from that drive. These settings are apparently "close enough" to boot up GRUB from the first partition. From then on, Linux ignores the BIOS and is able to access the rest of the drive just fine.

[irlandes]

I hope I can add that BIOS trick to my personal Linux commands and tricks file! Sounds good to know.

My new $74 external HD is formatted in NTFS. I think I am going to attempt dd using Puppy, to clone my old HD to it then if that works clone back to the new HD. If I screw it all up, I will reformat external HD as FAT 32, should be able to fix boot loader just like any HD, then format as Fat 32 for storage, using Qparted in Kubuntu 7.04 live or Mandriva install disk.

And, then, as I said, simply build up Samsung the hard way. :confused:

Of course, I am hoping I can also do my cloning. I try to have back-up plans.

Will let you guys know how badly I messed it up. :p

[banzaikai]

IsaacKuo
If it's a BIOS limitation, one way to get around it is to go into the BIOS and manually set the hard drive statistics.
I'm guessing you haven't played around with many Compaqs, huh? Unless it's an older DeskPro, the BIOS on those things are "auto-everything", since Compaq assumes their customers are complete blithering idiots.The last one I worked on (6000 series), I had used to test an old Fujitsu 10GB I had lying around. On the computer I pulled it out of (DFI), it would claim an HDD controller failure or ID the drive as a garbage string (F]]it+u4@&6). On the Compaq, it wouldn't even POST. At least the DFI mobo would let you know which direction to look. There was no way to preset the drive parameters, as it would just auto-detect at the next boot. So, irlandes may be able to punch in the right numbers, but if the thing auto-detects parameters that seem out-of-range for what he typed in, then the computer will halt with an error (or not POST at all).

Things about Compaq/HP that p!ss me off:
- "CS" cables for PATA drives
- Auto-everything BIOS (even disabling something doesn't always guarantee it's disabled)
- Diagnostic partitions
- Restore partitions
- Anemic power supplies (and the use of smaller "S"-class sizes, so it can't be easily swapped)

Now, for those reading who do have to punch things in manually, I offer the following:
520MB limit -> 1024 Cyl, 16 Hds, 63 Sect
2GB limit -> 4096 Cyl, 16 Hds, 63 Sect
8GB limit -> 16384 Cyl, 16 Hds, 63 Sect
32GB limit -> 65536 Cyl, 16 Hds, 63 Sect

Note: Some BIOS makers used the "zero point" when calculating cylinders, so you may have to subtract 1 from that value to get it to set (e.g., 1024-1=1023). It's rare, but that may also apply to heads, too.

irlandes
I bought a Maxtor (Seagate?) 160GB
I will reformat external HD as FAT 32
Ummm. I believe that FAT-32 has the 132GB per partition limit. You may be able to prep it as two 80GB, but I haven't used FAT-32 with anything that big (120GB was as big as I ever got around to trying), so I'm not sure if it'll even see the whole thing. Be warned.

OTOH, if you'll be throwing multiple OSes on this thing, then you're going to have some ext3/Reiser partitions, so it should work just fine (Windows won't see Linux partitions unless special utils are used). Newer distros now use the NTFS-3g utility to R/W, so you could go back to using that if you want.

Also, I've used external drives to copy/mirror computers I'm working on, and can attest that they do work fine with USB 1.1, albeit very slowly. We're talking "go find another computer to work on" or "pull 'Moby Dick' off the shelf" slow.

banzai "Call me Ishmael..." kai

[saikee]

I believe that FAT-32 has the 132GB per partition limit

I believe a fat32 partition has a limit considerably bigger, something like 2 terrabytes. I have used at least 320Gb fat32 partitions and possibly 400Gb before. I changed to ntfs because fat32 does not support a file larger than 4Gb.

[bwkaz]

I believe a fat32 partition has a limit considerably bigger, something like 2 terrabytes. I'm not sure what the limit actually is, but the 132GB thing is an XP-only limitation. You can create vfat filesystems much larger than 132GB, if you use win98 or Linux tools to do it, but if you try to create a filesystem larger than that size on XP, it'll tell you it has to be NTFS.

Which is completely wrong (the FS can support much larger partitions), but that's the decision they made. I believe they made it because the design of vfat made it really hard to use on really large partitions (e.g. it has no list of free blocks -- which didn't matter when FS sizes were 10 megs, but is a serious limit now that they're multiple hundreds of gigs...).

[saikee]

Is it not true also that the cluster size increases proportionally with the partition size in a fat partition, making it a ridiculous large and uneconomical?

[irlandes]

Right you guys were. (Of course.) After cloning/cloning my old 20GB HD on to the new Samsung, via a new external HD, I plugged the old 20GB Hitachi into the old Compaq, and it started to boot correctly into the old OS. Since I knew it would crash and burn when it tried to initiate nonexistent hardware, I shut it off. Proof that BIOS was the problem verified.

Thanks!

[bwkaz]

Is it not true also that the cluster size increases proportionally with the partition size in a fat partition, making it a ridiculous large and uneconomical? The cluster size does increase with the partition size, yes. I'm not sure of the point where it would qualify as uneconomical, though -- the only thing the cluster size affects is the granularity of allocations. If you have a 64K cluster size, then any file's size-on-disk will be a multiple of 64K.

And if you have a 200GB disk, I'm not sure that wasting (on average) 32K per file is going to matter all that much. Though it depends on what you put on the disk, of course.

So yes, maybe it'd be a problem, or maybe not. :)