PDA

View Full Version : Hard disk bad track 0, date of manufacture codes



Allen
January 20th, 2009, 05:54 PM
I've been trying to fix four 30 MB ESDI hard disks for an IBM PS/2 Model 50Z (another one). The original disk was bad, and so I acquired three more (two were new old stock, and one was used). All of them come up with invalid media--bad track 0 errors when I try to format them. Some give me error codes 1780 or 1790 on startup, but this isn't always consistent. The disks in question are WD-336R (three) and WD-336RT. Internet searches didn't help. If a hard disk has a bad track 0, does this mean there is no fix?

One last unrelated question: I have a Model 30 PS/2 (8086) and I'm trying to figure out the (roughly) the date of manufacture. There are no actual dates on the system, but the cover has an unusual date stamp--see attached picture. What is this date? Only the day is clear to me. The year must be 1987 or 1988, when these PS/2s were sold.

patscc
January 20th, 2009, 06:20 PM
Dumb question, you've tried another controller ?
patscc

Allen
January 20th, 2009, 06:43 PM
I suppose I should go into more detail on what I tried. First, there's nothing wrong with the ESDI controller or the PS/2 itself (I tried a good hard disk in this particular PS/2, and everything went fine). The system sometimes has trouble recognizing the type of fixed disk (hence error codes 1780 and 1790), but this isn't always consistent. The system tells me this when I try to view or run configuration with the reference disk.

Low-level formating with the reference disk gave too many sector errors. I've tried all possible options with the FDISK and FORMAT commands in DOS. Nothing seems to work. There are jumpers on these disks, so I tried closing a few. Sometimes I would get strange noises with some positions closed, others didn't seem to do anything (if any of you have a working WD-336R or WD-336RT, let me know if you have any particular jumper settings on yours).

I'm tired of dealing with these disk--I just want to be absolutely sure there is nothing else I can do. That's why I asked if "bad track 0" means a hard disk is dead, forever.

patscc
January 20th, 2009, 06:58 PM
There's a tool called Spinrite out there that does low-level formatting, above and beyond what some controller's low-level formatter do, but I've never used with an ESDI drive. (And also, unfortunately, don't have a copy anymore )

patscc

Terry Yager
January 20th, 2009, 07:55 PM
Best guess, 7th month, 14th day, 3rd year of manufacturing?

--T

Chuck(G)
January 20th, 2009, 08:35 PM
Yes, track 0 errors are bad news; that's where the master boot record is kept. I assume you're doing a low-level format, not the FORMAT command type.

It could be that the whole disk is bad. If I remember correctly, the format operation starts at cylinder 0--if it fails, the formatter calls it quits right there.

Allen
January 20th, 2009, 09:20 PM
I tried using Spinrite, but it couldn't do it's analysis (or anything else) because the disk wasn't formatted to DOS--and that is the problem, I cannot format any of the four disks.

I tried a low-level format using the PS/2 reference disk, but the program stopped because there were over 800 bad sectors, so I could not finish a low-level format. I could not start a normal format (using DOS command FORMAT) because of a bad track 0.

I appreciate the efforts in decoding the date of manufacture. I wish IBM made it simple to understand date codes.

Chuck(G)
January 20th, 2009, 10:15 PM
I tried using Spinrite, but it couldn't do it's analysis (or anything else) because the disk wasn't formatted to DOS--and that is the problem, I cannot format any of the four disks.

I tried a low-level format using the PS/2 reference disk, but the program stopped because there were over 800 bad sectors, so I could not finish a low-level format. I could not start a normal format (using DOS command FORMAT) because of a bad track 0.

I appreciate the efforts in decoding the date of manufacture. I wish IBM made it simple to understand date codes.

If the LL format failed from the reference disk, there's no point in following up with DOS FORMAT.

Out of curiosity, do the drives you're trying to format have the terminator resistor packs installed? Unlike floppy drives, which can get by without them, they're pretty important to ESDI functioning.

channelmaniac
January 21st, 2009, 07:03 AM
+1 on the terminator packs.

An alternative would be to download DR-DOS and use it to partition the drive. You can start the partition at a different cylinder than 0. I've had luck in using that with drives that reported a bad cyl 0.

But... with your drive failing the low level format it's either a terminator pack is missing or it's toast. As a last ditch effort to use the drive, you can try to resolder the connections on the cable and on the board where the heads plug in. Sometimes the solder cracks.

RJ

Allen
January 21st, 2009, 12:57 PM
I'm no electronics expert, so I am unfamiliar with terminator resistor packs. I tried looking for pictures on the internet for these, but nothing was clear. Attached are two pictures of the board of one of the ESDI drives. There is nothing interesting on the other side of the board. If these drives have terminator resistor packs, I assume it would be on the board.

channelmaniac
January 21st, 2009, 02:03 PM
Hmmm... Maybe active termination?

Check for a manual for the drive. It may have to do with the jumper settings.

RJ

Chuck(G)
January 21st, 2009, 02:08 PM
Can we get a close-up of the 5-position jumper block in the upper right corner of the drive PCB?

Thanks!

Allen
January 21st, 2009, 02:24 PM
A picture of the jumpers is attached. I could not find any documentation about this drive (WD-336R and WD-336RT) online, so I couldn't find any information about the jumpers. I did close two jumpers at a time (from top to bottom) and restarted the computer. At least three times, I got some real strange noises from the disk. But in all five tries, it didn't seem to change the status of the drive.

patscc
January 21st, 2009, 02:27 PM
Does that wide edge connector plug into a cable, or into a socket on a card ?
patscc

Allen
January 21st, 2009, 02:30 PM
The hard disk connector connects to a socket on a fixed disk adapter card connected to the system board (Model 50), but this depends on the PS/2 model. Some PS/2 models use a cable connected from the system board to the hard disk.

Terry Yager
January 21st, 2009, 03:00 PM
I dunno, man. Can't tell much from the pix, but when it comes to IBM and MCA, all bets are off.

--T

patscc
January 21st, 2009, 03:05 PM
Ok, my take on it is that the drive you have pics posted from plugs into a board, and therefore has no termination on it. The 'second stage', for want of a better word, controller probably provides any neccessary termination.
You're trying to use the drive with the plug-in type board, right ?
patscc

IBMMuseum
January 21st, 2009, 03:27 PM
Dumb question, you've tried another controller?

When referring to "DBA" (Direct Bus Attachment) ESDI, all of the controller electronics are on the drive itself. DBA ESDI is also a single drive system, no primaries, secondaries, or chains to worry about turning termination on or off. The following microchannel PS/2s are DBA ESDI either being present or a stock configuration (but can also have that supplanted by another interface like SCSI):

50Z (on riser card)
N51SX (smaller notebook-class dimensions)
N51SLC (ditto)
53SLC3 (on riser and cable, obsolete vestige, stock is planar IDE)
55SX (on riser and cable)
CL57SX (smaller notebook-class dimensions)
70 (on riser card)
P70 (from planar cable)
90 (obsolete vestige on 8590, planar connection x 2, connections removed on 9590)

patscc
January 21st, 2009, 03:37 PM
IBMMuseum, if I'm reading you right, that would mean that on Allen's problem termination is not an issue ?
patscc

Chuck(G)
January 21st, 2009, 03:49 PM
That explains why the drive in some lists is specified as "IBM MCA".

IBMMuseum
January 21st, 2009, 04:11 PM
...on Allen's problem termination is not an issue ?

Termination isn´t even an adjustable factor on DBA ESDI drives (¨ESDI¨ is a misnomer too). These 30Mb drives are very problematic nowadays (at over 20 years old, is it any wonder?). Much better to find the larger drive capacities (they came in 30, 40, 60, 80, 120, and 160Mb sizes) when you can´t switch to another controller type (read: SCSI) because of limited adapter slots.

Chuck(G)
January 21st, 2009, 04:14 PM
Calling them ESDI is indeed a red herring (http://ps-2.kev009.com:8081/ohlandl/misc/ESDI.html). See the last topic on this page.

So they're IBM-specific MFM/RLL drives. Look like a disaster.

patscc
January 21st, 2009, 04:51 PM
Chuck(G) said...IBM-specific MFM/RLL
Sounds like the marketing department trying to add some glamour to the word 'cheap'
patscc

IBMMuseum
January 21st, 2009, 05:22 PM
Calling them ESDI is indeed a red herring (http://ps-2.kev009.com:8081/ohlandl/misc/ESDI.html#Intergrated_HD_ESDI). See the last topic on this page...

Or change the tag like I did when I quoted you. The other sections of the page deal with the more standardized ESDI [EDIT: and MFM] adapter[s] found on the Model 60 and 80. There was supposed to be a way to query that controller (and I was going to try it on the DBA ESDI, to see if it would give a response) for its microcontroller version (because there was an ECA on the early ESDI controllers) much like "SCSILVL" (http://www.IBMMuseum.com/projects/SCSILEVL/).

Allen
January 22nd, 2009, 12:59 PM
I suppose I will just have to try to liquidate this second IBM PS/2 Model 50Z without a working hard disk. I really wanted to try to fix any of the four 30 MB drives I have, since everything else in this PS/2 works fine.

It sounds like bad track 0 problems are difficult or impossible to fix (and not something I'm really trained to do). The discussion about whether these are really ESDI or MFM makes me disappointed in IBM. IBM didn't make these PS/2s easy to handle.

I do thank all of you for your insights.

modem7
January 22nd, 2009, 02:03 PM
It sounds like bad track 0 problems are difficult or impossible to fix
I guess it depends on the drive.
In a 2006 thread, two forum members were able to get around a track 0 problem on their ST-412s.
The thread is http://www.vintage-computer.com/vcforum/showthread.php?t=3391 (from post #34 onwards).

Chuck(G)
January 22nd, 2009, 02:45 PM
I guess it depends on the drive.
In a 2006 thread, two forum members were able to get around a track 0 problem on their ST-412s.
The thread is http://www.vintage-computer.com/vcforum/showthread.php?t=3391 (from post #34 onwards).

The real problem is that the "track 0 bad" message doesn't contain a lot of information. All the message really tells you is that something is wrong very early on in the formatting process.

Some PS/2 BIOSes have LL format support built in under INT 13H, Function 1AH. It might be worth trying the INT 13H BIOS routine; e.g. AX=1A00, CL = 0, DL = 0080 and see what comes back in AH.

patscc
January 22nd, 2009, 03:04 PM
There's actually a call you can make to get extended ESDI error information, INT 13 serivce 1C08h
See here, http://lrs.uni-passau.de/support/doc/interrupt-57/RB-0617.HTM
if you want to start throwing stuff at it from debug.

patscc

IBMMuseum
January 22nd, 2009, 04:58 PM
Exactly, even though these drives were a kludge, they did give very good error codes to tell where they thought the problem was (and even saved it in a log many times, that could go through many power-cycles). As far as I went to diagnose many was swapping controller cards and platter assemblies, but you wouldn´t believe how many defunct parts I had to weed out. Just accept that the 30Mb versions are pretty crappy 20 years later, and the 80Mb+ types do much better.

IBMMuseum
January 22nd, 2009, 05:05 PM
...Some PS/2 BIOSes have LL format support built in under INT 13H, Function 1AH. It might be worth trying the INT 13H BIOS routine; e.g. AX=1A00, CL = 0, DL = 0080 and see what comes back in AH.

Or just do CTRL-A from the Reference Diskette...

IBMMuseum
January 22nd, 2009, 05:10 PM
http://lrs.uni-passau.de/support/doc/interrupt-57/RB-0617.HTM


The author needs to update a few calls, but yeah, I need to expand mine a great deal too: http://www.IBMMuseum.com/interrupts/INT15h/

Chuck(G)
January 22nd, 2009, 05:39 PM
The author needs to update a few calls, but yeah, I need to expand mine a great deal too: http://www.IBMMuseum.com/interrupts/INT15h/

The Phoenix books on the PS/2 CBIOS/ABOIS are very good for this sort of stuff and better than most online references that I've seen.

patscc
January 22nd, 2009, 05:49 PM
I was just wondering since these buggers seem to be more MFM than ESDI, how do they handle defect mapping ? Would there me a way to manually edit the defect table on these, or does the 'controller' (the thing the drive plugs into) store the defect table ?
patscc

Chuck(G)
January 22nd, 2009, 06:30 PM
The 1A format command allows the caller to specify an RBA table that can either be preset with values or updated as the format progresses.

What I don't know is if these units have the CBIOS ESDI support, since they're not really ESDI.

...or what an INT 13H, function 05h might do in any case.

patscc
January 22nd, 2009, 07:10 PM
Chuck(G) said...INT 13H, function 05h might do in any case
It sets ES:BX to "Your warranty has expired. For service, please ship to IBM along with the blood of your firstborn."

patscc

IBMMuseum
January 22nd, 2009, 08:23 PM
...What I don't know is if these units have the CBIOS ESDI support, since they're not really ESDI...

Once upon a time I was going to write a utility to determine the ESDI controller microcode version using one of those calls, and thus see how close the DBA ESDI was to the standard. But I haven´t even completed my version of ¨SCSILEVL¨, that does the same thing on IBM SCSI controllers. And now I have less time than before.

Allen
January 22nd, 2009, 10:05 PM
I did use Ctrl+A from the main menu of the reference disk to do a low-level format. I've done this many times with all four disks, and everytime the system stops formatting because it reaches a threshold of 800+ bad sectors. The bad sectors are probably a result of the inability of the system to read the fixed disk type (as it keeps telling me this). There doesn't appear to be an option to log errors during a low-level format.

When I test the fixed disk using the testing program in the reference disk, the program gives me error code 00175140 and tells me "error while reading fixed disk" and to "please replace integrated fixed disk C."

I have to admit I really don't know how to interpret or deal with "AX=1A00, CL = 0, DL = 0080 ..." I also don't have enough expertise and experience to physically fix these disks, if possible (as noted in other topics mentioned here). I will accept "that the 30Mb versions are pretty crappy 20 years later" as IBMMuseum says. I'm not going to fight 20 years.

Chuck(G)
January 23rd, 2009, 11:15 AM
IBMMuseum's probably right that this is beating a dead horse.

So much for the myth of the quality of construction of the PS/2 line, I guess.

Sorry that we couldn't help out more.

frozenfire75i
January 23rd, 2009, 12:29 PM
I would agree, the PS/2 line seems to be much cheaper then the PC line, I have run across lots of bad PS/2s and parts!



So much for the myth of the quality of construction of the PS/2 line, I guess.

Jorg
January 23rd, 2009, 12:48 PM
I tried using Spinrite, but it couldn't do it's analysis (or anything else) because the disk wasn't formatted to DOS--

What about Disk Manager?

channelmaniac
January 23rd, 2009, 03:07 PM
Resolder the connections on that plastic ribbon cable and on the drive's logic board that connects the 2 together.

If you have a cracked solder joint there you will get failed formats and other issues.

It's a 5 minute thing to do and I've used that to resurrect old drives in the past.

patscc
January 23rd, 2009, 03:13 PM
channelmaniac said...plastic ribbon cable and on the drive's logic board that connects the 2 together
I thought the drive's controller board plugged directly into the card ?
patscc

Chuck(G)
January 23rd, 2009, 04:05 PM
I thought the drive's controller board plugged directly into the card ?


Well, there is a short flex cable that connects the HDA with the drive PCB, but it's not soldered on the board end--it fits into a connector on the PCB.

patscc
January 23rd, 2009, 04:38 PM
Gotcha, I think. IDC-type cable with something that looks like a edge-card connector on one end ?
patscc

Allen
January 23rd, 2009, 05:48 PM
No one should apologize for not being helpful--you all have been a great help. I did try Disk Manager, and I got the same problem--bad this, bat that. In this case, too many cylinders were bad. The program tried to relocate the partition, but each time it would run into another bad cylinder. "Uncorrectable ECC [cylinder no.]" is the error I kept getting while trying to install the drive/create partition with Disk Manager. This program identified the hard disk as MFM/RLL, not ESDI.

For a picture of the fixed disk setup inside a Model 50, see http://john.ccac.rwth-aachen.de:8000/alf/ps2_50z. There seems to be a lot of questions about the connector from system to drive.

These disks are probably dead. They may even be physically bad, which would not be surprising after 20+ years.

IBMMuseum
January 23rd, 2009, 07:09 PM
...For a picture of the fixed disk setup inside a Model 50, see http://john.ccac.rwth-aachen.de:8000/alf/ps2_50z. There seems to be a lot of questions about the connector from system to drive...

Except that Arnold has some kind of unusual setup - the drive looks to be a DBA ESDI model, but the riser is different from anything else I have seen. It almost looks like the Model 50 MFM riser. Except for a couple BIOS chips on a riser version to run a DBA ESDI drive in a base Model 50 (no ´Z´), all of the other 50Z DBA ESDI risers (all riser diagramed at http://www.ibmmuseum.com/ohlandl/8550/8550_Controllers.html) have only circuit traces on the board (functionally equivalent to a ribbon cable).

IBMMuseum
January 26th, 2009, 08:47 AM
I would agree, the PS/2 line seems to be much cheaper then the PC line, I have run across lots of bad PS/2s and parts!

It really just depends on the PS/2 or the part. A few models, mostly early on, are known to be buggy in certain areas. Parts, notably a couple brands of 1.44Mb floppy drives, are known problems.

Later PS/2s are industrial-strength (and maybe a little too "new" to be on-topic here)...