need some help on a 8088

[modem7]

Thanks a lot for that, as soon as i have some time i will compare the content of my bios ship and your image file !

Your motherboard has a Phoenix BIOS, whereas what I pointed you to is the DTK/ERSO BIOS on my motherboard. The content will be quite different.

On the web page that I pointed to, there is what I think is an earlier version of your Phoenix BIOS, version 2.51. You have version 2.52

It is highly unlikly that the motherboard BIOS is the problem cause. And that is supported by the fact that the FileCard still did not work when you moved it to your IBM 5160, a system that I expect to be compatible with the FileCard (assumption: you first removed any existing hard disk controller).

for the moment i'm waiting for the replacement controller.

His only chance is to get data out of that old hard drive !
If there is any good software for this purpose please let me know ..

Because the WD1002-27X may be the faulty part, you are best to try the replacement WD1002-27X first. Set the jumpers the same as what is on the existing WD1002-27X. If no success, try swapping the ROMS between the two WD1002-27X cards (in case the ROMs contain different drive tables).

I presume that you have already disconnected the cables between the WD1002-27X and drive, and then reconnected them (to fix any poor electrical contact).

 

[g4ugm]

His only chance is to get data out of that old hard drive !

Even if he does get the data out it doesn't leave him very far forward. He is still using a machine for which parts will become harder and harder to source.

 

[modem7]

At [here] is DOS based software named SpeedStor.
It is not data recovery software. It is best described as hard disk initialisation and testing software.

With it, you can:
1. Get the controller card to perform a self-test, and display the result on-screen.
2. Get the controller card to step the drive heads to various cylinders.
3. Get the controller card to read sectors of data.

It can be useful in determining the amount of functionality that is working, something that can aid in diagnostis.

For your XT-class machines, use version 6.03


1. Get the controller card to perform a self-test, and display the result on-screen.

Select 'Diagnostics' then select 'Controller'.
Expect to see "Internal Controller test - PASSED" at screen bottom.
Even if you see that, the controller can still be faulty. It is a confidence test only.

2. Get the controller card to step the drive heads to various cyinders.

Select 'Diagnostics' then select 'SeekTest'.
You should hear the heads in the drive being stepped, and see the stepper arm/flag ([photo]) rotating accordingly.

3. Get the controller card to read sectors of data.

Select 'Diagnostics' then select 'ReadTest', then choose the desired tracks (track = the combination of a cylinder and a head).
SpeedStor will then attempt to read the sectors in the chosen tracks.

 

[evris80]

Western Digital 8bit MFM/RLL controllers have a build in format utility stored in their bios. Theese controllers supported custom drive parameters such as Cyls, Heads, WPcom, Step etc. As I remember, the WD controller stored that sequence of parameters on the first sector after the format utility ended ok, so that it can read theese parameters every time the PC powers up by probing the first sector. My WD1002-27X reports error 1701 if I connect a drive tha has not been initialized (LLF) with it's internal format utility. If sometimes you get the 1701 error and sometimes you don't, its probably due to deterioration of the drive's integrity (sometimes the controller manages to get data about the configuration from the first sector and sometimes not).

The reason you can't see drive C even if your system manages to start without 1701 err code, is maybe because this internal deterioration has affected crusial parts of the MS-DOS partition/file system organization (I assume your custom made sorting machine PC uses MS-DOS). So I think it's not a problem that has to do with your motherboard but with physical fatigue of the magnetic material on the surface of the disk due to age. You could try SSTOR as modem7 adviced to get a glimpse of how bat the situation is. I recall that some versions of spinrite could help recovering sectors reported as bad by reinitializing them without data loss but you can't be sure.

If you see no progress no matter what you try AND the data stored in this drive have a great value, then I would also recomend contacting a data recovery company.

 

[pearce_jj]

Just a thought, it might be worth changing the power supply, if the 12v rail is drooping enough it could be preventing the drive from spinning up in time or at all.

As an aside, if you can get the data off the drive I would gladly supply a bootable CompactFlash base storage solution for the machine, it's interesting to see such an old machine still in use!

 

[techtatan]

hello every body

sorry for delay, but i'm still busy at work ! also I'm still waiting to recieve the new WD controller card from Stone.

Just a thought, it might be worth changing the power supply, if the 12v rail is drooping enough it could be preventing the drive from spinning up in time or at all.

As an aside, if you can get the data off the drive I would gladly supply a bootable CompactFlash base storage solution for the machine, it's interesting to see such an old machine still in use!

I already replaced the power supply but no luck ... I would really apreciate some help to get it all on a CompactFlash card Thanks in advance

Western Digital 8bit MFM/RLL controllers have a build in format utility stored in their bios. Theese controllers supported custom drive parameters such as Cyls, Heads, WPcom, Step etc. As I remember, the WD controller stored that sequence of parameters on the first sector after the format utility ended ok, so that it can read theese parameters every time the PC powers up by probing the first sector. My WD1002-27X reports error 1701 if I connect a drive tha has not been initialized (LLF) with it's internal format utility. If sometimes you get the 1701 error and sometimes you don't, its probably due to deterioration of the drive's integrity (sometimes the controller manages to get data about the configuration from the first sector and sometimes not).

This is probably the case since the owner told me that the boot error has began since last year and he just restarted the machine a few times to get it work.
Since last month when he contact me he never was able to get it boot !

The reason you can't see drive C even if your system manages to start without 1701 err code, is maybe because this internal deterioration has affected crusial parts of the MS-DOS partition/file system organization (I assume your custom made sorting machine PC uses MS-DOS). So I think it's not a problem that has to do with your motherboard but with physical fatigue of the magnetic material on the surface of the disk due to age. You could try SSTOR as modem7 adviced to get a glimpse of how bat the situation is. I recall that some versions of spinrite could help recovering sectors reported as bad by reinitializing them without data loss but you can't be sure.

As far as i have understood the purpose of spinrite it needs to read MSDOS partition to work. It reads and write back data several times to get it more persistant. and reallocate bad sectors. I my sens it can not work without the ms dos partition.
I would try SSTOR as soon as i have some time.

If someone knows a software to get databack without MSDOS partition please let me know

Thank you all for your help, i come back with test results as soon as possible !

Kind regards

 

[SomeGuy]

As far as i have understood the purpose of spinrite it needs to read MSDOS partition to work. It reads and write back data several times to get it more persistant. and reallocate bad sectors. I my sens it can not work without the ms dos partition.
I would try SSTOR as soon as i have some time.

If someone knows a software to get databack without MSDOS partition please let me know

Thank you all for your help, i come back with test results as soon as possible !

Kind regards

I believe that is correct.

However, you really don't want to run Spinrite until you have retrieved the data. If track Zero is physically damaged, there is the risk that its low-level attempts could make it go from marginally readable to completely unreadable. It can't move the MBR, boot sector, the FAT, or the root directory. (Although ocasionally you can get lucky and suddenly everything becomes fully readable)

The Norton Utilities Disk Edtor can open "physical" drives that are unformatted, non-dos, or corrupt data-wise. Futhermore, you can select all sectors as object for editing and then write that object to a file on a secondary larger hard drive, a file on a network share (won't span floppies), or directly to an identically sized secondary hard drive. There are certainly others that can do similar.

BUT the catch is that the physical drive must be visible to BIOS (and if it still formatted then C: would automatically be visible to DOS anyway).

If bad/weak sectors happen to be in the FAT, or some other place that DOS needs to constantly re-read while copying files, then making a physical sector dump like that could be faster.

 

[Stone]

Are you sure that this is a DOS partition on the hard drive? I know that you have never seen it operate. What does your neighbor say about this actually being a DOS partition? A lot of these old systems used in factories and manufacturing operations were proprietary and did not use DOS as an OS.

 

[evris80]

I believe that is correct.

However, you really don't want to run Spinrite until you have retrieved the data. If track Zero is physically damaged, there is the risk that its low-level attempts could make it go from marginally readable to completely unreadable. It can't move the MBR, boot sector, the FAT, or the root directory. (Although ocasionally you can get lucky and suddenly everything becomes fully readable)

I think SpinRite version 6 can scan and attemp repairs on defective sectors at low level, without having an MS-DOS or other partition on the media (unfortunatelly I have no idea if this version works on pre-286 machines). I also agree that any attempt to try a repair using such tools could easily make an already bad situation worse as SomeGuy says. It could happen even if you go with a simple non-destructive media analysis test which doesn't make any repair attempts, only just because you stressed your drive some more.

 

[mR_Slug]

Sorry if this has been mentioned before, but have you tried leaving the system on for like an hour or 2, and then tried to reboot the system? Some old stepper motor drives, write tracks in different places depending on the expansion of the HD platter. It may work. What was the model no of the actual drive?

Before you start disassembling the drive i would at least get a quote from an Hdd recovery company. However if that's not an option i do recall reading about someone who actually swapped the heads from an identical drive. But i think your along way from that long shot. Can you hear it seek when you try any of the above programs?

 

[pearce_jj]

Where do MFM controllers write the bad sector map? Maybe we need to adjust the bios to skip initialisation and the grab a flat file export of the entire surface for analysis.

 

[evris80]

Where do MFM controllers write the bad sector map?

On some MFM/RLL controllers you could define a number of cylinders as spares for relocating (hiding) manufacturer's declared (defect list) and/or potentially new bad sectors from the useful space. This procedure is transparent and totally controller depended and you can enable it if your controller supports it just before you LLF your drive. The drive's capacity shrinks statically accordingly to how many cylinders you have defined as spares (at least that is the case with my KONAN KDC-230 MFM controller).

But I don't think that is the case with WD1002-27X, unless that disk-card has some sort of a special bios other than usual. In most MFM/RLL controllers, their built in bios accepts typing the entries of manufacturer's defect list as it is printed on a label on almost every MFM/RLL drive, and then performs LLF to initialize the media and make it accessible for R/W operations. This is totally controller specific. While formatting, the controller just skips every defect previously declared to speed up the process. If you then partition and high-level format the partitioned space, these defects will come up as lost allocation units (bad sectors). DOS then avoids using them by cross-referencing the FAT entries. This means that even though you declared the defect list just before LLF, these defects will show up as bad sectors when DOS-formatting occurs.

Maybe we need to adjust the bios to skip initialisation and the grab a flat file export of the entire surface for analysis.

AFAIK on XT machines the hd related BIOS is on the disk controller and I don't think there is a way to override it's startup self-test. After all it is a WD controller, it has to issue a read sector command to get the disk geometry stored at the first sector so that it can address the drive properly. I think the use of a partition cloning software to clone a drive sector-by-sector without stopping when a read sector fail occurs will be of use here, to salvage any data before any repair attempt.

 

[pearce_jj]

Yes I agree, eg pop the combo in a '386 and use dd to create an image file with some older version of linux. I was thinking though that the controller might need to be tweaked by changing the BIOS to get it to access whatever is left of the drive. The ROM chip could be image, pulled(/disabled) then modified and popped on a flash ROM board (eg this).

 

[techtatan]

hello every body

I think the use of a partition cloning software to clone a drive sector-by-sector without stopping when a read sector fail occurs will be of use here, to salvage any data before any repair attempt.

The problem is that i do not own a more recent computer that supports the western digital filecard i have tested it on a K6 boad tha has an isa slot but the computer hangs at bios startup screen (may some bad irq interaction with the graphic card)

Yes I agree, eg pop the combo in a '386 and use dd to create an image file with some older version of linux. I was thinking though that the controller might need to be tweaked by changing the BIOS to get it to access whatever is left of the drive.

This is a good idea but as i said above the card does not start in a more recent machine. also i'm not aware of a linux version that supports 8088 hardware, but i'm may wrong.

But i know that freedos has DD feature but it needs at least a 1.44 floppy. I never tried to plug a 1.44 floppy drive on an MFM controller could this work ?
As a 320 floppy drive works on a normal dikette port on "recent" computers (at least my K6).

if you can get the data off the drive I would gladly supply a bootable CompactFlash base storage solution for the machine, it's interesting to see such an old machine still in use!

With the new ideas about dd and freedos i need a destination drive where to store image file. Does your proposal still stands ?

The ROM chip could be image, pulled(/disabled) then modified and popped on a flash ROM board (eg this).

This is an interessting solution but i was not able to find the list of components they use as they just sell the blanck pcb as far as i understood... I will contact the site admin/owner .

I should recieve stone's controller card soon and test the disk with that one. But one way or another i will need a compacflash based solution that seems mutch more reliable.

Thanks for all this help and good ideas
Kind regards

 

[evris80]

The problem is that i do not own a more recent computer that supports the western digital filecard i have tested it on a K6 boad tha has an isa slot but the computer hangs at bios startup screen (may some bad irq interaction with the graphic card)

I've never tried to attach an XT disk controller to such a 'modern' motherboard, I don't know if the 8bit bus portion of it's full 16bit ISA slot is somehow compatible. I've successfully used a WD1006V-MM2 and a WD1006V-SR2 AT controller on a Pentium II board by disabling the onboard floppy and IDE controllers. If you want to try again anyway, I suggest that you disable the onboard IDE controller (at least the primary one) and specify NONE disk at the K6's BIOS hard disk table, to eliminate the possibility that an onboard disk controller and the WD1002-27X try to share the same I/O resource simultaneously. I'm also not aware of an AT WD RLL 2,7 controller to be compatible with WD1002-27X format for trying another read-disk approach.

Yes I agree, eg pop the combo in a '386 and use dd to create an image file with some older version of linux.

Yes indeed, something like that. But I can't think of a way for techtatan to bypass the 1701 error other than a risky mechanical/environmental change like aligning the drive horizontally/freeze (risky, maybe as a last effort if everything else fail?). I don't know if the XT controller can access the drive properly later from OS if it starts with 1701 err code, but I think is possible. SpeedStor v6.03 that modem7 mentioned has a selection of Re-Initialize sectors without wiping off their data, maybe you could try that for a few sectors at the beginning, hopefully without stressing your drive a lot. I hope that Stone's replacement will do the trick and save your time!

 

[SomeGuy]

I don't know if the XT controller can access the drive properly later from OS if it starts with 1701 err code

The only time I (vaguely) recall seeing that happen was when the "1701" error was caused by the controller incorrectly looking for a second hard drive.

I would expect that if the controller could not get the geometry information from the drive, then it would be impossible for BIOS to access the drive.

Now, if one were to attach an identical good drive during boot, and then switch to the bad drive, (only applicable to MFM/RLL drives) I would expect it to provide access to any undamaged areas - *IF* the damaged hard drive is entering the "ready" state at spin-up.

From my own experience, damaged MFM/RLL hard drives will fail to enter the "ready" state if the drive's logic board is damaged, or the heads can not seek. That makes it completely impossible to read data from the drive, no matter what the controller card does.

 

[g4ugm]

I think its time to get this to a data recovery expert. Messing with the drive is likely to destroy any remaining data, especially if there has been a head crash and there are bits of oxide floating round inside the drive. These get stuck under the head and remove more oxide. The effect is then like an avalanche as more is removed there gets to be more oxide snow in the drive and so more oxide is removed. A good data recovery shop can open the drive in a clean room and check for this before attempting recovery.

A 1701 is almost certainly a drive error, either its not spinning up to speed or the head isn't unlocking.

 

[pearce_jj]

Do you think anywhere has the hardware to effect recovery of MFM drives now? (I don't know the answer to that btw, it's not meant sarcastically!)

 

[Stone]

All it really requires is a fully functional drive of the same model and a clean room.

 

[mR_Slug]

The problem is that i do not own a more recent computer that supports the western digital filecard i have tested it on a K6 boad tha has an isa slot but the computer hangs at bios startup screen (may some bad irq interaction with the graphic card)

Did you disable the bios (jumper w3) on the WD1002-27X when you put it in the k6 board? The WD1002-27X is a dual XT/AT card that should work in at systems, although the K6 system is likely to be more problematic than say a 386.

If you disable the bios on the WD1002-27X, you will need to configure the settings for the HDD in the motherboard bios for most utilities to be able to do anything. Try using one of the previous utilities mentioned to "access" the drive. Not too familiar with them, i used to use a program called FXPREP.EXE (i think, long time ago). You could enter the drive parms then do a LLF, but you could also open the drive in a hex editor. I think Norton can do something similar. If you can get this far, then imaging the drive, should be possible.

Now I'm not entirely sure that accessing the drive like this is going to produce meaningful output, as the correct procedure probably requires the drive to be low level formatted by the AT bios. But you should hopefully be able to see there is something on the drive. The heads should be able to read what there even if its garbled.

BTW if you do manage to access the bios on the WD1002-27X it is a bit gung-ho, if your used to later stuff be careful. The manual is here: www.coleskingdom.com/files/WD1002.TXT. Link is dead for me , but if you past it into google, you can read it in the cache.