Hard drive formatting

[Brufnus]

Hi there!

 

I've just bought a non-functional Seagate ST1100 83 MB MFM drive very cheap (about 12 USD), in the hope that I might somehow get it to work. If so, it will be the largest MFM drive I've ever had for the Geneve.

According to seller, it formatted fine but was unable to accept a partitioning afterwards (DOS/Win, apparently).

 

So, I've tried different settings and also done some things I shouldn't have - such as taking it apart in far from ideal working conditions, cleaning the outer edge of the upper platter with isopropyl alcohol, and even tried demagnetizing the platters and heads very carefully. Heck, if it doesn't work, there's no harm done in ruining it further... and at that price, I thought I'd give it a shot.

 

The details are pretty weird: I've been able to format 10-20 cylinders with only 1 or 2 heads, and then actually copying files to the drive and even execute them from it. When I tried all 9 heads, it gave me a track 0 error every time... with 8 heads, it was a little better, but still no joy.

 

After the cleaning and demagnetizing, however, I'm able to format the drive successfully with MDM5 with all 9 heads, not a single sector shows up as faulty, but as soon as the formatting is complete and I try DIR'ing or something, MDM tells me the drive is not formatted.

CFORM gave me a lot of CRC/ECC and SYNC errors at some point, but if MDM5 is able to format the entire drive with not a single sector error, then I assume the heads and platters have to be healthy... or?

 

Does any of you have any experience with such a drive, and/or such a weird behaviour?

It's no big deal if it doesn't work, especially considering the cost... but I'd love to have that drive, also because it's specs make it sound like a pretty fast drive (15 ms seek time, etc)

 

Edited September 27, 2021 by Brufnus

[+9640News]

OK, the Myarc HFDC Controller only supports 8 heads unless the card is modified.

 

Myself, I would use MDM5 for formatting the drive and do not use CFORM.  Do not enter anything more than 8 heads.

 

From the drem.info manual where a lot of development work on drive emulation, there is this note:

 

NOTE: Myarc HFDC controller supports drives with maximum 8 heads. This is because HFDC is using Pin 2 signal on control interface as RWC (reduced write current) instead of Head 2^8 select signal. HFDC may be modified to handle drives with up to 16 heads: - Cut the trace from Pin 5 of chip U9 to Pin 5 of chip U17 - Solder a jumper from Pin 12 of U9 to Pin 5 of U17 ATTENTION: AM26L32PC chip is buggy and can cause 'quirky' system problems. If you have one of these chips it is suggested to replace it with a AM26LS32AC.

 

Beery

[Brufnus]

Ha ha, okay... only 8 heads, I've never read or been told that, and I just assumed it would be in order as long as the drive didn't exceed 134 MB. Yes, I usually also use MDM5; in this case CFORM was for testing.

 

Anyway, the result is the same - when I format with only 8 heads, it formats fine with MDM5 and reports no sector errors, but it seems it fails to write bitmap etc. to track 0. I just don't understand how track 0 can be bad, if all sectors are okay... if so, I think it sounds more like a calibration issue...?

[Brufnus]

It's probably the weirdest MFM drive I've ever encountered. I wasn't able to format it with MDM5 on my "main" Geneve, thus the usage of CFORM. On the other one, it formatted but yesterday caused the controller to become unresponsive - I thought I had damaged it, since I couldn't boot from any hard drive, but then I got it working again.

Now I've just installed it on the "main" Geneve again, formatted it once again with CFORM - it works well until it reaches cylinder 680-690, then it throws quite a few CRC errors, all on the same platter, it would seem - the rest are error free.

 

Anyway, now I'm suddenly able to format it with MDM5 on that one... I wonder if there could be some resident old data on the platters, which are hard to get rid of. It didn't work very well when I entered a value for reduced current cylinder, so I retried with 0 for both that and write precomp (I kept the latter one in MDM5 now, however).

 

Could it be possible that the heads are simply getting a too low current for writing? In that case, I'd guess it's just a matter of replacing a bad cap or something like that... but then again, I'm far from an MFM hard drive expert, ha ha

 

In any case, it does seem to improve with every attempt I make. c",)

 

Oh, by the way... since there's a head # limit, is that true for cyl # as well - and, if so, what limit? Hopefully the HFDC can handle the 1072 cylinders the drive has...

Edited September 28, 2021 by Brufnus

[+9640News]

Your questions are not something I can answer myself.  Better left to other experts.

[Brufnus]

I think track 0 is bad, despite the fact it reports no bad sectors on that one. The bitmap is not updated now; so the disk name and available sectors (wrong count) remains the same, no matter what I do.

 

Perhaps I should just let it go, it's no big deal... but I was really hoping I could get that drive working again.

[Brufnus]

21 minutes ago, 9640News said:

Your questions are not something I can answer myself.  Better left to other experts.

Yes, that's okay. I'll see what I can figure out. c",)

[+mizapf]

3 hours ago, Brufnus said:

Anyway, now I'm suddenly able to format it with MDM5 on that one... I wonder if there could be some resident old data on the platters, which are hard to get rid of. It didn't work very well when I entered a value for reduced current cylinder, so I retried with 0 for both that and write precomp (I kept the latter one in MDM5 now, however).

 

Oh, by the way... since there's a head # limit, is that true for cyl # as well - and, if so, what limit? Hopefully the HFDC can handle the 1072 cylinders the drive has...

The HDC9234 (chip) on the HFDC can handle up to 2048 cylinders. Also, you could experiment with Reduced Write Current; the number is the cylinder where the setting gets activated.

[wierd_w]

Personally, I would try running it through an old PC with spinrite on a repeated testing cycle for several hours/days, just to make sure the magnetic domains on the platter have been good and exercised.

 

 

[+InsaneMultitasker]

Something to check is whether the hard drive is configured with terminating resistors, and that the termination is correct for the cables and drives.   Incorrect termination can cause some weird results - especially when there is one or more floppy drives connected at the same time as the hard drive(s).

 

Try formatting the drive again with 20-30 cylinders and fewer than 9 heads. Does it still work?   When I test/repair HFDCs, I usually format 50 cylinders with MDM5, and once the card is working successfully, I'll exercise the HFDC and drive at the maximum cylinder/head count for the test drive.  As stated earlier, unless your HFDC is modified it cannot support more than 8 heads; and if you do modify the card, this affects how Reduced Write Current operates. 

 

 

[Brufnus]

Sorry for wasting your time with this. It's so da** heart breaking what has happened... It didn't work initially which was the reason I got it so cheap.

After having it disassembled I actually managed to format it properly, but it simply refused to update the bitmap in the end, despite several attempts (showing some 806 sectors available from a test formatting with 1 and 2 heads and a few cylinders).

 

So, inspired by my initial success I took it apart again to clean the platters a bit more than I already had, and from there things got worse and now the drive is in pretty bad condition. It turns out it's possible to ruin the platters by tightening the screw at the cover which goes into the actuator magnet - I've learned some things but ruined the drive in the process. I'm surprised though, that I was lucky enough to hit the sweet spot on that screw during my first attempt.

Anyway, thanks for your suggestions anyway; I may try and repair it later on - I do have one advantage, namely the fact that one of the surfaces doesn't have a head, so the worst can be placed at the headless arm.

 

It's still a riddle however, why it can format with MDM5 and show only one bad sector (at the end of the drive), and still refuse to update the bitmap. Weird things happen... c".)

 

Oh, and by the way... yes, the termination can cause strange behaviour, but I've also experienced that the proper termination doesn't always work as it should. With certain drive combinations I need to leave the terminator pack installed on both drives in order to make them work. I even had problems once with formatting a Seagate ST-251; it simply refused to format... so I played around and did things I wasn't supposed to, since it didn't work anyway...  and eventually when I shorted two legs at the terminator pack, it started formatting! c".= lol It won't read the disk with these legs shorted, but works perfectly fine without. I do need to short them if I want to format it, though. Very strange hard drive! c".)

Edited October 7, 2021 by Brufnus

[+mizapf]

10 hours ago, Brufnus said:

So, inspired by my initial success I took it apart again to clean the platters a bit more than I already had, and from there things got worse and now the drive is in pretty bad condition.

Ouch ... The platter compartment is normally filled with some gas to keep dust away, so you should only open it when you want to put that drive on your shelf afterwards. The read/write head hovers some fractions of a millimeter over the platter, so dust can actually wedge between it.

[Brufnus]

Yes, I know... but it was my only shot at getting it to work. It wouldn't format properly until after I had it opened and cleaned the upper platter which had some debris on it. As soon as I reassembled it and tried again, it formatted properly, except for that bitmap update, as mentioned.

[Brufnus]

I've never heard they should contain gas, though. They do have a breathing hole in the cover, so I find that highly unlikely...?

[+mizapf]

In newer drives they use helium, as far as I know. Could be that earlier drives had just a pressure valve, but don't they have some dust filter behind? I remember to be warned never to open the cover, as even invisibly small dust particles may damage the drive.

[Brufnus]

Yes, I agree. However, my philosophy is that old 20-80 MB MFM drives with huge heads and low platter speed are less prone to damage from, say, dust, than a modern 12 TB drive. Usually I would never open up a drive, though, unless it's not working in the first place, of course. If that's the case, then I have nothing to lose by opening it (and I was actually quite close to having that Seagate working after doing it). My only regret now is that I opened it once more and apparently messed up the calibration when reassembling it, unlike a lucky first-time-disassembly. Perhaps I could've resolved the bitmap issue by formatting it several times, perhaps just track 0 to speed things up.

 

I'm in the same situation as before I purchased that faulty drive, albeit a few Euros and some time poorer, but I can live with that. I also have a number of working MFM drives at my disposal, and I'll probably use TI99-PC to backup all our club archives from 5 1/4" diskettes instead of storing it directly on the Geneve; that will save me a load of disk space.

What really breaks my heart though, is the fact that I was apparently SO close to get that drive working, and then I messed it up nonetheless... c".) haha

 

[+mizapf]

I have a ST-225 which I regularly show in my lectures, and at first it still did its calibration on spin-up, but I messed up something, and now it does not even turn the platters ... but well, impressive enough, I guess.

[Brufnus]

Could it be a bad cap? I'm far from an expert when it comes to electronics, but it seems these small monkeys are quite prone to quit doing their jobs properly and build an internal resistance, denying the motors etc. the juice they need to do THEIR job (I have a suspicion that's what has happened to my Amiga CDTV drive, so I'll probably replace the caps on that one during winter).

I've been trying to get a Tandon 262 working during the last couple of days, didn't spin up either - until I realized it didn't return the head stack to it's parking position when powered down. When I did that manually, the drive suddenly did spin up again... I haven't been able to format it, though.

But these Tandons are not the best design, IMHO.

 

My collection consists of a quite fast Rodime 43 MB drive, a 225, a 125, a 251-1, a clunky full height IBM, an Olivetti, a Miniscribe 8425 and a Kyocera RLL drive, which shouldn't work with the HFDC but does nonetheless (perhaps because it was never taught to read and therefore is ignorant about this fact). B-)

My favorites are the Rodime, the 125 and the Kyocera - quiet and relatively fast. The 225 does a good job, too.

 

Still - a WHT or SNUG SCSI controller would be such a nice upgrade.

Edited October 7, 2021 by Brufnus

[+9640News]

1 hour ago, Brufnus said:

Yes, I agree. However, my philosophy is that old 20-80 MB MFM drives with huge heads and low platter speed are less prone to damage from, say, dust, than a modern 12 TB drive. Usually I would never open up a drive, though, unless it's not working in the first place, of course. If that's the case, then I have nothing to lose by opening it (and I was actually quite close to having that Seagate working after doing it). My only regret now is that I opened it once more and apparently messed up the calibration when reassembling it, unlike a lucky first-time-disassembly. Perhaps I could've resolved the bitmap issue by formatting it several times, perhaps just track 0 to speed things up.

 

I'm in the same situation as before I purchased that faulty drive, albeit a few Euros and some time poorer, but I can live with that. I also have a number of working MFM drives at my disposal, and I'll probably use TI99-PC to backup all our club archives from 5 1/4" diskettes instead of storing it directly on the Geneve; that will save me a load of disk space.

What really breaks my heart though, is the fact that I was apparently SO close to get that drive working, and then I messed it up nonetheless... c".) haha

 

The best, and easiest way to backup, and be able to secure the backup, is probably with the TIPI setup.  It has unlimited capacity as far as the TI is concerned, and it easy to make a backup of the drive to another computer with relative ease.  With the MFM drive, your only backup options are to another drive and/or to floppies.  My "backups" of my HFDC drives go to a TIPI path and then I back everything on my TIPI to a Windows folder.

[Brufnus]

4 minutes ago, 9640News said:

The best, and easiest way to backup, and be able to secure the backup, is probably with the TIPI setup.  It has unlimited capacity as far as the TI is concerned, and it easy to make a backup of the drive to another computer with relative ease.  With the MFM drive, your only backup options are to another drive and/or to floppies.  My "backups" of my HFDC drives go to a TIPI path and then I back everything on my TIPI to a Windows folder.

Yes, problem is... I don't have a TIPI... ;-(

[+9640News]

1 hour ago, Brufnus said:

Yes, problem is... I don't have a TIPI... ;-(

Cha-Ching......

[Brufnus]

1 minute ago, 9640News said:

Cha-Ching......

[Brufnus]

By the way... is piping somehow possible with MDOS? I was thinking about creating a batch file, which reads the contents of my directories and then pipes it to SFBACKUP, which in turn saves the compressed output files to 1.44 MB floppies or a smaller hard drive. I don't know if that's possible, though...

[+9640News]

23 minutes ago, Brufnus said:

By the way... is piping somehow possible with MDOS? I was thinking about creating a batch file, which reads the contents of my directories and then pipes it to SFBACKUP, which in turn saves the compressed output files to 1.44 MB floppies or a smaller hard drive. I don't know if that's possible, though...

I think SFBACKUP is the progrm Al Beard wrote that embeds the path of the file into the compressed file for restoration.  Last time I used that program if memory serves me, was about 25 years ago.  I thought it ran some kind of front end program with one program calling the other as one program read the directory, and then the other program compressed the file.  If that is the case, then pipes I don't think is necessary and you can let the program run "as is".

 

What I don't is if your batch file changes to various directories, and then you execute the frontend program, if it uses the current path and subsequent subdirectories, or if it goes back to the devicename with no path.  I don't recall what the command line arguments are with the program, however if any line exceeds 80 characters, you will likely have an issue.

 

Another option to use the HFDC, which is more costly than backing to the PI/TIPI, is the use of a DREM emulating a MFM drive.  Then, you can just pull the SD card and make a backup of the image file that way. That is how I backup my hard drive images with the HFDC as I do not have physical MFM drives anymore, rather emulated MFM drives.

 

The other issue is with 1.44 MB floppies and questioning whether you can truly use 1.44 MB's, or if you should limit yourself to 720K floppies if you still want to go the floppy route for backup.

 

Beery

 

 

[Brufnus]

I guess what you're referring to, is the Backup Utility from LGMA, it sounds pretty spot on what you mention. I bought a copy back in the 90's; however I was foolish enough to NOT make a backup copy of that floppy in time, and now there are a couple of the files it simply won't read. So in practise, I only have the SFBACKUP standalone program at my disposal.

 

The 80 character limit I was not aware of. Yes, I know the 1.44 MB format has been questioned a bit; personally I haven't had any serious issues using that, but you're probably right that 720K would be a safer bet. I'm still thinking about what would be the best possible solution.. many years ago I had quite a few setbacks now and then, if the MFM drives for some reason weren't configured properly - more than once I've witnessed that ruining the contents of the drives and required a formatting. For some reason, I haven't (yet) experienced that during my last many disk swaps, reconfigurations, etc... but perhaps I've become more careful or wiser or something since then (or perhaps has MDOS improved on that situation as well).

 

Anyway, my first priority at the moment is to get a backup made of all our club archives from the 80's and 90's, which are stored on a quite large number of 5 1/4" floppies right now. It may not be needed, but I've always preferred to have too much rather than too little. I think that's a good principle when it comes to (old) computers. c".)

I found a couple of functional 5 1/4" drives upstairs some time ago, so one of them is now installed in a Pentium III PC with a floppy controller, which works fine with TI disk formats on both 5 1/4" and 3.5" drives (running Win98). So once I'm ready I'm going to make DSK images of a couple of hundred floppies; perhaps I'll discover something in between, which may be of use to the community today, who knows? I do have both a web and an FTP server available, so if I find anything worth sharing, I'll let you know.

Edited October 7, 2021 by Brufnus