PDA

View Full Version : ST-412 resurrection



romanon
April 30th, 2014, 08:33 PM
Hi, i bought this disk on ebay, yesterday i tried to run it, i connected it on my 286 test computer, original Xebec IBM controller. Disk on first time booted without any problems to german IBM PC dos. First problems begins after DIR command, disk wrote a reading error. So i booted with my ms-dos 3.30 floppy and launched norton disk doctor. He found some errors in FAT but he could not repair it, notice me to run Calibrat. So i tried Calibrat and he also found error in FAT but couldnt repair it, allegedly i must low level format disk. I restarted computer but disk couldnt boot, drive C was inaccessible and fdisk wrote "error reading disk". So i booted with my PC-DOS 2.0 floppy, launched debug, G=C800:5, computer not respondet, next try with dos 3.30, nothing. Ok, another attempt with software SSTORE, program identified disk parameters but couldnt low level format it, wrote "adress mark error" or something...My the last ace in the hole, was program Disk Managere. Program correctly identifyed the drive parameters, but after low level format request wrote this (attachments) Any idea for rescue this disk? Thanks

Chuck(G)
April 30th, 2014, 09:04 PM
You'll probably have to perform a low-level format to get the disk usable.

romanon
April 30th, 2014, 09:28 PM
if you read my post carefully, i wrote that i tried it..

Chuck(G)
April 30th, 2014, 10:31 PM
if you read my post carefully, i wrote that i tried it..

The problem is that it isn't clear to me that a low-level format was actually performed. The controller was an 8-bit XT-style controller? Have you tried an AT-style 16-bit controller? Note that the disk drive does not furnish drive parameters to the controller--they're obtained from the BIOS extension ROM on the controller.

archeocomp
April 30th, 2014, 10:34 PM
launched debug, G=C800:5, computer not respondet, next try with dos 3.30, nothing.

My guess is, Xebec conroller has no built in format utility at address C800:5. Try low level format with another controller. For example Longshine.

SpidersWeb
April 30th, 2014, 11:12 PM
My guess is, Xebec conroller has no built in format utility at address C800:5. Try low level format with another controller. For example Longshine.

If it's the original PC XT 8 bit controller then you're correct.
When I did mine I just used the PC XT Advanced Diagnostics diskette - SSTOR should do it ok though?

romanon
April 30th, 2014, 11:19 PM
SSTORE wrote "adress mark error" or something, i will try it with another controller.

Stone
May 1st, 2014, 02:53 AM
My guess is track zero (0) is unusable and that makes the drive impossible to LLF. I have a stack of FH drives with this exact problem and no remedy available.

modem7
May 1st, 2014, 03:32 AM
ST-412 resurrection ..., original Xebec IBM controller
That combination is known to work together.
Documentation, cabling diagrams, low-level format procedures, etc. are [here (http://www.minuszerodegrees.net/ibm_xebec/ibm_xebec.htm)].


original Xebec IBM controller
By "original", do you mean the first variation of the controller ("Variation #1" as shown at [here (http://www.minuszerodegrees.net/ibm_xebec/ibm_xebec.htm)]) ?
If not, then which variation do you have?
If it is variation #3, then for the ST-412, the switches need to be set for table address 0.


i connected it on my 286 test computer, original Xebec IBM controller
The Xebec-made-for-IBM controllers supplied in the IBM 5160 are intended for a 4.77 MHz XT-class computer computer.
I have never seen any Xebec documentation for those controllers, and so I don't know if they will run reliably (repeat: reliably) in the faster ISA bus of a 286.


Disk on first time booted without any problems to german IBM PC dos.
So, we can rule out a cabling problem.


SSTOR should do it ok though?
Yes. SpeedStor (SSTOR) is what I use for my Xebec-made-for-IBM controller and ST-412 combinations.


SSTORE wrote "adress mark error" or something
A low-level format of the ST-412 will take a few minutes, and during that time, you will see the stepper flag (http://www.minuszerodegrees.net/ibm_xebec/st412_cyl_0.jpg) (damper, interrupter) rotate.
The ST-412 is a very old drive, and there is possibly going to be areas of the platter surface that SpeedStor (or anything else) will not be able to low-level format.
Did you see "adress mark error" part way through the format, or did you see it at the very start ?

Stone
May 1st, 2014, 04:09 AM
The ST-412 is a very old drive, and there is possibly going to be areas of the platter surface that SpeedStor (or anything else) will not be able to low-level format.
Did you see "adress mark error" part way through the format, or did you see it at the very start ?Sounds like you're trying to determine if the bad part(s) of this drive are in the Track 0 neighborhood. So I'd like to point this question directly at you. I've asked it in general several times in other threads and am still searching for a solution, if there is one.

Specifically, is there anything left to try when a LLF cannot be successfully performed on an old, FH drive like this? As you may know (from my other posts) I have quite a few of them with this problem, i.e., LLF failure which is likely due to a bad Track 0. I have one ~ 70 MB drive with a 'flakey' Track 0 that I was somehow able to partition into a 2 MB and a ~ 68 MB partition. I ignored the first (2 MB) partition and made the second (68 MB) partition active and now I can use it successfully. Ironically, the first partition cannot be used but the partition table, itself, is valid. I've also tried the method of opening the platter housing and adjusting the arm so the heads land over another area for Track 0, etc., as described in another, older thread. I tried this on two drives without any success. So I'm hoping that you might have another solution that might do some good that I might try. My pile of FH doorstops is getting unmanageable and I'd like to try something else to see if there is a chance of bringing any of them to life, again.

romanon
May 1st, 2014, 07:46 AM
By "original", do you mean the first variation of the controller ("Variation #1" as shown at [here]) ?
If not, then which variation do you have?
If it is variation #3, then for the ST-412, the switches need to be set for table address 0.

I have Variant #2, with "IBM 1501492" label


A low-level format of the ST-412 will take a few minutes, and during that time, you will see the stepper flag (damper, interrupter) rotate.
The ST-412 is a very old drive, and there is possibly going to be areas of the platter surface that SpeedStor (or anything else) will not be able to low-level format.
Did you see "adress mark error" part way through the format, or did you see it at the very start ?

i saw it at the very start

Stone
May 1st, 2014, 08:13 AM
i saw it at the very startOnce again, it sounds like a Track zero issue and from my experience it's nearly impossible to work around that kind of drive problem.

fatwizard
May 1st, 2014, 11:35 AM
On an ST412 you can change the head position slightly by moving the interrupter on it's shaft. On my 412, track 0 was a problem and the interrupter would just bang against the stop. I loosened the set screws and shifted the interrupter arm slightly. Now the drive works reliably even though Spinrite complains about track 0 still. Track 0 issues many times cannot be resolved (not by me anyway). An ST412 would only need to be stored for years with the heads unparked to cause track 0 problems, or any drive that doesn't self park.

I have had success with LLF on the IBM Xebec controllers using the IBM diagnostic disk, but I have ofter had issues when using SSTOR. Many members use SSTOR, but I often have problems with it. Maybe I'm doing something wrong.

romanon
May 1st, 2014, 11:26 PM
ok i tried debug agin, but he dont start formating only wrote this at the very start

18549

im trying to analyze media surface in SSTOR but...wtf? All tables bad??

18550

modem7
May 1st, 2014, 11:55 PM
Sounds like you're trying to determine if the bad part(s) of this drive are in the Track 0 neighborhood. So I'd like to point this question directly at you. I've asked it in general several times in other threads and am still searching for a solution, if there is one.

Specifically, is there anything left to try when a LLF cannot be successfully performed on an old, FH drive like this? As you may know (from my other posts) I have quite a few of them with this problem, i.e., LLF failure which is likely due to a bad Track 0. I have one ~ 70 MB drive with a 'flakey' Track 0 that I was somehow able to partition into a 2 MB and a ~ 68 MB partition. I ignored the first (2 MB) partition and made the second (68 MB) partition active and now I can use it successfully. Ironically, the first partition cannot be used but the partition table, itself, is valid. I've also tried the method of opening the platter housing and adjusting the arm so the heads land over another area for Track 0, etc., as described in another, older thread. I tried this on two drives without any success. So I'm hoping that you might have another solution that might do some good that I might try. My pile of FH doorstops is getting unmanageable and I'd like to try something else to see if there is a chance of bringing any of them to life, again.
I think that an easy trap to fall into is the belief that a drive has a track 0 problem, when in fact, it is a different problem.

For example. I get my good ST-225, and remove resistor R15 from its logic board. My ST-225 can no longer write to any part of the platter surface. I run a low-level format program. The program starts at track 0. The controller goes through the actions required to put down, on track 0, the required pattern of ones and zeroes. But the controller and drive are oblivious to the fact that a fault is preventing actual ones and zeroes getting written to the platter surface.
The controller that I'm using is one that reads the low-level sector after it is written, and based on that result, passes back to the BIOS either a success or fail status. It fails to read the (non-existant) sector and passes a fail status to the BIOS, which in turn goes back to my low-level format program. My low-level format program, knowing that the first sector on the first track is critical to just about all operating systems, aborts the low-level format, displaying "Critical error on track 0".

It doesn't always work like that. Different low-level format code and different controller (chips) may behave differently, but my example shows how people can be led to believe that they have a track 0 problem.

So I think the first thing to do is prove whether or not the problem is a track 0 one.

Having a look at SpeedStor, I can see that it has bounded low-level format functionality, named 'boundInit'. I have never used it myself, but I can see that it would be of use to you. On your 'suspected track 0 problem' drives, use the boundInit to low-level format, say, cylinders from 5 to 10. If all cylinders fail, then clearly, the problem is not a track 0 problem.

modem7
May 2nd, 2014, 12:01 AM
I have had success with LLF on the IBM Xebec controllers using the IBM diagnostic disk, but I have ofter had issues when using SSTOR. Many members use SSTOR, but I often have problems with it. Maybe I'm doing something wrong.
On the SpeedStor page of my web site is:

"I usually tell people that version 6.03 be used for PCs and XTs, and that version 6.5 be used for ATs.
Why is that?
In the early 90s, I was using version 6.03 successfully. At some time, I received version 6.5 and consequently stopped using version 6.03
But then I noticed that in some XT clones, version 6.5 didn't work properly but version 6.03 did."

Were you using version 6.5 ?
Per my own rule, I use 6.03 on my IBM XT.

Now, I'm not saying that 6.5 will not work on an IBM XT, just that it may be a problem.

modem7
May 2nd, 2014, 12:09 AM
im trying to analyze media surface in SSTOR but...wtf? All tables bad??
What is interesting about that result is that only cylinders 0 to 18 are showing as bad, suggesting that the surface anaylsis is good for cylinders 19 to 305.

romanon
May 2nd, 2014, 01:31 AM
It doesnt look good...

18551

modem7
May 2nd, 2014, 02:30 AM
It doesnt look good...
At this point, you still do not know if the drive is the problem, or the controller is the problem, or perhaps a cable problem has developed.
As Chuck wrote, it would be good to try the drive on a different controller (which will probably require that the drive be lw-level formatted on that controller).

One thing that I see on my ST-412 from time to time is that the power connector needs to be reseated.

romanon
May 2nd, 2014, 03:15 AM
Finally success!
I replaced IBM controller with my Longshine controller, then LLF with debug, it works without problems, so i agin replaced Longshine with IBM (Xebec) controller and launched LLF, it WORKS! Then i create FAT table with fdisk, formated and transfered PC-DOS 2.0, all works fine, Norton disk doctor not foundet any errors and any bad sector! Thanks for help from all!

18552

Beast and and his pure performance ;)

18553

Al Kossow
May 2nd, 2014, 07:17 AM
Finally success!
I replaced IBM controller with my Longshine controller, then LLF with debug, it works without problems, so i agin replaced Longshine with IBM (Xebec) controller and launched LLF, it WORKS!


Interesting data point, that there are two MFM controllers that share sector formats. Which models of Xebec and Longshine controllers do you have? Could you take a picture of the Longshine
so that the hard disk interface chips are visible?

Shadow Lord
May 2nd, 2014, 08:03 AM
I am glad to see you got your drive working! I am amazed how often some of this vintage equipment which is supposedly not working works flawlessly after a bit of due diligence. A few months back I was about to write off my CMI drive and it turned out to be stupidity on my part. While I agree that everything fails, specially mechanical HDDs, I am constantly surprised by the resilience of this equipment!

Eudimorphodon
May 2nd, 2014, 08:33 AM
Beast and and his pure performance ;)


That disk should perform far better than a transfer rate of 25.4 Kb/s. What interleave did you format it at? The Xebec controller as configured by IBM used... 6:1, I believe, but could usually handle up to 5:1, and at the correct interleave it could do between 80-100Kb/s.

Backing way up... were you using this in an AT, again? Is there any reason why you really want to use the Xebec in an AT? An AT controller will perform better and be more compatible with the native BIOS. Semi-amusing side story: the first hard drive I ever set up was an ST-412/Xebec I was given as a hand-me-down. (We were poor at the time.) The best computer we had was a 12Mhz AT motherboard stuck in an XT clone case (and that was the only AT part in there, for a floppy controller and serial/parallel port I was using an XT multi-I/O card), and I had a *really* bad time trying to get the Xebec to play nice at the 8mhz bus speed. So I gave up and spent my pennies on a WD-1006, an AT hard drive controller that had a track buffer allowing you to use a 1:1 interleave. Formatted that way an ST-412 gives just shy of 500Kb/s, letting that humble old 10MB drive land somewhere around the 4.0 mark on that Norton benchmark. (Average seek time would still be dead slow but the overall performance was still pretty darn impressive.)

Chuck(G)
May 2nd, 2014, 08:47 AM
Interesting data point, that there are two MFM controllers that share sector formats. Which models of Xebec and Longshine controllers do you have? Could you take a picture of the Longshine
so that the hard disk interface chips are visible?

That's not what I get--in both cases, he mentioned low-level format (LLF) with both controllers, no?

Stone
May 2nd, 2014, 09:06 AM
So I gave up and spent my pennies on a WD-1006, an AT hard drive controller that had a track buffer allowing you to use a 1:1 interleave. Formatted that way an ST-412 gives just shy of 500Kb/s, letting that humble old 10MB drive land somewhere around the 4.0 mark on that Norton benchmark. (Average seek time would still be dead slow but the overall performance was still pretty darn impressive.)I have a NIB, NOS WD1006V-SR1 if anybody's interested in it. It's an RLL controller. I use one just like it to get 26 sectors or 50% more usable space out of my FH Maxtor tanks.

romanon
May 2nd, 2014, 11:18 AM
Interesting data point, that there are two MFM controllers that share sector formats. Which models of Xebec and Longshine controllers do you have? Could you take a picture of the Longshine
so that the hard disk interface chips are visible?

I will take picture of both controllers on monday ;)


That's not what I get--in both cases, he mentioned low-level format (LLF) with both controllers, no?

yes i did LLF twice, first time on LGS controller, then on Xebec, because after LLF on LGS didnt work on Xebec without LLF


Backing way up... were you using this in an AT, again? Is there any reason why you really want to use the Xebec in an AT? An AT controller will perform better and be more compatible with the native BIOS.

AT machine is only for test, in future, i want to build original IBM PC XT, with original components, so i need Xebec controller and ST-412 ;) (IBM PC is already constucted, so XT is on next plan)

romanon
May 4th, 2014, 10:57 PM
Mentioned photos of those two controllers..

18584 18585