• Please review our updated Terms and Rules here

Recovering data from MFM drive if paired controller is dead

Zare

Zare

Experienced Member
Joined
Mar 12, 2015
Messages
383
Location
Croatia
Is this possible? National Computer NDC5127 controller and a Seagate drive. The controller is hard to come by.

Don't need to run this drive in a PC, just extract data off it, any kind of hardware dumping tools apply too.
 
Depends. If the controller is 8-bit, chances are low, since each controller did things differently (CHS hard-coded, stored on track 0 etc.). If it is 16-bit, you may have luck by using e.g. a WD1006. At that time, they pretty much all worked the same - using the CHS values stored by the BIOS in CMOS.

Can you show a picture of the controller card?

If the data is important (either to you or in general), try to get a replacement controller of the same type, even if it takes time. Trying with any other controller can in the worst case destroy the data beyond restorability.
 
It's 8 bit.


The option rom is loaded and disk is seen but on boot the machines BIOS reports illegal interrupt. The jumpers on the card set the CHS and judging by one Youtube video the value is set correctly for the drive, e.g. not been tampered with.

Is there chance of this controller not doing an ondisk format since it's jumped to specs?

I have another same XT machine, same result. P5AB board I have does not boot with this card in. I have earlier chipset Pentium in which I can try but I'm not holding my breath.

The disk does a loud click on poweron and performs into booting. I have no guarantees that this disk is correct but I would like to peek at its data as it may contain the OEM software for the machine, which I lack.
 
Every MFM/RLL controller low-level formats the drive, which is one of the main issues with compatibility. The fact that it has dip switches for the CHS values means you can't really use any other controller to read the data from that drive. Most 8-bit ones either had fixed CHS values (which are unlikely to match by coincidence) or wrote the CHS values to track 0 of the drive itself. Trying to use a controller that does the latter would damage the data on the drive, as setting up such a controller to work with the drive would wipe track 0.

So really your best chance to read the data is finding an identical controller card.
 
Identical might be an issue but similar cards are available here and then.
 
That would be best solution for me but it's quite quite pricey. $240 without shipping insurance to Europe and since one would not risk such money, it's $280 insured.

257 euros is not cheap for one-shot case where I don't even know what's on the drive or does it work correctly, but merely suspect what's on it.
 
I don't, and I don't have anything else to try it against.

The illegal interrupt is just in the boot phase. It might be that ondisk format cannot be read by the controller.
 
Yes it did. But if the disk got damaged that would maybe explain what is happening.
 
I'm going to double check that. Btw. it's not a divide error, it's an "illegal interrupt" banner that's printed from BIOS just prior to boot strap.

Is a loud click on power on typical for Seagate drives?
 
Timo W. said:
Depends. If the controller is 8-bit, chances are low, since each controller did things differently (CHS hard-coded, stored on track 0 etc.). If it is 16-bit, you may have luck by using e.g. a WD1006. At that time, they pretty much all worked the same - using the CHS values stored by the BIOS in CMOS.
Click to expand...
Am I right in thinking that controllers that don't have any jumpers or BIOS configs for the drives (e.g. WDXT-GEN) would write the CHS information to track zero?

If a drive, LLF'd with a jumpered controller, were attempted to be read with a different (but compatible) controller without jumpers or BIOS config, it would fail to be read because such a controller would look for the information on track zero?
 
It would fail because unless it's a very rare case the two controllers would always be different firmware and part revisions, which would in their own ways alter the unique way each controller reads and writes to a drive. It might position the head on the right tracks but the interleave might be hard-coded different, for example.

The MFemulator has unfortunately been the only device of its kind in the last ten years to exist and provide both emulation of an MFM drive and to act as a tool to capture the raw data off any MFM-style drive so you can reconstruct it later. The price in part because it requires a rather specialty CPU has never really gone down in price. The actual NEED in the community for such a device has also been so low (in a PC especially most people switched to devices like the XT-IDE if their original drive died) it wasn't practical for them to enter mass production and bring the price down in volume of sales.
 
Ah so the controller pairing is actually about ensuring the controller is the same model, revision, and firmware version of the card that originally formatted the drive not that it's literally the same card?
 
I call it unintended copy protection. Back in those days the cost for an intelligent drive was so high and it was so rare to swap drives between machines that it was just assumed the average person would never run into this problem.
 
Vin Digit said:
Ah so the controller pairing is actually about ensuring the controller is the same model, revision, and firmware version of the card that originally formatted the drive not that it's literally the same card?
Click to expand...
Correct. Same revision etc are interchangeable. Seen a couple cards where different firmware versions are incompatible but frequently doesn't matter. Some brands different similar models are interchangeable. No list exists to document what is compatible. If things work a lot of people don't report anything so I don't have that much data on what is compatible. I get the doesn't work which then leads to another new format in the tool.

Track 0 is a frequent place for controllers to take to store format data but people are way to creative so other methods have been used.

The MFemulator guy.
 
NeXT said:
The actual NEED in the community for such a device has also been so low (in a PC especially most people switched to devices like the XT-IDE if their original drive died) it wasn't practical for them to enter mass production and bring the price down in volume of sales.
Click to expand...

True - I don't need it any more, the software I wanted to search for has been made available online in the meantime.

The controller is not cheap on euro ebay, it goes for over 80euro. I wonder if that price is unrealistic due to being 'rare' and 'very vintage'
 
You must log in or register to reply here.
Back
Top