I was reading that the ST-506 requires slower pulses (3ms) by the controller. This makes me wonder: what controller would've been used back in the day with this drive? I understand that it never came from IBM, as the ST-412 was the first one. The IBM Xebec controller also doesn't support it.
- VCF South West - June 14 - 16, Davidson-Gundy Alumni Center at University of Texas at Dallas
- VCF West - Aug 2 - 3, Computer History Museum, Mountain View, CA
- VCF Midwest - Sept 7 - 8 2024, Schaumburg, IL
- VCF SoCal - Mid February 2025, Location TBD, Southern CA
- VCF East - April 2025, Infoage Museum, Wall NJ
-
Please review our updated Terms and Rules here
Period-correct controller for a Seagate ST-506
- Thread starter sergioag
- Start date
famicomaster2
Experienced Member
The ST-506 does not support buffered seek and must have controller-generated write current reduction at cylinder 128 for correct operation. Many drives in their old age suffer from media degradation and will often benefit from earlier write precompensation or write current reduction, I suggest starting at 128 and moving to 64, then 0 if all else fails.
The ST-506 shows up before IBM even released the 5150, let alone gave it support for BIOS extensions. It would have been very unusual to see a hard drive at all in such a machine, and even the early Davong setups used 10MB mechanisms, I believe provided by Disctron. Most XT controllers expect a minimum number of 306 cylinders.
It's not "Period correct" 1981/1982 hardware, but the Western Digital WD XT-GEN with the Super BIOS is always a favorite card for running ST-506 drives, as it supports unbuffered seek mode and write current reduction.
There are plenty of controllers which can generate this signal and have unbuffered seek, my personal favorite being the Zobex 4HDC, one of very few quad drive controllers. It supports both RWC and unbuffered seek, as well as the CNGCAR signal used by the early removable cartridge drives, like the DMA 360 or SyQuest SQ-306RD.
Something I am on the lookout for is an AT controller which supports unbuffered seek (You can tie the RWC pin to ground on the drive and operate it at reduced write current at all times). Unfortunately, I'm not personally aware of any card which can do this!
For the most part, the ST-506 itself would have seen most of it's application in either a translated scenario (running through a SASI controller like the SMS Omti 3100 or Adaptec ACB-4000) or with a dedicated control card, like the Convergent Technologies AWS workstations and many industrial controllers. It was not until the much, much faster ST-412 came around that the interface and drives became popular in the PC world.
The ST-506 shows up before IBM even released the 5150, let alone gave it support for BIOS extensions. It would have been very unusual to see a hard drive at all in such a machine, and even the early Davong setups used 10MB mechanisms, I believe provided by Disctron. Most XT controllers expect a minimum number of 306 cylinders.
It's not "Period correct" 1981/1982 hardware, but the Western Digital WD XT-GEN with the Super BIOS is always a favorite card for running ST-506 drives, as it supports unbuffered seek mode and write current reduction.
There are plenty of controllers which can generate this signal and have unbuffered seek, my personal favorite being the Zobex 4HDC, one of very few quad drive controllers. It supports both RWC and unbuffered seek, as well as the CNGCAR signal used by the early removable cartridge drives, like the DMA 360 or SyQuest SQ-306RD.
Something I am on the lookout for is an AT controller which supports unbuffered seek (You can tie the RWC pin to ground on the drive and operate it at reduced write current at all times). Unfortunately, I'm not personally aware of any card which can do this!
For the most part, the ST-506 itself would have seen most of it's application in either a translated scenario (running through a SASI controller like the SMS Omti 3100 or Adaptec ACB-4000) or with a dedicated control card, like the Convergent Technologies AWS workstations and many industrial controllers. It was not until the much, much faster ST-412 came around that the interface and drives became popular in the PC world.
Chuck(G)
25k Member
The IBM Xebec controller does support the slower unbuffered step rate of the ST506, but not with the stock ROMs. I bought my Lovely Wife an XT clone that came with the enticement that a hard disk would be included free gratis with the purchase. Said disk turned out to be a 506. I was determined to make the combination work, so I did some patching to the ROM and got it working. She used it until a larger drive turned up. Still have the 506 drive.
Mine still works with the manufacturer parameters, which is quite astonishing to be honest.
That may explain why I had so much trouble making it work initially :/
Good point, I'll try to get a WD XT-GEN. I have other cards from WD (WD1002-WX1 and WD1002S-WX2), but they didn't work with this drive. The SuperBIOS is nice though.
Even back in the day RWC was a controversial topic, where some sources suggest it doesn't make any difference except if you want the absolute reliability, which in our current use cases probably we don't care anyway.
This makes me think that my question wasn't correct to begin with, as this drive predates the 5150, thus there can't be a "period-correct" controller.
Yes, that's the route I ended up going. It works reliably with it. I'm using the version 1 controller though. I have no idea if it works with the later versions. In case you find it useful, I took the source listings from the IBM manual, translated them to be useful with NASM, corrected a few mistakes (compared to binary files) and published them in Github, which you can find here.
Dwight Elvey
Veteran Member
I may be dumb but what is a buffered seek? Is that the auto step?
I use a ST-506 on my NC4000 setup. I used an XT controller. I don't know which one. Since there is no software for the NC4000 to run a hard drive, I wrote my own and it has been working fine. 5 megs is more than I'll ever used.
Dwight
I use a ST-506 on my NC4000 setup. I used an XT controller. I don't know which one. Since there is no software for the NC4000 to run a hard drive, I wrote my own and it has been working fine. 5 megs is more than I'll ever used.
Dwight
It basically means the controller can issue a seek command even if the drive is not ready yet. On a ST-506 drive, this would fail, because the drive can not buffer the seek command - the controller would have to wait a certain amount of time between commands instead.
Dwight Elvey
Veteran Member
Thanks, that makes sense. I guess I always wait for the drive to be ready before asking it to do anything else.
No seek before its time.
Dwight
famicomaster2
Experienced Member
Not quite what it means. Buffered seek means the controller can command the drive to move, say, 40 positions inwards by just sending 40 pulses as fast as it can. The ST-412 sees it's being asked to seek, starts moving the heads, and just counts how many pulses came in to know when to stop moving the heads.
The ST-506 does not have this, all it can do is move one position at a time. To move 40 positions, the controller needs to send one pulse, wait until the drive says the seek was complete, send another pulse, etc etc 40 times. The drive is "dumb".
Dwight Elvey
Veteran Member
I should mention that the correct period controller for the ST-506 is actually the WD1000. The only issue is that it would require an interface board to connect to an IBM PC. You also need to run a really old version of MSDOS as, at one point, they stopped supporting anything smaller than a 10Meg disk.
I just thought I'd mention those minor details.
Dwight
I just thought I'd mention those minor details.
Dwight
famicomaster2
Experienced Member
I'm not sure this is true, as I've had plenty of good luck working with my ST-506 and even damaged drives which have become smaller via DOS 6.22.
Same here. Both my ST-506 and ST-412 have worked correctly with MS-DOS up to 6.22, both booting from it or accessing the drives after booting a floppy. However, These days I normally run MS-DOS 4.0.
That's a very good clue. After your tip, I've looked at it and indeed it looks as if it could work with an adapter card which could contain also an appropriate BIOS. I will add it to my project queue
famicomaster2
Experienced Member
I like to use my ruined Tandon TM-252 as a benchmark. 4 heads, but only one surface is good. 2.5MB and DOS has no trouble with it.
Here's how I think of it, why would DOS not be able to access such a small drive? It can still use 360K floppies as boot media.
Dwight Elvey
Veteran Member
The disk controller used for the TRS80 5Meg drive is essentially a WD1000. It does require some data bus buffers but I used one of those, with minor changes to be a WD1000 for my Olivetti M20 project.
Real WD1000s are really hard to find.
Dwight
Real WD1000s are really hard to find.
Dwight
famicomaster2
Experienced Member
Really! That's very interesting, what mechanism is in the TRS80 modules? I see 5, 10, 15MB units appear from time to time and I always thought about buying one to find out what actual drive was inside. Just an ST-506 for the 5MB unit, or is it the infinitely more interesting ST-406 single platter drive? I imagine the 10 and 15MB units are ST-412 and ST-419, respectively.
andy
Experienced Member
famicomaster2
Experienced Member
Wow, an SA-1004? Now I have to go buy one.