• Please review our updated Terms and Rules here

I used an RLL controller on an MFM drive

cr1901

Veteran Member
Joined
Dec 28, 2011
Messages
817
Location
NJ
For those who haven't been following my recent posts, I recently got a Compaq Portable I. It's been quite a bit more work to upkeep than the other vintage machines I own. However, some issues I've been having are completely my own damn fault.

One of the gems of this past night was trying to install a hard drive into the portable using one of my two remaining Hard drive controllers. After screwing up the twist of the MFM cable multiple times (this gets it's own thread!), I successfully got my drive (Tandon TM-502) recognized by my hard drive controller (Seagate ST11R)... only to get the good ol' "Drive Parameter Activity Failed (0x07)" in both the Seagate BIOS and SpeedStor formatter. Furthermore when this Tandon drive was paired with this controller, the actuator carriage assembly would never move at all when I finally got the twist in the cable correct. Then I attached a Seagate ST-4038 to this controller, and while formatting still failed, the hard drive would respond to park, seek, and even respond to a (botched) media analysis in SpeedStor which suggested all sectors were bad.

On my second RLL controller (WD 1002-27x), no matter which drive was attached, and regardless of the twist in the cable, a 1701 error was triggered. I believe this means controller failure, but I've seen this occur even when the controller is just fine (known to work). I can't help but wonder- is it possible that the WD 1002-27x was able to sense drive parameters and detect that an MFM drive was attached? And subsequently choked and died with 1701? I'll have to try again...

And why did the programmable Seagate drive fail when it supports 17 sectors per track? My only educated guess is based on something vwestlife said:
MFM and RLL drives are mechanically identical. RLL drives were simply tested and certified by the manufacturer to meet a higher data reliability standard than MFM drives, because RLL squeezes 26 sectors per track onto the disk while MFM uses only 17 sectors per track, thus RLL increases the data density being stored on the magnetic surface of the platters.

The Seagate ST-11R controller card was designed for RLL, but you can tell it to format the drive with only 17 sectors per track. Whether that actually switches it into MFM mode, or if it's just only formatting the first 17 out of RLL's 26 sectors per track and leaving the rest blank, I have no idea. Maybe a program like SpinRite could do a surface scan of the disk and tell you how the sectors are laid out on the tracks.

I'm thinking the latter is the case. Before tonight, I thought an MFM drive would tolerate an RLL format, but there was no guarantee data would be written reliably thereafter. My WD 1002-27x controller also supports 20MB MFM drives... I'll try that next.

Why such wildly varying behavior between drives (drive attempts seek vs drive doesn't respond), and controllers (1701 consistent on one controller)? Weren't both drives being sent the SAME commands barring the drive geometry parameters? And I don't know this for sure, but I thought the drive geometry was stored on disk during low level format... shouldn't have the Seagate controller detected this condition and choke with 1701 like the WD controller? Or is that 1780 and I'm confused.

And now that I've verified that the drives are actually functional in my PC/AT, I'm wondering- was the AT ever come stock with an RLL controller and BIOS support for it?

Sorry for all the questions people XD... I'll try to break it up better when I have a moment.
 
I've got a Seagate ST-238R that is running fine with a WD1002-27X. If you wanna get together for the ultimate test we can swap your controller for the one that's in the machine now and then there won't be any guesswork.
 
After screwing up the twist of the MFM cable multiple times (this gets it's own thread!), I successfully got my drive (Tandon TM-502) recognized by my hard drive controller
For a single drive, set as DS0 I believe there is no twist in any of the cables. For a secondary drive, if you set it to DS1 there is also no twist.
 
RLL uses an identical interface to "MFM" drives, but it is higher density. The disk surface of many MFM drives can not accommodate the higher density, but once in a while some can.

By coincidence, I also have a Tandon TM502 10MB MFM drive. I have had it formatted using an 8-bit WD 1004-27X, and it works perfectly formatted to 16MB. I also have an IMI/Corvus 5MB hard drive using the same model of controller formatted fine to about 8MB.

From my experience, Seagate MFM drives like the ST-225, and many others will not work. What happens in these cases, if you use the BIOS formatter that does not verify, it will look like it formatted perfectly. But a disk tester will reveal increasingly weak/faulty sectors on the inner tracks. Usually about half to a third of the disk will be unusable.

1701 is a very generic error that basically means the controller can't "talk" to the drive. This failure can be in the drive, drive select settings, cables, or controller. (For me it is usually the darn cables!)

Once formatted you will absolutely want to run Spinrite, because weak sectors usually don't show up immediately.

Hmm, I've always used flat cables with these drives.
 
(2,7) RLL encoding employs the same maximum flux-transition rate, (frequency) as MFM and FM, but requires that there be less ambiguity in the exact time that transitions occur. Plated-media drives generally are less "fuzzy" in this respect than those that use thicker oxide coatings, but that does not mean that a coated-media drive cannot muster the performance required for (2,7) RLL. That is, the actual pulses on an RLL-encoded drive are no more dense than MFM or FM, but that their timing is more critical. In my own experience, I've found very few ST412 drives that would not admit to some success using a (2,7) RLL controller.
 
I've found very few ST412 drives that would not admit to some success using a (2,7) RLL controller.

Have you observed differences in write pre-compensation between RLL and MFM controllers? I haven't dug around to compare RLL and MFM preamps, did
they do some tweeks to the read channel on the drives between the RLL and MFM versions of the drives, or did they just change to plated media?
 
I can't say either way on that count, Al. I do know that some manufacturers separated the non-RLL-capable drives from the "certified for RLL" during the QA process, but whether it was on the basis of numbers of bad sectors or read channel jitter, I couldn't say.
 
I've got a Seagate ST-238R that is running fine with a WD1002-27X. If you wanna get together for the ultimate test we can swap your controller for the one that's in the machine now and then there won't be any guesswork.

Yea, it's about that time of year that I get together with you, isn't it? I'll PM you when I have a free block of time (probably not for a few days).


In any case, I found the manuals for both controllers:
ftp://ftp.seagate.com/techsuppt/controllers/st11m-r.txt
http://www.coleskingdom.com/files/WD1002.TXT

BOTH of them go out of their way to say to NOT use an MFM drive with an RLL controller.

The internal BIOS in the WD controller decided to behave today and I attempted an LLFormat; the controller BIOS returned an error code 20, which indicates internal controller failure. So now it's quite probable that the controller IS in fact dead/beyond repair.

As for the Seagate, I have no clue why none of my drives would work properly with that controller, considering that ChuckG stated he got the second PC hard drive ever created (ST-412 before RLL was standardized?) to accept a LLFormat. As a little fun fact, I recently found out that the "4" in ST-412 means "thin film platters", where the "5" in ST-506 means metal oxide platters... assuming the TM-502 used metal oxide platters- I don't know for sure/need to find the manual- could that make a difference?
 
In any case, I found the manuals for both controllers:
ftp://ftp.seagate.com/techsuppt/controllers/st11m-r.txt
http://www.coleskingdom.com/files/WD1002.TXT

BOTH of them go out of their way to say to NOT use an MFM drive with an RLL controller.
Well, that is SOP for MFM/RLL drives, isn't it? But as Chuck has attested to and I have to agree with from my own experience ... 20 to 25 years ago when MFM and RLL were common and IDE was up-and-comming I regularly used MFM drives with WD RLL controllers with great if not 100‰ success. I used the WD1006V-SR1/SR2 (which I had many of and still have some NIB) on all the Maxtor Full Height tanks I had, e.g., XT-1140 and XT-2190 to get that extra 50‰ or so of space out of them. And every one of them worked flawlessly year after year in that configuration. So, not being in a corporate position of having to back up or stand behind my words for any profitable enterprise or assume any liability whatsoever I will resoundingly refute that statement! :)
 
I did a low level format with a WD1002A-WX1 (different model, same brand/vintage) the other day and got an error 20 - it was just having trouble controlling the drive (controller was actually fine - just confused). So I wouldn't sign off on the controller just yet. Are you using the same cabling on this setup, as you were when you tested the drives?

Does your WD controller have jumpers? If so, what drive type is it set to? Did it's BIOS ask "Are you dynamically configuring the drive?"?
 
I did a low level format with a WD1002A-WX1 (different model, same brand/vintage) the other day and got an error 20 - it was just having trouble controlling the drive (controller was actually fine - just confused). So I wouldn't sign off on the controller just yet. Are you using the same cabling on this setup, as you were when you tested the drives?

Does your WD controller have jumpers? If so, what drive type is it set to? Did it's BIOS ask "Are you dynamically configuring the drive?"?
Yes, don't remember the drive type, and yes it does ask.

You know, I'm asking this on a hunch, but- is it possible for an AT controller's BIOS to send commands to an older-model hard drive that it simply doesn't understand? Now that I think of it, I've only performed a low level format successfully on either the PC, XT, or AT (in various combinations of SpeedStor, the controller BIOS and the AT BIOS), and NEVER on a clone...

My Tandon TM-502 is not seeking at all (stepper motor is not moving) on my 486 (no IDE or onboard controller otherwise), and that drive was just working a few days ago in my AT. And SpeedStor claims a Seek Test succeeds anyway! And yet, I successfully formatted it a few days ago on my AT...
 
Forget it, I just got to accept it- every single one of my spare MFM drives (all 4 of them) are dead. None of them seek (well, except 1 which can't read the voice coil track any longer). Granted, they all got their use at some point (namely the Tandon), but I suppose this is a lesson for me... to test them out once every couple of months just to make sure they still work, instead of keeping them on a shelf until I need them... at least I have the extra space now.

I do not wish to buy an XT-IDE... until they are exceedingly rare, I prefer to use the real MFM drives from the era.
 
You know, I'm asking this on a hunch, but- is it possible for an AT controller's BIOS ...
XT class computer: BIOS support for HDD resides in the HDD controller (no support in motherboard BIOS)
AT class computer: BIOS support for HDD resides in the motherboard BIOS

In an AT class computer, you have the option of using an AT class controller that has it's own BIOS (e.g. for SCSI drives). In such a scenario, the hard drive type in the AT's motherboard SETUP is usually set to zero, so that the motherboard BIOS does not think that it has a hard drive that it needs to control.

So, maybe your "AT controller's BIOS" should have been "AT motherboards's BIOS".

to send commands to an older-model hard drive that it simply doesn't understand?
'MFM' drives do not receive 'commands' in the same sense that 'IDE' drives do. In the 'MFM' world, most of the intelligence is in the controller. The 'MFM' drive is very basic, and so are the controls/signals used between controller and drive. Diagram [here].

Head stepping (for 'MFM' drives with a stepper motor) is explained [here].
 
Forget it, I just got to accept it- every single one of my spare MFM drives (all 4 of them) are dead.
Over the years, I've seen many 'MFM' drives fail, and it was rare to encounter one of those that would not at least seek. You have four drives that were previously known working, and now all them no longer seek. It sounds to me like 'something else is going on'.
 
Over the years, I've seen many 'MFM' drives fail, and it was rare to encounter one of those that would not at least seek. You have four drives that were previously known working, and now all them no longer seek. It sounds to me like 'something else is going on'.
Agreed, seems suspicious. I'd at least double check those. I think the only time I may have seen that symptom would have been on a drive with a completely dead logic board. And it seems odd multiple ones would fail the same way.
 
As a last resort, I tried the Tandon in my AT, which has the original 1984 version 1 BIOS and WD controller (off of which ATA is based, so I'm told)... formatted perfectly with no bad sectors, and transferred DOS 3.3 to it successfully.

Okay, I'm thoroughly confused now... WHY did it start working correctly all of a sudden, when it wouldn't even seek the night before?! o.0;

Testing my Microscience HH-825 will have to wait until I replace a bad trace on the PCB.
 
So... anyone have ANY ideas on this one? I'm still thoroughly confused... the drive went from completely dead and "no seeking- not even back to Track 0 on power up!" to working perfectly.

The only thing I can think of at this point is "dodgy power supply"... it's the only factor that remained constant throughout all these tests before today. But that's just weird to me.

Note: As some of you might be aware from previous threads, I have a 5150 motherboard and an original 63 watt power supply that I use for testing code I write. In this case, I was not using that power supply to power up the Tandon, but rather a spare AT-class supply that can deliver about 250-300W. I'm considering trying it with the 63W supply with a half-height floppy, thinking that the supply may be able to handle it.
 
So... anyone have ANY ideas on this one? I'm still thoroughly confused... the drive went from completely dead and "no seeking- not even back to Track 0 on power up!" to working perfectly.
The only thing I can think of at this point is "dodgy power supply"... it's the only factor that remained constant throughout all these tests before today. But that's just weird to me.
Some drives (perhaps most) will not move the heads if the spindle is not up-to-speed (i.e. not safe because only at up-to-speed are the heads considered to be flying above the platter surface).
If the +12V to the drive is inadequate/unstable, it affects spindle speed.
Check the power connectors on the drive (both plug and socket). Poor connections there are not unknown.
 
Some drives (perhaps most) will not move the heads if the spindle is not up-to-speed (i.e. not safe because only at up-to-speed are the heads considered to be flying above the platter surface).
If the +12V to the drive is inadequate/unstable, it affects spindle speed.
Check the power connectors on the drive (both plug and socket). Poor connections there are not unknown.

I don't have an oscilloscope to check whether it's unstable- can I use the AC component of my multimeter to check? If so, is there a rule of thumb for how small the AC component of the input power should be (+/- 5%)?

My other concern is that in order to check, I have to run the drive upside-down (PCB-side up)... not exactly a great idea, so I recall reading from old HD manuals.
 
I don't have an oscilloscope to check whether it's unstable- can I use the AC component of my multimeter to check?
No. That could be used to check AC ripple, but it won't reveal (as an example) the line sitting at 11.5V then 200 miliseconds later dropping to 9V and then 100 milliseconds later up to 12V, etc.

But is investigation needed? You wrote, "to working perfectly." So from my understanding, the TM-502 wasn't working and then it was, and you are looking for a possible cause. If the TM-502 is working now, then the +12V must now be adequate/stable and consequently, any 12V measurement pointless.

Or is it the case that the TM-502 always works in your IBM AT, and always fails to seek in your 486 based clone ?
 
Back
Top