Apple Hard Disk 20 repair: A success story

[BadGoldEagle]

Hey guys/gals

I bought a week ago sight unseen an Apple HD20 for 28€ shipped. I figured, what the heck, if it doesn't work at least it'll look good as a prop. A lot of these don't work anymore, and sure enough, when it finally showed up, it was DOA. 

I looked online and found this video by Adrian Black... I thought, why not give it a try? Plus, if you look down in the comments, you'll notice that Adrian wasn't the only one to bring one of those stepper motor drives back to life.... 16mmDJ managed to revive his 20mb miniscribe inside his SE using the same method.

I ordered some Hetman oil (Lubricant 11, Light Rotor) a week ago and it arrived today. It's actually made for brass instruments, but apparently it can be used to revive old hard drives as well.

I set about disassembling this thing. It's actually a lot easier to take apart than the 20SC. I decided to remove and clean every part inside since quite a few dust bunnies had made their home inside the casing. Both the inside and the outside were really filthy. I don't know where or how this was stored but I actually thought I wouldn't be able to restore this cosmetically, but it actually turned out great. It looks new.

Anyway, back to the hard drive itself. The stepper motor's shaft is located underneath the drive, so you definitely need to remove the four screws that secure the drive to the chassis first. If you want to know how to take it apart properly, then I suggest taking a look at the HD20 service manual available here. 

I added a couple of drops of lube and lo and behold... nothing happened. I gave a couple of knocks on each side. Still wouldn't fire up. I tried powering it on and off a couple of times and the PCB started getting really hot! But there was no time for electrical or mechanical sympathy. After a while I heard a click coming from the drive. That was good news. I powered it off and gave it a few more knocks. It went back to sleep.

At that point, anger was flowing through my system. I thought I had spent 15 bucks on some lube that made no difference whatsoever and that I would never use again (I'm no trumpeter). So I gave it a really good wack while it was still powered on. And it sprung back into life. It sounded terrible at first but after it warmed up a little, it sounded a lot better. There still wasn't any head movement though. But the motor had sucked up all the lube I had previously applied on its shaft. So I added a couple more drops, and after yet another restart, I saw the heads move a little. It was then I knew I would be able to bring it back. :rambo:

About 10 minutes after that, the drive was as good as new. It now spins up and down like in the olden days. And better yet, it mounts right on my performa's desktop! So now, I have a perfectly working and mint looking HD20. I'm the happiest man on earth right now. I actually mended an old hard drive. I had previously tried to revive an old sony but that ended catastrophically...

Here's a terrible picture of my partially disassembled HD20:



I was in the process of backing it up before formatting it... Notice the use of the Floppy EMU in HD20 mode connected to the HD20... HD20-ception!

So, TL;DR

Have an old HD20 with a Rodime HD or an SE with a Miniscribe? Give the drive a good wack and the stepper motor some lubricant and it could spring back into life!

PS: I was quite surprised to hear that the Performa 200 (Classic II) had built in HD20 support. I really wonder now why Apple decided to add that back to the ROMs after they deliberately removed it from the SE/30...

 

[Juror22]

Congrats on getting it running!  Nice tip about the oil.

...in answer to the ROM question, I'm guessing that it was because of the extra space in the ROM on the ClassicII?

 

[Dog Cow]

I have an HD20 mechanism that spins up and seeks, but it won't mount on the desktop. When I run the NishaDiag program, I can get it to read sectors off the disk using the Diag menu and I get recognizable data, such as the HFS catalog, or parts of programs and documents. But it's intermittent, and can sometimes take 4 or 5 read commands to get data back that's not garbage.

However when I use the System Read, I get a Comm error. Also get errors when I try to Format, Zero, or otherwise write to the disk.

When I run the other HD20 Test program and press Command-D to get the advanced interface, I get a Self Test failed error.

Would love to know what to look at next: is it a controller board issue or a drive issue?

 

[just.in.time]

That’s awesome, good job!

 

[BadGoldEagle]

Thanks!

...in answer to the ROM question, I'm guessing that it was because of the extra space in the ROM on the ClassicII?

I guess you're right... Perhaps people complained to Apple at the time that they couldn't use their old HD20 with their then new SE/30? Imagine someone who bought an HD20 with a 512Ke in '86, who then upgraded to an SE in '89 before upgrading it again to an SE/30 a year later... A lot of HD20s were still in use in 1990, and I can understand the frustration of those who spent $$$$ on an SE/30 upgrade kit and had to get rid of their HD20.

Dog cow, that looks to me as if the heads can't reach some parts of the disk... Even if the shaft isn't completely stuck, if it cannot get to its optimum speed, it won't be able to travel as far as it should go. The drive's own controller board will tell the stepper motor to move by say 20° which corresponds to applying that amount of voltage for that amount of time. If the shaft is slightly stuck, it won't move by 20° but perhaps 15 and that will can cause serious read/write errors. Mine would also fail the HD20 advanced test before I re-lubed the stepper motor's shaft. That's what I recommend you to do. I had to import mine over from Germany, that's why it costed me 'so much' (actually it's not THAT much but still...), maybe you can pick one up for real cheap?

I'm going to try and revive my Hyperdrive now. Its MiniScribe also has a stepper motor and I suspect it might be the cause of my problem.

 

[PB145B]

Very nice! I revived a Miniscribe 3650 (MFM) 5.25” hard drive by oiling the stepper. Before I oiled the stepper it would only get partially through the seek test and the the LED would flash an error code, but when I powered it on after oiling it passed the seek test perfectly and I was able to format the drive! Speaking of Sony hard drives, those things are VERY prone to stiction. I’ve got a 40MB one that locks up if it sits for more than a week or so. What exactly was wrong with your one?

 

[BadGoldEagle]

That's exactly what happened with my miniscribe 20mb MFM drive! I suspect a couple of drops of oil will bring the 512K Hyperdrive back to life! More on that tomorrow.

I am starting to get quite a sizeable collection now and over the last 5 years, I've only seen two SCSI hard drives go bad. Both of them are Sonys. Strangely enough, all my Quantums work flawlessly. One Sony suffers from god awful friction. I opened it up and spun it by hand a couple of times. I fired it up with the lid off (and that was when everything went wrong), the platters got covered in some oily glue that definitely killed the drive. The other sony has bad blocks everywhere and it ran out of sectors...

 

[Dog Cow]

Thanks!

Dog cow, that looks to me as if the heads can't reach some parts of the disk... Even if the shaft isn't completely stuck, if it cannot get to its optimum speed, it won't be able to travel as far as it should go. 

Maybe. I took the top off the drive and watched it operate. The heads will park and seek across all 315 cylinders, from the edge of the platter into the spindle.

It may be that they're not landing at the correct track, so your suggestions are something I will look into when I have time to work on this mechanism some more.

 

[olePigeon]

Oooo, nice.  I need to get one of these.

 

[CC_333]

Last I knew, my HD20 still worked. Which reminds me, I should check it....

If it doesn't spin up, I now know of something I can try to revive it!

Do you know if 3-in-1 light motor oil would work? It's all I have....

c

 

[BadGoldEagle]

All I know is that Triflow is a good substitute but any light motor oil should work.

My Hyperdrive is having Analog board blues right now, so I can't test my new improved MFM drive... Shame.

 

[CC_333]

It's probably bad caps....

You'll get there!

c

 

[BadGoldEagle]

It's currently in a flup flup flup phase... The display was cutting out intermittently and now everything is pitch black.

I had a look at Larry Pina's The Dead Mac Scrolls, and I checked CR20 but my multimeter is showing some strange stuff, I probably need to retire it, it's from 1987 after all... Just to be extra sure, I'll order a new one as well as caps for this mac. GCC modified the A/B a bit so the caps aren't the traditional ones. I need to list them one by one...

While we're on this subject, I might as well ask you guys about this: I noticed when I opened it up before plugging it on that I had forgotten to screw the ground wire back on. But I remember using this machine a month ago so the wire wasn't on back then either. Could that be the source of my problem?

 

[CC_333]

I think I accidentally forgot to put the ground back once. As I recall, the machine worked, but I'd hear a zapping noise when I turned it off (CRT discharge not making it to ground, perhaps?).

c

 

[BadGoldEagle]

Was that the black and white ground wire or the green and yellow one? I'm talking about the latter...

 

[techknight]

I tried this on a couple MiniScribe SE drives about a year or so ago, I had one where the stepper was completely frozen. Lubing fixed it along with LLFing the drive, but there is a bad sector right where the stepper had frozen. Sadly, that sector is on track 0 so when the SE tries to boot, itll hunt around for a minute or two before it boots. 

2nd miniscribe drive, didnt do squat. So I left it be. 

a few years ago I had to do this on my IBM 5155 floppy drives, had to tear out the hubs and drive bearings and redo them all. 

 

[Dog Cow]

Mission Report:

I applied lube to the platter motor and the head stepper motor. This had no useful effect other than to make operation of the drive quieter. After reading the Nisha documentation some more, I came across this gem (paraphrased):

"If the Nisha does not pass its self-test, all use of System commands will return a Communication error and only the diagnostic commands will work."

That seems to be exactly my problem, as Scavenger Mac reports that self test failed on the drive, and only the diagnostic read/write commands will work. Using HD Diag, I am able to find the boot block (LK), volume header (BD), and HFS catalog, as well as other data that I can recognize within the System/Finder files, and so on.

I also seeked way out to cylinder 291, head 2, and wrote an entire track (diag write_track) then wrote my name in sector 0. After power cycling the drive, I can read that back OK. Interesting thing is that if I seek to neighboring tracks, then I get a read error!

Final interesting point: I cannot find any spare tables! The Nisha documentation says they're at sector 0, head 0 of every 8th track. Well when I seek to the first 3 or 4 of these, I get a sector full of zeros. So this may be why the self test is failing!

I need to read the Nisha documentation some more and find out what the self test checks, so maybe I can determine why it's failing.

 

[techknight]

Does nisha have a built in LLF command? it may need that to re-prep the drive. The issue your having could simply just be magnetic fade over the years. 

 

[Dog Cow]

Does nisha have a built in LLF command? it may need that to re-prep the drive. The issue your having could simply just be magnetic fade over the years. 

HD Diag has two commands, one is Zero disk and the other is Init Mac directory. I don't know exactly how these two commands are implemented, but I think they use the System read/write commands so they won't work as long as self-test fails. I've already tried these commands, and they're not working for me.

As far as magnetic fade.... maybe. I'd like to try to recover the data from this drive first before I wipe it.

Reference for others: This is the the Nisha firmware documentation that I'm referring to. It is also the documentation for the HD Diag program, as that program implements all commands. Nisha Operation Summary, December 1984 is also important.