• Please review our updated Terms and Rules here

Safe to open up an mfm hdd to try and fix it?

Twospruces

Veteran Member
Joined
Dec 9, 2009
Messages
779
Location
Canada
Forgive the dumb question especially if this is asked and answered.

I'm contemplating opening up a kyocera kc20b mfm drive. If I do, is it basically bound to fail due to contamination?

Thanks
 
Do you have access to a proper clean room? If not, the drive is very unlikely to work after being opened.

Why are you considering opening the drives up?
 
Get a cheap large clear bin from walmart, cut 2 holes on one side and duct tape on some rubber gloves, like dish washing gloves. And in the rear, or top lid, throw a 2" opening for a vaccum cleaner/shop vac hose. Viola, you have a tiny clean room for less then 10 bux.

Anywho, if you are just trying to dump what is on the drive, it will work long enough to do the job. Just don't expect it to be a long lasting drive afterward if you dont use a cleanroom.
 
If the drive is not working right now, you won't make it any more bad by opening it.

These drive are not as delicate as modern ones. It is possible to open and fix them without having a clean-room. But you need to be very careful. I would suggest wearing lint-free gloves and using something like this:


to remove any dust before re-installing the lid.
 
To think these Kyocera drives were around $1300 back in the day (1990's)
Not sure what you're hoping to gain from opening the drive. Most likely immediate failure if you mess with the platters.
Unlike more current drives, opening these MFMs isn't a death knell right off the bat, but really not much to accomplish in doing so unless you're maybe replacing the stepper or head to hopefully get it working again enough to take data off it. There's quite a few videos on YouTube showing how to replace the motor or heads on these older drives, etc. The proper tools would be important - see here (https://hddsurgery.com/)

If you want to see the path your adventure would most likely take though, this video always cracks me up for some reason

 
Thanks for all the comments.

I have one kc-20b with a stuck stepper which never becomes ready for obvious reasons, and a second drive which has a working stepper but also never becomes ready. Out of curiosity I am thinking about trying to identify what night be wrong and trying to fix one or both.

So really just for the challenge of debugging these old drives.
 
The stepper is likely on something on the control board. There isn't much active electronics inside the spinning part of these drives.
 
I had a very dead MFM drive (the kind of dead that involves screeching noises as the heads crash into the platters) so took it apart just for fun. Do not under any circumstances touch the platters with your fingers. The beautiful shiny surface is a magnet for fingerprints!
 
The stepper is likely on something on the control board. There isn't much active electronics inside the spinning part of these drives.
Since I have 2 identical drives, I can swap the controllers. No difference.

For the strange acting stepper, I did buzz out thr coils. Shortage in coils where expected. I didnt check for a shunt to ground however.

So, not the electronics. Maybe the stepper.

I also realized that the axles spin very freely when the drive is depowered. So not seized at all.

So... why do these drives not become READY...
 
I guess I was thinking maybe there would be something mechanical wrong .. like a sensor, which might need a repair, under the hood.

Since I have two identical drives and the swapping of the electronics changes nothing...I feel like the issue is with the drive body.
 
Presumably the electronics/logic board of the drives have some aluminium electrolytic capacitors, maybe even tantalums. Open circuit, or out of tolerance?
 
there are 9 aluminum electrolytics; none have the appearance of leaks. I scoped them all and they have stable voltages.. maybe not the right voltages, but non zero.
I'm really fixating on - why that stepper doesn't move in one direction. the coils appeared ok, and the drive signals are there and look ok.
Since I can manually rotate the axle freely, there can't be a mechanical conflict in there that is preventing rotation.
If the coils had a shunt to ground, those drive signals would not look normal; they'd be ..ground. So..
Why would the stepper drive signals be inadequate to torque the stepper...
And the problem tracks the mechanical not the controller board. So any defect is not related to the circuitry on-board.
 
In any electronics this age, electrolytics are always a prime suspect but to check them properly you really need an ESR meter. They're all over Ebay and not expensive.
 
DRIVE #1 - STEPPER MOTOR RELATED

From the other thread:

* the drive with the "stuck stepper" isn't stuck at all, it just only seems to go in one direction. The coils buzz out ok
* turns out I can rotate the steppers easily when de-powered. Both drives move easily.
* If I rotate all the spindles counter clockwise, both drives rotate the spindles clockwise till they bottom out.
* then the "bad" one just rattles around at the bottom while the "good" one swings the arms quickly rail to rail back and forth.

I have also swapped the controller boards between the 2 drives. no difference. so the issue with the stepper tracks the mechanical.

That kind of reminds me of the following. Going back a few months, I brought some 1.2M floppy drives out of years of storage, for the purpose of 'exercising' them. One drive failed. My label on the drive indicated that I had tested it good about 4 years ago. I noticed that it stepped in okay, but intermittently stepped out (or maybe it was the other way round). To my surprise, a clean+lubrication of the carriage rails fixed the problem. A surprise, because I thought inadequate rail lubrication would have resulted in poor stepping in both directions.
 
DRIVE #2

At power-on, you see the stepper move, in both directions. Surely, that has to be part of the drive's self test. And you should be able to determine that by hooking up only power to the drive.

If I, in the day, had been tasked with writing the self test code for an MFM/RLL hard drive, the gist of it may have been something like the following.

Step 1: Test CPU/MPU. If not as expected, trigger 'Bad CPU' error.
Step 2: Verify checksum of CPU/MPU code ROM. If not as expected, trigger 'Bad code' error.
Step 3: Test RAM buffer. If not as expected, trigger 'Bad RAM' error.
Step 4: See if spindle rotating (via index sensor). If not rotating, trigger 'No rotation' error.
Step 5: Wait a certain period of time for platters to get up-to-speed. If not up-to-speed after that timeout period, trigger 'Did not reach speed' error.

At this time, platters are up-to-speed (i.e. heads no longer contacting platter surface). It is now 'safe' to move the heads.

Step 6: If track 0 (cylinder 0) switch not activated, step heads out until the track 0 switch activates. If still not activated after [drive's cylinder count] steps, trigger 'track 0 switch' error.

At this time, heads indicated (via switch) as being on track 0.

Step 7: Step in 5 times (i.e. heads expected to then be over track 5)
Step 8: Verify that the track 0 switch is no longer activated. If not as expected, trigger 'track 0 switch' error.
Step 9: Step out 5 times (i.e. heads expected to then be over track 0)
Step 10: Verify that the track 0 switch is activated. If not as expected, trigger 'track 0 switch' error.

At this time, the track 0 switch (and related circuitry) is verified as working.

Step 11: LONG SEEK - Repeat steps 7 to 10, but instead of track 5, go to the last track. On failure, trigger 'long seek problem' error.



Test the track 0 (cylinder 0) switch ?
 
I did this once to free a stuck head latch, The drive I fixed worked long enough to recover the data. I think if you made a diy clean cupboard you might get it work for a long while.
 
If you plan to let it go to dumpster, then there’s nothing to lose by opening it. That said, I would find a relatively clean room to do it. Make sure you use some sort of compressed air to blow away the dust before you put the case back together. I did this to a few drives and they have been working well for at least a few years now.
 
DRIVE #2

At power-on, you see the stepper move, in both directions. Surely, that has to be part of the drive's self test. And you should be able to determine that by hooking up only power to the drive.

If I, in the day, had been tasked with writing the self test code for an MFM/RLL hard drive, the gist of it may have been something like the following.

Step 1: Test CPU/MPU. If not as expected, trigger 'Bad CPU' error.
Step 2: Verify checksum of CPU/MPU code ROM. If not as expected, trigger 'Bad code' error.
Step 3: Test RAM buffer. If not as expected, trigger 'Bad RAM' error.
Step 4: See if spindle rotating (via index sensor). If not rotating, trigger 'No rotation' error.
Step 5: Wait a certain period of time for platters to get up-to-speed. If not up-to-speed after that timeout period, trigger 'Did not reach speed' error.

At this time, platters are up-to-speed (i.e. heads no longer contacting platter surface). It is now 'safe' to move the heads.

Step 6: If track 0 (cylinder 0) switch not activated, step heads out until the track 0 switch activates. If still not activated after [drive's cylinder count] steps, trigger 'track 0 switch' error.

At this time, heads indicated (via switch) as being on track 0.

Step 7: Step in 5 times (i.e. heads expected to then be over track 5)
Step 8: Verify that the track 0 switch is no longer activated. If not as expected, trigger 'track 0 switch' error.
Step 9: Step out 5 times (i.e. heads expected to then be over track 0)
Step 10: Verify that the track 0 switch is activated. If not as expected, trigger 'track 0 switch' error.

At this time, the track 0 switch (and related circuitry) is verified as working.

Step 11: LONG SEEK - Repeat steps 7 to 10, but instead of track 5, go to the last track. On failure, trigger 'long seek problem' error.



Test the track 0 (cylinder 0) switch ?
Sounds reasonable. Not sure where to start though? There is a set of cabling coming off the armature. I dont have a pinout but I'd there was going to be a hardware switch based on armature position then it should be there.

I may poke around on that a bit. Thanks.
 
There are many drives from this era that use a rubber bumper that turns from a shock absorber into an adhesive. Going in and freeing the head assembly will usually get you enough service life to scrape the bits and if you clean the gunk off and cover the bumper with kapton tape, then it can last for years.
As to a cleanroom: Before I had easy access to a cleanroom, I opened up several 5-40MB drives and did work on them with very good success. I don't recall any that looked decent inside when I opened them and I was able to get working that failed subsequently. Wrap the part of the drive that isn't the lid in plastic or foil (after vacuuming the drive carefully and/or blowing it off). Then line a cardboard box or other frame with clear plastic as a glove-box and work inside that. There is a filter in the drive, so if you don't get silly with it, the drive does have some ability to clean itself.
In short, if you are careful, you should be fine.
I'll be cracking a couple RD53s later today to deal with rubber parts, but I'll be doing that in a cleanroom, just because I happen to have one.
 
Back
Top