• Please review our updated Terms and Rules here

My theory on Kalok drive failures

ST251

ST251

Experienced Member
Joined
Oct 19, 2020
Messages
146
Location
Northeastern US
Just rambling a bit here. I was tinkering with my dead Kalok Octagon MFM drive today (apparently can't find track 0 as the heads just slam against the internal stopper for a few seconds and then the drive spins down) when I noticed that there appeared to be no Track 0 sensor anywhere in the drive. I started doing a little digging on the internet and came across this patent (filed by none other than the thread's namesake).

https://patents.google.com/patent/US4949202

Now in my experience, drives of MFM vintage don't seem to store data reliably in the long term. That made me come to the conclusion that media degradation (more specifically degradation of already stored data) is the reason they fail. I think it is (was) a silly design as it essentially eliminates the one big plus of most MFM drives, and that's their ability to be low level formatted even if every sector is unreadable.
 
That patent is interesting as they tried to use a technique that later drives use (servo information) in a limited manner to detect track 0. This patent describes a mechanism for detecting track 0 using the special pattern. Modern drives are effectively doing the same thing all of the time, with the servo information embedded in the data areas on every track.

In theory this is fine and the reduced complexity is a plus. But as you point out, unless the low level format routine can lay that pattern down on the special track the drive is eventually doomed. (Then again, these things were not designed to last more than 5 years.)
 
mbbrutman said:
That patent is interesting as they tried to use a technique that later drives use (servo information) in a limited manner to detect track 0. This patent describes a mechanism for detecting track 0 using the special pattern. Modern drives are effectively doing the same thing all of the time, with the servo information embedded in the data areas on every track.

In theory this is fine and the reduced complexity is a plus. But as you point out, unless the low level format routine can lay that pattern down on the special track the drive is eventually doomed. (Then again, these things were not designed to last more than 5 years.)
Click to expand...

Since they filed it in 1985 in Japan, I think they were using it already when they made the particular drive I have (KL320, 1989 date stamp). There's no Track 0 sensor in the drive (opened it and looked) and its symptoms fit what I think would happen if it couldn't find the track (constant back stepping). Yep, modern drives do it but they also don't seem to lose data from non-use like MFM drives do. I have an XT that came with the original controller card and drive and it was completely unreadable from sitting unused for so long.

That's the problem, it can't re-make that track. According to the patent, it has to be done in the factory which sounds just like modern drive servo writing. With no secondary positioning sensor I don't see how it would be possible to re-establish a track with that special pattern.
 
Factory is correct--I recall seeing a servo writer setup at the Tandon R&D lab. Used a laser, interferometer and a pneumatic cylinder all mounted on a 5 inch thick slab of granite.

There were also drives with embedded servo bursts on every track. You couldn't LLF those either.
 
Chuck(G) said:
Factory is correct--I recall seeing a servo writer setup at the Tandon R&D lab. Used a laser, interferometer and a pneumatic cylinder all mounted on a 5 inch thick slab of granite.

There were also drives with embedded servo bursts on every track. You couldn't LLF those either.
Click to expand...

When you think about it, Kalok was way ahead of the times with that patent seeing as how servo writing would become the future of drives. If I'm not mistaken, early voice coil drives had to reserve an entire platter for servo information and later drives had to mix it in with data thereby reducing space for data. Their system had only one track reserved in the negative cylinder space and it also had a section that served as an index sensor. I also theorize that if magnetic media in the 80s was more reliable in terms of long term unmodified data storage, those drives would possibly be alive today.
 
I recall the "shoebox" IMI 7710 10 MB drives. Lift the front of the drive about an inch off the table (i.e. incline the drive) and the voice-coil servo would go nuts trying to seek to track. I think that Corvus used them for a time. Bitsavers brochure. I think Priam had a similar drive.
 
ST251 said:
Yep, modern drives do it but they also don't seem to lose data from non-use like MFM drives do. I have an XT that came with the original controller card and drive and it was completely unreadable from sitting unused for so long.
Click to expand...
Define "for so long". How can you know a modern drive would not lose its data when sitting unused for the same amount of time? You would need to be a time traveller to know...

I had many modern drives (year 2010 and newer) that would no longer find the server tracking information and became useless. If that fades, data does, too. Have to say these were 2.5" drives mainly, but a few 3.5 ones as well.
 
Chuck(G) said:
I recall the "shoebox" IMI 7710 10 MB drives. Lift the front of the drive about an inch off the table (i.e. incline the drive) and the voice-coil servo would go nuts trying to seek to track. I think that Corvus used them for a time. Bitsavers brochure. I think Priam had a similar drive.
Click to expand...
That's quite an interesting thing to see the use of servo information on a drive first produced in 1979. I'm not surprised to hear it would act like that though, I'm sure that was a lot of weight to put on such an early voice coil system.

Timo W. said:
Define "for so long". How can you know a modern drive would not lose its data when sitting unused for the same amount of time? You would need to be a time traveller to know...

I had many modern drives (year 2010 and newer) that would no longer find the server tracking information and became useless. If that fades, data does, too. Have to say these were 2.5" drives mainly, but a few 3.5 ones as well.
Click to expand...

Well, from my understanding, the IBM in question sat for I believe 20 years or so. I have a drive from around 1998 that I just checked and can still read information.

I'm assuming that the failures of the recent drives could possibly be related to the heads and not the platters themselves. In my experience, the heads on these newer drives can just up and stop working for no reason. I had that happen to me on a drive that I bought a few years ago. It was a backup drive and one day I noticed the read error SMART attribute increasing rapidly, then came the current pending sectors and then the drive was dead.
 
You must log in or register to reply here.
Back
Top