Page 1 of 1

Seeking algorithm for negative tracks

Posted: Mon May 28, 2012 10:28 pm
by TeaRex
May I propose a different seeking algorithm for "negative" tracks? I just modified a non-Panasonic drive, a Mitsumi/Newtronics. It steps to -8 very well, but then it doesn't step up again until "negative space" is exited. So it will read track 0 of a normal flippy disk backside fine, then it will still read track 0 when it wants to read 1...4, then read 1 when it should read 5, 2 for 6 etc.

I suggest that every "negative" track is seeked to directly from track 0. So for reading tracks starting at -8 as typical:

* Disable Track 0 bypass.
* If light barrier is sensed as blocked, seek upward one step at a time until it's not, then you know you're on track 1
* Now seek down step by step until the barrier is blocked. Now track 0 is established.
* Enable bypass
* Seek down 8 tracks, to -8
* Read track
* Disable bypass
* Seek to upward to track 1 again (check unblocked)
* Seek down to track 0 again (check blocked)
* Enable bypass
* Seek down 7, to -7
* Read track
* Disable bypass
* Seek upward to track 1 (check unblocked)
* Seek down to track 0 (check blocked)
* Enable bypass
* Seek down 6, to -6
etc.

This would be slightly slower but I think it should work with this non-Panasonic drive, while the current DTC does not. Also it would avoid the drive (not just this one but all drives) getting stuck in "negative" space as it often does at this moment.

Whenever you seek back to 0 after an operation is finished, you'd first disable the bypass and then, if the barrier is sensed as blocked, seek up to track 1 and only then down to 0, to make sure the head is really in a known position.

One more thing that's separate from this: if there is a read error on the minimal track (e.g. track -8 if -b-8), the head shouldn't ever "wiggle" downward, only upward. There might not be enough room for downward wiggling.

What do you think?

Re: Seeking algorithm for negative tracks

Posted: Tue May 29, 2012 12:28 pm
by IFW
When retrying a bad track DTC seeks one step below and above the track that shows a read error; this should help eventually settling the head in a better position.
However when reaching either the minimum or the maximum track this stepping is limited to valid tracks only, ie one direction - exactly as you described.

As for stepping it's a possibility, but would require a drive for testing - these things usually don't work in practice like they should in theory :lol:

Re: Seeking algorithm for negative tracks

Posted: Tue May 29, 2012 1:03 pm
by TeaRex
IFW wrote:As for stepping it's a possibility, but would require a drive for testing
Unfortunately the Mitsumi/Newtronics isn't my own :( , I've just sent it back to the owner for whom I modified it. Otherwise I'd of course lend it to you.

But anyway I'd think it would be a good thing for all drives if you always extablish that you're not in fact already on a negative track before going there. Even for Panasonic drives, if you break the dtc program while the drive is in negative, it won't go back up to 0, and next time it will smash rather violently into the stop or whatever is left of it.

Seeking algorithm for negative tracks

Posted: Tue May 29, 2012 3:43 pm
by mr.vince
Ok, maybe I can help. I think I have several of these mitsumi/newtronics drives. Could you repeat the mod on one of those? This could then be shipped to IFW...

Re: Seeking algorithm for negative tracks

Posted: Tue May 29, 2012 7:58 pm
by TeaRex
Sure, why not. I'll PM you my address. But no hurry, I don't think I'll get around to it until maybe the weekend after the next.

Re: Seeking algorithm for negative tracks

Posted: Tue May 29, 2012 10:04 pm
by mr.vince
Sweet.