Page 1 of 1

Dumping flippy in non-flippy drive using fewer than 2 passes?

Posted: Tue Oct 23, 2018 1:45 am
by rcaputo
I have some fragile Apple DOS 3.3 diskettes that I may try to dump in a non-modded drive.

It seems like I should be able to minimize a donut's spindle time by dumping it in flippy mode, flipping it in the sleeve, and then dumping only the tracks that were missed in the first read.

1. Is that even possible? (It doesn't sound impossible.)
2. Would it reduce the theoretical risk of damage to the donut?
3. If anyone has done this before, is the process already documented?


Re: Dumping flippy in non-flippy drive using fewer than 2 passes?

Posted: Tue Oct 23, 2018 3:20 pm
by SomeGuy
I assume you mean flipping it over in the jacket? Sleeve is usually the term used for the paper/tyvek thing the disk is stored in.

When doing professional data recovery, that is actually the preferred way to do things. You remove the cookie/donut/brown magnetic disk from the jacket, carefully clean the cookie, possibly "bake" the cookie in an oven to temporarily prevent oxide shedding, then place it in a fresh clean jacket (such as from well stored unused higher quality new old stock disks), and then make your read attempts.

Of course, that renders the disk unfit for future reuse in vintage systems, and destroys the value of factory pressed software titles.

I'm sure most of the Kryoflux folks would insist you use a flippy modded drive. I had looked in to flippy modding myself recently but modding such a drive requires tools I don't have, irreversible metal grinding, and the high probability of a klutz like me destroying the drive. Not worth it for a small handful of disks.

If you have the jacket cut open anyway then yes, flipping the cookie over will work perfectly and eliminate the need for a modded drive.

If your goal is serious data recovery, then grime/residue/dirt/sand inside the old jackets is a real, serious, issue. While I personally usually just wash the entire disk, placing the cookie in a fresh clean jacket greatly reduces the risk of things getting scratched up.

Another option to consider is punching a second index opening in the jacket (not the cookie!). That does risk some damage, but that was actually the approach I took since my main goal was to WIPE the data, test the disks, and re-use any good ones on an Apple II system. I used an empty jacket and pencil to mark the location of the hole on each disks, I used a hole punch with tape over the hole catching side instead of the metal cover to help reduce scratching, and then carefully inserted the hole punch via the hub and punch the marked area.

Back in the day I used to punch holes like that to make flippies for single sided systems that required the index, such as the TRS-80, TI-99/4a and single sided IBM PCs.

As for the actual archiving pass, I usually suggest starting with a flux preservation dump. You can feed those back in to the Kryoflux software to decode them later. If the first dump does contain errors, then follow up with an interactive dump where you can more easily retry tracks, clean specific spots, press down on the head a tad, or do whatever else is needed to coax the bits off of the disk.

Without a f lippy modded drive, you will want to just concern yourself with only one side at a time.

Re: Dumping flippy in non-flippy drive using fewer than 2 passes?

Posted: Wed Oct 24, 2018 12:01 am
by rcaputo
Thanks for the response and the good advice.

You're right; I meant "jacket" not "sleeve".

I feel like the main question wasn't answered. I probably asked it poorly.

Assuming a modded drive is out of the question...

If the first dump is two-sided, could I get away with only dumping 16 more tracks after flipping the donut/cookie, and merging the extra 16 .raw files back into the original dump?

Something like this seems feasible (note, I may have got offsets wrong):

dtc -g2 -s0 -e79 -fdouble -i0
(flip the donut)
dtc -g0 -s0 -e7 -fmissing-0-7 -i0
dtc -g1 -s72 -e79 -fmissing-80-87 -i0

Then the .raw files from all three dumps would need to be merged.

This might reduce wear on the medium and time spent spinning it against the grain, compared to doing two full single-sided reads.

I've read most of the KryoFlux forums. Merging .raw files from mulitple dumps is supported. I haven't seen anyone mention optimizing the second dump, though. Maybe it's a terrible idea for reasons I don't yet understand?

Re: Dumping flippy in non-flippy drive using fewer than 2 passes?

Posted: Wed Oct 24, 2018 5:14 am
by SomeGuy
Had not tried that before, so I just tried switching around a RAW track like that and as far as I can tell that won't work.

When decoding, the tracks on each side of a flippy disk are read in opposite directions. Moving a RAW track from one side to the other, the KF has no idea it should decode that track backwards.

You could capture preservation streams like that, but you would have to merge IMG files AFTER decoding.

I would just capture full 0 to 79, flip it over and capture a second dump 0 to 79 again. From my own experience If the disk is going to shred itself, it will usually do it in the first pass and a second pass won't do much more damage. Hopefully most of the time you will get a good and easy decode, then in the few times you don't you can check the first dump to see if it can decode the bad sector(s).

Re: Dumping flippy in non-flippy drive using fewer than 2 passes?

Posted: Wed Oct 24, 2018 7:20 am
by brightcaster
That's actually my usual workflow with really bad looking disks! I do a format guided Raw dump of all tracks first. After that I atempt a retry to only the bad tracks to another target directory. If any get good reads after several retries I copy them to the first directory replacing the former bad files.

Many times I get alle tracks fine in the end. In a last step I decode the streamfiles to the target image format in offline mode...