Writing dumps, how to increase stability?

All questions about how to use KryoFlux go here.
SLC
Posts: 55
Joined: Mon May 04, 2015 10:03 pm

Writing dumps, how to increase stability?

Post by SLC » Fri May 08, 2015 8:04 pm

Hi

I'm one of the (many?) who purchased the KryoFlux with the intent of going both ways (preserve and master). I've ran into a few problems I was hoping you'd be able to shed a light on.

Games using "fat tracks" seems to be rather difficult to write back, same with a more recent Vorpal. Regarding fat tracks, sometimes it works, sometimes it doesn't. Recent Vorpal I'm unable to write fully functional at all and it seems that it is different tracks failing from time to time. I have two drives, one Epson with a fake index sensor and one Panasonic which is flippy-modded (and obviously the one I use for dumping disks). Both done by the excellent ZrX.

Any help and/or explanations appreciated :)

Feltzkrone
Posts: 255
Joined: Sat Jun 18, 2011 8:03 pm

Re: Writing dumps, how to increase stability?

Post by Feltzkrone » Fri May 08, 2015 8:24 pm

Do you write the stream files directly or do you use G64 files with mastering info as the source?

SLC
Posts: 55
Joined: Mon May 04, 2015 10:03 pm

Re: Writing dumps, how to increase stability?

Post by SLC » Fri May 08, 2015 8:28 pm

Mainly G64 files with mastering info (the G64s are fine, they work perfectly in WinVICE.. every time). Was told streamfiles would be even more unstable so haven't tried that...

User avatar
IFW
Posts: 3079
Joined: Mon Nov 08, 2010 2:42 pm

Re: Writing dumps, how to increase stability?

Post by IFW » Fri May 08, 2015 8:48 pm

The usual stuff:
- use good disk
- degauss the disk
- only use dd or sd disk, never hd
- use the Panasonic drive
- most importantly: make sure you add the correct -dd mode to the command line; this is very important due to coercivity during write and filtering during read/verify
- use verify

That's it really - although I may have forgotten something from this checklist, so feel free to add items :)
I think the manual explains the problem:
- the actual footprint difference between the tracks written with a 48 TPI head vs 96 TPI head
- a disk you write with a 80 track head (96 TPI) should never be written with a 40 track (48 TPI) head, otherwise the written area will have crosstalk and random readability
For reference the drive you use for mastering is 96 TPI, the 1541 is 48 TPI. Therefore if the 1541 ever writes to the disk that part will be unstable, unless you can properly degauss it.
Old 8 bit systems used 80 track mechanism mixed with 48 TPI head, meaning larger track footprint that you'd have with a properly paired 80 track+96 TPI. Every second track (the "half-tracks", but there is no such thing for real) will have issues due to the 48 TPI (wider) track footprint.
This was used as alignment was not reliable and a track could be possibly read this way even on a misaligned drive if adjacent tracks were tried if the target track failed to read.
This would also make it pointless to have a track 0 sensor... (since you have to microstep depending on alignment) so there was not one for a long time.
Not only CBM was this cheap; BBC, Apple and Atari were all using the very same "technique" :)

SLC
Posts: 55
Joined: Mon May 04, 2015 10:03 pm

Re: Writing dumps, how to increase stability?

Post by SLC » Fri May 08, 2015 9:09 pm

I use fresh disks when I do this. Will I need to degauss even unused and unformatted floppies?
Of course I'm also using only DD and SD disks. :) I've tried with both -dd0 and -dd1, but neither seems to affect the result much. It's still pretty random whether or not it works.

Will later versions of DTC prove more stable or is this as good as it gets when attempting to remaster?

Also, the floppies that fail are passing verify tests every time...

This is surely a confusing topic for someone who's used to tapes only :)

User avatar
IFW
Posts: 3079
Joined: Mon Nov 08, 2010 2:42 pm

Re: Writing dumps, how to increase stability?

Post by IFW » Fri May 08, 2015 9:19 pm

Writing is actually very reliable - reading it back can be problematic, again, due to the different head footprint. So it's very important to have the 1541 drive very well aligned, otherwise you'll see random failures.

If the verify on the Panasonic succeeds, but the disk fails on a 1541 that is because of crosstalk or misalignment.
Remember that the Panasonic writes at 96TPI and reads at 96 TPI as well - hence it is capable of perfectly reading (in this case, verifying) all of the tracks, regardless of their (pre)formatting or the alignment of the target 1541 drive.

So if the verify works without any problem the things worth checking are:
- are you sure the disks were not preformatted to e.g. IBM?
- if you are certain the disks are empty and were never ever touched by a 48 TPI head, that means that 1541 needs to be aligned.

Also keep in mind that many of these loaders actually do no checksums during reading... so known good data (and KF software will only produce good data) will always load without any problem under emulation, as the virtual disk will always read correctly.
A real disk can read bad data fairly easily on the other hand... :)

SLC
Posts: 55
Joined: Mon May 04, 2015 10:03 pm

Re: Writing dumps, how to increase stability?

Post by SLC » Fri May 08, 2015 9:31 pm

The alignment of the 1541s should be OK, but I don't know how I can further test that? The closest I am to having an alignment disk is the Vorpal Utility Kit. I've also tried the disks on several 1541s, and the ones that fail fails on them all. It's also worth noting that I've sometimes tried re-writing on the same disks several times, and ended up with a working copy after 2 or 3 tries. However, for the Vorpal version used on for example Street Sports Basketball, I've not been able to make a stable copy.

Normal DOS and protections using weak bits has never failed me.

If there are any software packages for 1541 to align using the Panasonic written disks, I'd be happy to hear about it. If I need a scope and a real alignment disk I'm afraid I'll be out of luck :(
But again, there is definitely differences between writes done on both the Panasonic and the Epson since sometimes it works, sometimes it doesn't.. same images and both fails and successes on both new disks and disks I already wrote a few times, so writing does definitely not create the exact same results every time.

SLC
Posts: 55
Joined: Mon May 04, 2015 10:03 pm

Re: Writing dumps, how to increase stability?

Post by SLC » Fri May 08, 2015 11:59 pm

Quick update. I have now tried some calibration of the 1541. I wrote back the Free Spirit tool using the same drive I wrote another image with. Alignment according to Free Spirit is good, and the game still doesn't load. I also checked with my original Vorpal Utility Kit disk, and it's also according to that in alignment. I also tried writing back this disk with the KF, and the results are the same. While this is not a professional calibration, I feel this indicates the drive not being out of alignment, so there's something else here which sabotages me!

User avatar
IFW
Posts: 3079
Joined: Mon Nov 08, 2010 2:42 pm

Re: Writing dumps, how to increase stability?

Post by IFW » Sat May 09, 2015 7:25 am

You actually need "real" alignment, a 1541 with a disk is far from that.
For example Tommi could help you I think.
The suspicious part is that you said that the verify never shows any error.
Try dumping the same disk then after writing, and see if it's still error free.
If it is, still the above two things are the causes:
- the disk was written with a 48 TPI head before
- 1541 out of alignment

You may also want to disable/remove any cartridges, extended roms etc when running a V-MAX! disk, as many of them check even for memory extensions within the 1541.

There are also a few V-MAX disks that you'll have a problem running under emulation as well as a real C64 setup, due to the way how they check for cartridges.
If they think there is a cartridge present, they load for a short while, then either go back to the command prompt or reset.

Exactly which game it is?

Feltzkrone
Posts: 255
Joined: Sat Jun 18, 2011 8:03 pm

Re: Writing dumps, how to increase stability?

Post by Feltzkrone » Sat May 09, 2015 7:49 am

István has been quicker but nevertheless... that's what I would have suggested:

Just to be sure:
1) Verification in DTC really says that it's ok?
2) Resulting disks which are not working properly in 1541 drives remain fully readable with DTC?
3) G64 files of re-read disks still work fine in VICE?

It would be fine if we first could rule out any problem with DTC. If you can answer all of these questions with 'yes' then a problem with DTC becomes very unlikely and if you can answer any of these questions with 'no' then we at least have something to start with. Either way, do you already have access to the contrib FTP server and would mind uploading dumps and logs (zipped) of the offending disks/games there? There's also KRYOFLUX_TESTDUMPS directory where you could upload a dump of a problematic disk previously written with KF so we could take a look if necessary.

Oh and some more:
4) Have you used disks from different brands?
5) Are standard CBM DOS disks without protection fully readable when written with KF and read with 1541?
6) Does that still apply when you are using disks which have been written to using a 1541 previously?

XEMAG might fail if RPM of the drive used with KF is rather unstable. A dump of a previously written disk should reveal if that's the case or not. Why Vorpal fails is still a puzzle. Does it affect all games with later Vorpal (many sectors per track) that you own or only particular titles?

Post Reply