Is my Floppy Disk Image Good?

A forum about Atari protected floppy disks analysis, preservation, emulation, tools

Moderators: DrCoolZic, Brume

User avatar
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 2258
Joined: Mon Oct 03, 2005 7:03 pm
Location: France

Is my Floppy Disk Image Good?

Postby DrCoolZic » Tue Jan 14, 2014 11:19 am

Important Note: This thread is located in the KryoFlux subforum, but it could have been located in the SuperCard Pro subforum as currently the Aufit program read flux images from these two devices.

This thread is to discuss the questions around the quality of Floppy Disk images created with devices like KryoFlux or SuperCard Pro. Normally these questions should be somewhat independent of the imaging device as I only want to discuss problems related to the floppy drives and floppy disks conditions.

First you want to have a clean floppy drive head and a spotlessly clean floppy disk (sorry but finding the right English terms is not that easy). What I mean is that your floppy drive and floppy disks need to be in the most perfect conditions you can get.

I used to image my floppies without cleaning the floppy drive head or cleaning the diskette and indeed I found out that cleaning can make a big difference.
About cleaning technique refer to : viewtopic.php?f=102&t=25841&p=243278&hilit=brume+cleaning#p243273

Personally I have bought two head cleaning kits that contains a cleaning floppy disk and cleaning liquid. When I need to image diskettes I first clean the head of the floppy drive following the provided instructions.
When I find problem with a floppy disk image I clean the floppy with a Q-Tip impregnated with head cleaning liquid. Normally this liquid evaporate quickly but you can use the other side of the Q-Tip to help dry the liquid. Then I redo the image of the diskette and check the result again.

Here is an example of the results I have obtained just by cleaning the diskette. viewtopic.php?f=102&t=25854&start=25#p243549

Important thing to know is how to find the correct side of the diskette to clean. If you look at a floppy diskette you have one side with the etiquette and this is the “second logical side” (side/head 1) of the diskette and if you flip the floppy you have the side with the hub that engages with the drive motor and this is the “first logical side” (side/head 0) of the diskette.
So if you find problems on side/head 0 you need to clean the side without the etiquette (what I refer as the back side) and if you find problems with side/head 1 you need to clean the side with the etiquette (what I refer as the front side). For cleaning you need to open the metal shutter and keep it open (for that matter I use a scotch tape).
Track 0 is the most external track and track 84 is the most internal track

Now let’s move on the heart of the subject: How do I find if the image I have done is of good enough quality. I am not an expert on floppy disk magnetization degradation and therefore I will need feedback from people that know.

For looking at problems I use my Aufit program.
Important note: Currently this program is only designed to analyze Atari diskette so apart from the scattered chart of the raw flux transitions all the other charts are processed assuming an Atari MFM Double Density diskette. So do not expect to use it to analyze diskette from other platforms like Amiga or C64/C128 …

The program is currently in early development stage so be nice if you find problems. To get latest information about Aufit program please refer to this thread viewtopic.php?f=102&t=25906

The first thing is to read into the program the image you have created (for that refer to Aufit doc).
Once the program has read the image it gives some useful information in the Track grids as the background color for each track is already an indication of possible problem: A “normal” track has a light green background, an unformatted track has a light gray, and a track with potential problem is displayed in yellow.
At this stage only a quick Shannon Entropy analysis is done of the flux transitions histogram. A yellow color usually indicates tracks with partially unformatted areas but also tracks with problems.
For examples here is the result of reading Sherman M4 diskette.

We can see that we have several tracks with a yellow background. If we click on these tracks we will find some track are in fact correct. But if we click on track 66 we can see than sector one has a lot of “dirty” flux transitions. Thanks to the DPLL these variations can be corrected and we are able to read the sector correctly.

But if we select the track 73 and zoom on the first sector area we can see that we also have dirty flux transitions, but here the DPLL is not capable to decode the information correctly: we can see that the first sector is read with fuzzy bytes (orange box in data chart and indication in the information window). In fact is we further analyze and look at the fuzzy mask buffer we find out that only ONE byte at position 214 vary between revolutions! This also why it is recommended to sample at least four revolutions (Aufit always checks fuzzy bytes using all available revolutions).

It seems obvious that this error is not part of a protection but is a problem of the floppy diskette.
First try to clean your floppy and try to image again.
I leave to the expert the “magnetism” explanation
You do not have the required permissions to view the files attached to this post.

User avatar
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 2258
Joined: Mon Oct 03, 2005 7:03 pm
Location: France

Re: Is my Floppy Disk Image Good?

Postby DrCoolZic » Mon Aug 31, 2015 5:05 pm

Just a quick information.

I am currently working on better understanding the data separator used to decode data in a bit stream.

I have found an interesting information that I want to share. You may have noticed on some image when you look at the Flux transition chart that some transitions forms two lines one on top of the other (I know it is not clear but wait…). I always wondered why!
I think that now I have an answer.

Thanks to a new version of Aufit that has the capability to use one color (black) for odd transition and another color (red) for even transition. And you get this kind of result.

As you can see in the 4, 6, and 8 µs regions we have a blue line and a bit above a red line.

I am currently writing a paper on data separation (that I may publish) and here is the interesting information in it:

“read channel can be asymmetrical: The drive’s peak detector may be unbalanced, resulting in a bit shift. This could cause positive going peaks to appear earlier than negative going peaks or vice-versa. Also, a long cable between the disk drive and the disk controller may contribute to bit shift”

Why is this important because some drive may be more asymmetrical than other and therefore you should prefer a more symmetrical one. And try not to use long cable.

Is it bad to have very asymmetrical bit transition?:
Yes even though the DPLL is supposed to deal with this situation it will inevitably degrade the decoding of the bits.

Note that the above example is not that bad if you zoom you can see that the blue dots range from 5.7-5.9 and the red one from 5.9-6.2
You do not have the required permissions to view the files attached to this post.

Social Media


Return to “Floppy Disk Preservation”

Who is online

Users browsing this forum: No registered users and 2 guests