No flux transitions

[DrCoolZic]

It is not possible to reproduce the exact magnetic state of this protection with a standard drive. Consumer drives were not designed for erasing, let alone this type (DC) of erasing. Because this is a limitation at the drive side, a custom controller can't help here.

However, it is possible to emulate the protection behaviour, at a certain degree, with some trickery. I can't say for sure, though, how much reliable this is. It probably depends on both drives (the writing and the reading one), and on the disk as well.

Quite interesting. I did not thought too much about the way to write this kind of "no flux transition area" (NFA). Of course there is no way of "writing" NFA ! Therefore the only way is to erase the NFA.
If people from SPS/KryoFlux are looking at the thread I would be interested to hear from them on the subject. Will the kryoFlux device in write mode be able to reproduce such area and how?

[ijor]

However, it is possible to emulate the protection behaviour, at a certain degree, with some trickery. I can't say for sure, though, how much reliable this is. It probably depends on both drives (the writing and the reading one), and on the disk as well.

...
If people from SPS/KryoFlux are looking at the thread I would be interested to hear from them on the subject. Will the kryoFlux device in write mode be able to reproduce such area and how?

My quote above applies to the Kryoflux and similar controllers (Catweasel and Discovery Cartridge).

[mr.vince]

It is not possible to reproduce the exact magnetic state of this protection with a standard drive. Consumer drives were not designed for erasing, let alone this type (DC) of erasing. Because this is a limitation at the drive side, a custom controller can't help here.

However, it is possible to emulate the protection behaviour, at a certain degree, with some trickery. I can't say for sure, though, how much reliable this is. It probably depends on both drives (the writing and the reading one), and on the disk as well.

Quite interesting. I did not thought too much about the way to write this kind of "no flux transition area" (NFA). Of course there is no way of "writing" NFA ! Therefore the only way is to erase the NFA.
If people from SPS/KryoFlux are looking at the thread I would be interested to hear from them on the subject. Will the kryoFlux device in write mode be able to reproduce such area and how?

Yes, it does. Stunt Car Racer has such no flux areas, as well as Treasure Island Dizzy (first release). Both write back perfectly. And fail on the same Amigas that don't play the original disk as well.

But since Ijor says this can't be done, we must be doing magic here.

I'd say that strong DC erasure is nonsense, at least for these titles. I don't want to say this is impossible, but both titles were protected by Rob Northen with the old Copylock (also found on Stormlord Amiga btw). Rob Northen used Trace replication, so there was no "strong DC erasure" available. Just standard Trace commands and standard drives. And yes, you can generate no flux areas with these. Be creative! /
|-)

[DrCoolZic]

I Do not want to reply for Ijor but you must agree that it is not possible to write no flux reversal on a FD with a normal drive.
Therefore to mimic a no flux reversal area you have to play some trick. I believe that I know at least one way to do that but yes it has to be magic

[Hippy Dave]

Hmmm. An erase signal is usually a high-frequency AC signal (i.e. fast enough that the magnetic domains are left in a random state), but a constant value at the head is a DC signal, so I can see that the effects it would produce on the disk should be different...?

I would have thought Kryoflux could in theory completely erase a track by writing very high frequency data to it, but that might be filtered out by the drive electronics of course.

I Do not want to reply for Ijor but you must agree that it is not possible to write no flux reversal on a FD with a normal drive.
Therefore to mimic a no flux reversal area you have to play some trick. I believe that I know at least one way to do that but yes it has to be magic

What Dio said is possible on the Commodore 64 1541 floppy drive.
Is it not possible on the Atari?
If the Atari can do this then it is time to put flux reversals to rest.

[DrCoolZic]

Hmmm. An erase signal is usually a high-frequency AC signal (i.e. fast enough that the magnetic domains are left in a random state), but a constant value at the head is a DC signal, so I can see that the effects it would produce on the disk should be different...?

What Dio said is possible on the Commodore 64 1541 floppy drive.
Is it not possible on the Atari?
If the Atari can do this then it is time to put flux reversals to rest.

With KryoFlux the drive is directly driven by the board and no Atari involved. And yes this is probably the trick used with the KryoFlux board
I have already submitted several games with NFA but I have not yet received any IPF for them

[mr.vince]

I say it is... You generate no flux areas with a standard drive. Writing of DC is - to stay polite - nonsense. Basic recording theory will tell you that you can't write DC and particles magnetised with a permanent magnet won't generate an AC; they also don't generate a DC at readback. All you can do is add a DC to an AC, which will give an offset - that's it. Leave out the AC, and you lose the DC as well. Presto.

Be creative. I am embarrassed to hear no better explanations. At least Ijor should know... Really!

[ijor]

But since Ijor says this can't be done, we must be doing magic here.

I didn't say exactly it can't be done (I said that years ago when I first found this protection, but I corrected myself long ago), did I? Please read again what I said and you just quoted yourself.

..., but both titles were protected by Rob Northen with the old Copylock (also found on Stormlord Amiga btw). Rob Northen used Trace replication, so there was no "strong DC erasure" available. Just standard Trace commands and standard drives.

I don't know about the Amiga versions, but I doubt Rob Northern has anything to do with this protection in the ST (including Turrican, Dizzy and others). It is very possible that some of these titles used the RN famous trace exception encryption. But this is the software side of the protection, no direct relation with the on-disk copy protection.

And even if it was done by RN, and even if these titles were duplicated by Trace, so what? It doesn't mean anything. In first place, because they could always ask Trace to implement "non standard" commands. In second place because it could be done at a separate pass on a different equipment. I'm not saying that this is what really happened, just that it is possible.

And lastly, of course that Trace (as most other duplicators) didn't use standard drives. They were very special drives that costed many times as much as a regular consumer drive.

Be creative. I am embarrassed to hear no better explanations. At least Ijor should know... Really!

Just in case you didn't note, this is (sort of) my home. It is the subforum dedicated specifically to my own stuff. You are welcome to post anything you consider relevant. But whatever you say, please do that with respect.

I know how you can write back this protection. I intentionally didn't elaborate about the specifics in public, and did that on PM, precisely out of respect to you. At least it seemed to me you didn't want to make this public.

[mr.vince]

And even if it was done by RN, and even if these titles were duplicated by Trace, so what? It doesn't mean anything. In first place, because they could always ask Trace to implement "non standard" commands. In second place because it could be done at a separate pass on a different equipment. I'm not saying that this is what really happened, just that it is possible.

Rob Northen did many ST protections (he even started on the ST and then ported this code), I can't speak of particular titles, IFW would know. Rob Northen did not need special commands, he used what Trace offered which was enough to create disks uncopyable on regular equipment.

And lastly, of course that Trace (as most other duplicators) didn't use standard drives. They were very special drives that costed many times as much as a regular consumer drive.

Just like our DVD ROM does also write discs with CSS on special authoring media... This is due to a special firmware, mechanically the drive is the same. But it makes the drive expensive. The sources we have indicate that Trace had drives produced for them, but based on standard drives, with small modifications, like the one for one-pass flippy duplication. It's really nice to have such things because you can charge a lot more money for this.

In a production process (which sounds to me you never experienced this) the publisher would have kicked you for adding a second step during duplication. Putting disks into drives and getting them out, automated or not, means someone carrying the disks from machine A to B and you would pay for two processes instead of one. It's more error prone, takes longer... so no. This was unusual. Rob Northen worked for the biggies, these relied on cheap duplication and quick turnaround times. If you like, we'll ask him and that's it.

Just to add to this: Yes, there were disks that were duped in two passes, e.g. some Thalion titles, or Little Computer People. But that was not the standard and we found no evidence for this on disks that have the no flux protection.

Just in case you didn't note, this is (sort of) my home. It is the subforum dedicated specifically to my own stuff. You are welcome to post anything you consider relevant. But whatever you say, please do that with respect. I know how you can write back this protection. I intentionally didn't elaborate about the specifics in public, and did that on PM, precisely out of respect to you. At least it seemed to me you didn't want to make this public.

I did. But this is a public forum. Being your baby does not mean people have to remain silent when when they notice something odd. So that does not stop me from saying strong DC erasure sounds like bull to me. This is no disrespect and I don't feel I have to remain silent to respect you. This is a discussion and I'd be pleased to hear your facts.

I just entered the discussion when you made an assumption and KryoFlux was involved. As you see we shared the source to the decoder lib with the world, so we don't have a problem with sharing information. I am just curious if you know, too.

What we know about the no flux protection: It must have been problematic, because it fails on certain brands of drives, so later versions of the same game (e.g. Treasure Island Dizzy) were delivered with the later, density based, protection. This was a step back because the no flux thing is really uncopyable, regardless of which analogue copier you would take.

[DrCoolZic]

To better understand a possible solution I have been trying to find schematic for 3 ½” drive. I thought this would be a piece of cake on the Web but I did not found any!
The only schematic I have found is for a 8” drive.

Based on what I have seen on this schematic it seems like the “write chain” is relatively simple: the digital signal toggle a flip flop before being amplified with transistors to go to the head. I am guessing that passing a relatively high frequency signal to the write circuitry should go through and be recorded on the track without problem.

The read circuitry is much more complex and is composed of a preamplifier, a filter, a differentiator, zero crossing detector and digitizer. The documentation does not mention ACG but it might be there (remember this is a very old 8” drive).
My guess is that if the linear signal input frequency from the head is fast enough (reading the high frequency signal) the differentiator/detector will not detect bits and this will result as no transition on the read data signal?

Is this correct assumption and is this the way it is done with the KryoFlux board ?
Thanks
Jean

[Hippy Dave]

The read circuitry is much more complex and is composed of a preamplifier, a filter, a differentiator, zero crossing detector and digitizer. The documentation does not mention ACG but it might be there.
Jean

The differentiator and zero crossing detector behave as an AC AGC would at low signal level; with high gain AC zero crossing.

[DrCoolZic]

The read circuitry is much more complex and is composed of a preamplifier, a filter, a differentiator, zero crossing detector and digitizer. The documentation does not mention ACG but it might be there.
Jean

The differentiator and zero crossing detector behave as an AC AGC would at low signal level; with high gain AC zero crossing.

Thanks for the info. I must admit that I feel much more comfortable with digital circuitry

For those interested I have updated my page of FD by adding several sections and document references.
It seems that the answer to my questions are in the few sections starting at http://info-coach.fr/atari/hardware/FD-Hard.php#basics

[ijor]

Rob Northen did many ST protections (he even started on the ST and then ported this code), I can't speak of particular titles, IFW would know.

Well, if you have any evidence that RN made the on-disk copy protections of these ST titles, then please post it.

The sources we have indicate that Trace had drives produced for them, but based on standard drives, with small modifications...

Duplicator drives, depending on the model, had such features as writing to both sides at the same time, 2x, or even 4x RPM speed. Does this sound as "small modifications" to you? They were heavy duty, as any industrial grade instrument, designed to be working 24 hours a day. And they had a a wealth of analog features.

Quoting from Trace brochures:

Window margin and dropout testing. Azimuth, radial and amplitude built-in testing. Built-in drive cleaning. Automatic adjustment for head wear. Recording amplitude testing and control, etc, etc.

As you can see, they were very special drives. No way just standard drives with small modifications.

In a production process (which sounds to me you never experienced this) the publisher would have kicked you for adding a second step during duplication... This was unusual...Just to add to this: Yes, there were disks that were duped in two passes, e.g. some Thalion titles, or Little Computer People.

Seems like you are answering to yourself. It was unusual, it wasn't the standard, but it happened for different reasons as you admit yourself.

I am not saying I'm sure that this protection was performed in a separate step. I am just saying this is a posibility. And if you look carefully at the protected tracks, and compare the different titles and multiple copies of the same title, then it sounds even as probable.

What we know about the no flux protection: It must have been problematic, because it fails on certain brands of drives, so later versions of the same game (e.g. Treasure Island Dizzy) were delivered with the later, density based, protection.

It might have been problematic in the Amiga. But seems it was quite reliable in the ST. A fair amount of titles were produced with that protection, spawning a not so small period. And if they changed the protection on specific titles, it might have been for other reasons, including because this protection might have been more expensive to produce.

This is a discussion and I'd be pleased to hear your facts.

I don't have any hard facts or proofs about this, and I don't think you don't have either, or at least you didn't show any so far.

Being your baby does not mean people have to remain silent when when they notice something odd. So that does not stop me from saying strong DC erasure sounds like bull to me. This is no disrespect and I don't feel I have to remain silent to respect you.

I didn't say you have to remain silent, did I? I didn't say you can't debate or contradict my posts, did I? I said that whatever you say, do that with respect. Of course that disagreeing with me is no disrespect. You know exactly what you said that was disrespectful.

As you see we shared the source to the decoder lib with the world, so we don't have a problem with sharing information. I am just curious if you know, too.

Sorry, I don't follow the meaning of the question (if there is a question here). If I know what?

Rob Northen ... If you like, we'll ask him and that's it.

(I intentionally moved this to the bottom, to end on a more positive tone)

Yes, I'll love it. Please do it if you can.

Ask him which on-disk protections he made (if he remembers)? And of course, ask him specifically about this one. Did he make it? How it was produced?

[mr.vince]

Duplicator drives, depending on the model, had such features as writing to both sides at the same time, 2x, or even 4x RPM speed. Does this sound as "small modifications" to you? They were heavy duty, as any industrial grade instrument, designed to be working 24 hours a day. And they had a a wealth of analog features.

Quoting from Trace brochures:

Window margin and dropout testing. Azimuth, radial and amplitude built-in testing. Built-in drive cleaning. Automatic adjustment for head wear. Recording amplitude testing and control, etc, etc.

As you can see, they were very special drives. No way just standard drives with small modifications.

This is all stuff Trace ordered OEM from drive manufacturers, and yikes, 2 x or 4 x RPM... Didn't I post the duplicator information for some games a while ago?! It's unreadable with all other controllers I have seen so far. Should I be surprised that a decade after 5.25" was introduced and many years after 3.5" was introduced drive electronics was capable of writing at double speed? It's like the evolution of CD ROM, it just never happended for the average customer because to lower the price, parts were made more cheaply, e.g. die-cast chassis exchanged for cheaper (but more fragile) sheet metal.

The rest of the features is where drives went, you just did not notice. There was ZIP and Super Drive, and these exactly had all these neat features, again, not accessible for the consumer. But these features aren't responsible for trickery with protections. They add convenience and reliability, thus e.g. enabling a higher density and so forth.

I am not saying I'm sure that this protection was performed in a separate step. I am just saying this is a posibility. And if you look carefully at the protected tracks, and compare the different titles and multiple copies of the same title, then it sounds even as probable.

No. We can see if a track was written in one pass or if there are any splices or the like. No two drives have the exact same speed and again, you would see the splice. The games I was talking about just suppress flux reversals with a clever trick DrCoolZic explained above.

I didn't say you have to remain silent, did I? I didn't say you can't debate or contradict my posts, did I? I said that whatever you say, do that with respect. Of course that disagreeing with me is no disrespect. You know exactly what you said that was disrespectful.

Well, I don't. I just don't feel I should keep silent when I see someone "talking" about KryoFlux and how it works that hasn't the deeper knowledge of the product. If I'd be writing something offensive or something that was meant disrespectful, you would notice for sure. The above isn't. Feel free to PM me and give me a hint.

Rob Northen ... If you like, we'll ask him and that's it.

(I intentionally moved this to the bottom, to end on a more positive tone)

Yes, I'll love it. Please do it if you can.

Ask him which on-disk protections he made (if he remembers)? And of course, ask him specifically about this one. Did he make it? How it was produced?

Did ask him this afternoon. I hope his email is still correct. It did not bounce so... fingers crossed.

[ijor]

This is all stuff Trace ordered OEM from drive manufacturers, and yikes, 2 x or 4 x RPM...
It's like the evolution of CD ROM, it just never happended for the average customer because to lower the price...
The rest of the features is where drives went, you just did not notice. There was ZIP and Super Drive, and these exactly had all these neat features, again, not accessible for the consumer. But these features aren't responsible for trickery with protections.

The relevant point here, and seems you agree now, is that the duplicator drives weren't standard drives, neither standard drives with just minor modifications. If they used some of those features for protections purposes, and specifically for this one, I don't know. Certainly some of these could have been used, conceivably. And besides the mentioned features, those drives had erase heads, mechanism, and erase control, very different than standard drives. This could have been used as well.

Once again, I am not claiming that the protection was actually done thanks to the special features of these duplicator drives. I don't know that. I am saying that the argument "couldn't have been done because they used standard drives", doesn't hold.

No. We can see if a track was written in one pass or if there are any splices or the like.

Not necessarily. You see splices only if recorded data overlaps in two (or more) passes. But this case is special, it is not normal data. It is normal data overlapped with no flux transitions (whatever the cause for this is).

And actually, depending on the disk, track and drive, I can see lot of splices on these protected tracks. Some disks much more than the other, which is interesting, and suggests that perhaps this process (again, whatever it was) was improved at some point.

Another interesting fact is that the area without flux transitions is not exactly aligned on every case. Some tracks have more bits (more normal transitions) between the last sync mark and the start of no flux area.

... suppress flux reversals with a clever trick DrCoolZic explained above.

Speaking about his comment, that transitions might be too fast for the read channel to detect them. I am not convinced that the behaviour here is an effect of the read channel, as he posted. As I commented to him by PM, I think that the write channel is involved here as well. An interesting test would be to read a disk fooling the HD detector somehow.

I just don't feel I should keep silent when I see someone "talking" about KryoFlux and how it works that hasn't the deeper knowledge of the product.

That is not true. I didn't say here anything specific to the Kryoflux, and certainly nothing that requires a "deeper" knowledge of it. What I said was generic (and I actually said explicitely it was generic), to any similar controller like the Discovery Cartridge or the Catweasel as well. I might have been wrong, and I might still be wrong. But if I am wrong, then I am as wrong regarding the Kryoflux as I would be wrong (in this matter) regarding the DC. And if I was wrong all the time, then it wasn't because of lack of deeper knowledge of the Kryoflux.

[DrCoolZic]

... suppress flux reversals with a clever trick DrCoolZic explained above.

Speaking about his comment, that transitions might be too fast for the read channel to detect them. I am not convinced that the behaviour here is an effect of the read channel, as he posted. As I commented to him by PM, I think that the write channel is involved here as well. An interesting test would be to read a disk fooling the HD detector somehow.

However it seems that one working solution is to write an area with a relatively high frequency signal. But the reason why it works is indeed not related to the bandwidth limitation of read channel.
This is (almost) explained in the 2 sections that I have added on FD http://info-coach.fr/atari/hardware/FD-Hard.php#reading

[mr.vince]

Ijor, I think the major point is that we read it, analysed it and wrote it. We do have a working chain. What more would I want to demo?! We (IFW and me) worked in the games industry in the 90s. Now you can believe this, or don't, or can continue debating. I already emailed Rob and if I manage to get a reply, I'll post it here. I will also get you both in touch, so you get first hand information.

I will also call the guys in the pressing plant next week and see if they still have someone around that did floppy replication. Maybe they still have manuals or parts; although I doubt that.

Enjoy. It's an awesome fall over here today.

[Hippy Dave]

How To -- Write with flux reversals suppressed (degauss) midway through sector:

Write some data, then write a high frequency signal, then continue writing more data.
Simple.

The erase signal should be of a frequency high enough to put a few cycles within the head-gap at the speed the media is moving.

[ijor]

Ijor, I think the major point is that we read it, analysed it and wrote it. We do have a working chain.

I agree completely, you do have a working chain. That was never the debate, and I never argued about that. And I actually think it is great. In one of our last emails I already said something like "It is very good that you can write back Turrican", didn't I?

We (IFW and me) worked in the games industry in the 90s. Now you can believe this, or don't, ...

Of course I believe that. Again, I'm not sure what relation this has with the "debate" ???

I already emailed Rob and if I manage to get a reply, I'll post it here.

Thanks, it would be very interesting if he replies. Appreciate it.

He might not remember the details, this happens many times in these cases. You can see in his interview published long ago, that he already was completely wrong about the modified bitrate of the "copylocked" sector on the ST. But I'm sure it would be interesting anyway.

How To -- Write with flux reversals suppressed (degauss) ...

This is not degauss. Degauss is technically a different magnetic condition, the one that disks (except pre formatted ones) came from factory. A degaussed disk reads back as noise (random transitions).

[Hippy Dave]

Q) This is not degauss...
A) The definition of degauss is well known, I'll leave it as an exercise for the reader.

Here is a of what happens:

If writing a one to the media results in a north pole on the media, then writing a zero results in a south pole (or vise versa). The read head detects change in flux to produce voltage. The biggest change is at the flux transition (reversal in this case) from north to south, and south to north.
Writing an erase signal (degauss) results in demagnetized media (no north, no south) and noise is superfluous in this case because zero flux is halfway between north and south.
Note that there are THREE levels of flux AND that FLUX TRANSITIONS occur when you change between the levels. Imagine what you can do with this knowledge!

[DrCoolZic]

The definition of degauss is well known, I'll leave it as an exercise for the reader.

I have found this http://en.wikipedia.org/wiki/Degaussing ... rage_media

[ijor]

Writing an erase signal (degauss) results in demagnetized media (no north, no south) and noise is superfluous in this case because zero flux is halfway between north and south.

Dave, we can argue about the magnetic properties. But it is a fact that when you read a demagnetized track (or disk), you get random transition pulses from the drive. You can argue they aren't actually real magnetic flux transitions, but an artifact of the read channel. It doesn't matter for our purposes, what matters is that the controller gets those pulses. And this is different than the protection we are talking here, that doesn't produce any transitions pulses.

[Hippy Dave]

Writing an erase signal (degauss) results in demagnetized media (no north, no south) and noise is superfluous in this case because zero flux is halfway between north and south.

Dave, we can argue about the magnetic properties. But it is a fact that when you read a demagnetized track (or disk), you get random transition pulses from the drive. You can argue they aren't actually real magnetic flux transitions, but an artifact of the read channel. It doesn't matter for our purposes, what matters is that the controller gets those pulses. And this is different than the protection we are talking here, that doesn't produce any transitions pulses.

I agree. I believe there is more misinterpretation than disagreement here.
The voltage on the read-head cannot be DC (other than close to zero volts) and the digital circuitry will read either one or zero from this weak signal, producing random transition pulses.
Note that a weak signal close to zero volts will be produced at the read head whenever the flux stays constant at any level (saturated north or south or at any level between), it is the change in flux that produces voltage.
A constant flux track can be written with digital circuitry by writing a constant one, or a constant zero, or a high frequency string of ones and zeros (This may or may not be possible on the Atari ST).
If a constant flux track is made by writing a constant one level after a zero, reading this track will show the 'zero', then the 'one', and the voltage producing the 'one' will decay towards zero volts linearly in time and in direct proportion to the read head inductance (and the circuitry may make random transition pulses eventually).

[mr.vince]

Ok, coming back to this "strong DC erasure" thing. Recording theory says you can't record this unless you use a permanent magnet. Recording theory also says you can't read DC, as only AC generates an alternating field, thus inducing a current.

If this is correct, the "strong DC erasure" protection would generate no signal in the read head. No signal would mean the drive turning up the "volume" (AGC) which should pick up random noise only. The effect should be a track full of noise, which means sporadic flux changes. Under no circumstances should it be possible to generate "no flux" areas by using "strong DC erasure".

So what do you do to deliver proof? You build a "strong DC erasure" drive. This is what our friend Herzi over at a1k.org did. It's a modified floppy drive that has both heads replaced with two strong neodymium magnets. The electronics have a small circuit added that will monitor the disk insertion switch. As soon as it is activated, the motor will spin and step from 0 to 83 and back. Guess what... this erases the disk.

Drive electronics:

eraser_unten1.jpg

Head replacement:

eraser_kopf_unten.jpg

Ok, and here's a dump of a few cyclinders in STREAM format made with KryoFlux. I also made a screenshot of the GUI to show how this is looking:

dcerasure_scatter.jpg

I am sure DrCoolZic will take a look and provide us with some graphs, too.

My conclusion is that "strong DC erasure" is great for erasing floppy disks, but it does not create "no flux" areas.

[DrCoolZic]

This is interesting! The track seems to contain random values but different from unformatted track.
It is also interesting to note that we do not find any flux transitions below 3µs ??? This may be due to the read channel that filter transitions below this value ???