Catweasel Driver / Updates

[Loedown]

Jens is posting a fair bit again, so I wonder if he could answer a couple of enquiries about Catweasel?

1. Will new software be made available for the multiple issues that people have commented on in various threads and is there any chance of getting Catweasel to copy HD disks / protected disks?

2. What's the opinion of Kryoflux, clearly a competing device?

[8bitbubsy]

3. Will there be any 64-bit driver for Windows 7?

[papa_november]

64-bit drivers were promised over a year ago and nothing has happened since.

[Schoenfeld]

Originally Posted by Loedown

Jens is posting a fair bit again,

...which does not mean that I'm giving support for my products in a public forum. If you have a specific question like this, your way to go is the usual support chain: Ask your reseller, and if he does not have an answer, then my employees will be included in the loop. If that second step is not successful either, I'll answer the question. No part of this chain is happening in a forum for legal reasons (mainly our obligation to archive customer communication!).

I'm here on a private basis. Individual Computers does not give support in any forum. I may or may not read a request like this. Better assume I'm not reading it, given the workload I've had lately.

Originally Posted by Loedown

1. Will new software be made available for the multiple issues that people have commented on in various threads and is there any chance of getting Catweasel to copy HD disks / protected disks?

64-bit drivers for Win XP64 and Win7 have been in the works for quite some time, but the Microsoft 64-bit DDK keeps giving us headaches. We've had help from companies with full Microsoft support (which we don't have, as that's way too expensive), but even they haven't been able to resolve the large number of problems that the DDK is giving. Development is not stalled, but development time is unpredictable at this point.

Originally Posted by Loedown

2. What's the opinion of Kryoflux, clearly a competing device?

I won't comment on this. It has gotten ugly in other forums due to false claims and wrong assumptions about their device, and I don't want to get into this here.

Compare speed, compare price, compare features, then make your decision on what to buy.

Jens

[Dic_Ray]

Originally Posted by Schoenfeld

I won't comment on this. It has gotten ugly in other forums due to false claims and wrong assumptions about their device, and I don't want to get into this here.

Compare speed, compare price, compare features, then make your decision on what to buy.

Jens

I really admire your creations for the whole Amiga community. These days, the Amiga market is a very small special interest market and so every developer must fight to survive. So I relly comprehend, that every perhaps a little bit frivolous allegation voiced claim can endanger the future of your business.

But at the moment I can't decide for your great Catweasel, cause I only have computers without PCI slots and so I'm dependent on an USB solution to read out all my old data floppy-disks. I wanna do this for the last time and so I'm excited about KryoFlux.

I am sure, that the Amiga market, does not need this type of grave conflicts in the least. So how do you think about this or even would you see a little chance to cooperate with the SPS/KryoFlux developer-team?

Essen would be a great place to start a cooperation, wouldn't it?

[Schoenfeld]

Originally Posted by Dic_Ray

Essen would be a great place to start a cooperation, wouldn't it?

It would at least be a good place to apologise about making false claims about the Catweasel's capabilities. While SPS has always been claiming that the precision of the Catweasel is not high enough, the reality is that the Catweasel's precision goes beyond the precision of the floppy interface. This is even true for the old ISA version, which can still be bought at a very low price. However, I wouldn't like to support their hardware, as it's way too slow to be considered "viable" - waiting over two minutes for a disk to be backed up is like going back to the old datasette days.

Jens

[IFW]

Er... reading a disk is limited by the drive speed. Unless you artificially increase the drive speed there is no way of reading disks faster than the maximum data rate.
A 8-bit counter (or 7 as the index signal is bundled, can't remember) for flux transitions (CW) is not very helpful, when you need considerably better, unless the resolution you sample is not too high.

[IFW]

You need good precision to analyse data, especially that is suspectible to bit shifting.
You don't necessarily need that kind of precision for everyday use, such as backing up demo disks.

[IFW]

FYI: KryoFlux reads disks at the rate the data is read by the drive.
It won't get any faster than that for sure - but feel free to increase the speed of the drive attached... it will automatically adjust itself, and will indeed ingest the same disk twice as fast say from a 600RPM drive vs a 300RPM drive.

[IFW]

The current KF hardware has a (soft) limit of 32 bit counter at about 40ns per tick.
Signalling is NOT bundled or mixed in any way with counters - this is the true resolution.

A next generation KF device will have an even better resolution.

[Schoenfeld]

Originally Posted by IFW

The current KF hardware has a (soft) limit of 32 bit counter at about 40ns per tick.

Catweasel = 35ns.

Jens

[Schoenfeld]

Originally Posted by IFW

FYI: KryoFlux reads disks at the rate the data is read by the drive.

Catweasel reads an Amiga disk in less than 50 seconds.

Kryoflux advertisement video on youtube:
[ Show youtube player ]

Check the video at 1:37, where the speed of Kryoflux is advertised with 2 minutes and 12 seconds per disk. Also, you advertise a C64 disk in 38 seconds. The Catweasel reads it in less than 11 seconds.

I can see that you're reading the single track at rotation speed, but your slow USB interface limits you. This is where you find the advantage of PCI.

Jens

[kipper2k]

I wouldnt care if it took 10 minutes to copy an ipf or one of my original discs as long as it produces a 1:1 copy. Adf's can be copied on any Amiga that can load an adf write utility so that is not a concern

[mr.vince]

How pathetic. The predecessor to KryoFlux - Cyclone20 - was made because Rich Aplin and me were missing some clever modern hardware that would read and write whatever disk you'd feed it with. Why would we - and lateron Softpres (which I became involved with earlier that year) - spend hundreds of hours on development of a different solution if we could just mail-order something for a few bucks?

@Jens: I understand these are sad days for you and your product. I offered you to join the party when C20 development started. You turned us down. Ok for me.

But: No one here is apologising for not buying, using nor praising your product. You know I once bought a Catweasel personally, was not satisfied, sold it and decided I'd go for an alternative. What more should I say? You said the customer should compare and then decide. Now don't be embarrassed if he does.

I have respect for your work and for many of the designs you have made. But you aren't god. People are allowed to have other hardware designers beside of you. Get a grip, reduce your ego, let's have a beer and talk in Essen. But don't expect anyone to knee down. We're adults.

[Schoenfeld]

Originally Posted by mr.vince

@Jens: I understand these are sad days for you and your product.

Actually, it's not. All the fuzz that you're making got people to compare and buy a Catweasel MK4plus. Sales have picked up between february and june of this year, so all your online marketing work got people interested in the topic and found the solution that's already on the market and that has a 14-year track record of soft- and hardware updates.

I have no problem with the people buying Kryoflux. The only problem I have is you violating EU regulations about directly comparing advertising and making false claims about my hardware. Anywhere you're writing, you imply that the Catweasel's hardware is inferior to your microcontroller-based solution. However, if you'd care to follow EU regulations about comparing advertising, you'd have to admit that your superiority is located in connectivity only. You lack speed, you lack write support and lots of disk formats aren't there yet. This is why people buy the Catweasel.

I have to say thanks for that, and.. yep, let's have a drink in Essen, you're doing great work for my products.

Jens

[IFW]

Nope, it's not the USB speed, since the track data must be sent over while it's being read.
However a track should be read (at least) twice if you want to have all the samples possible without interruption no matter where the very first data starts on a track.
It's easy to see why, e.g. if the data writing started immediately before the index signal you'll need 2 complete revolutions to sample that no matter what, otherwise the data will be broken at one point.
That, and the protocol being 100% reliable.

It is possible to read only one revolution only if you are expecting a certain format to be read, such as say Amiga, not anything - while this is not an issue for CW it is for KF; as they have very different purposes.

KF could be made to run even faster by having format specific transfer methods, but it's a complete waste of effort.

I can't see any advantage of PCI (there is a slight advantage of practically no protocol overhead), especially as it is a legacy interface soon, does not work on my laptop and it's next to impossible to support the drivers for PCI unless you have an army of sw engineers who work on Windows certification, signed drivers in kernel mode and the like.
BSoD, intermittent and undebuggable driver problems etc are difficult to manage when dealing with PCI.

35ns is fine, too bad there is only a 7 (8?) bit counter for that.
I think you have to set the sampling rate as well in advance, and probably 35ns is not one of the rates you want to set given the counter resolution, unless expecting to sample a HD/ED disk.
Knowing in advance what and how to read is fine for standard disks though and a good shortcut - but dealing with standard disks is a side effect of our product, not the result.

KF always samples at the highest resolution possible, no shortcuts taken that are specific to any format - you can always go back to a dump (if you kept it) and see all the original flux transitions as if you were just reading the disk.

As I said, very different purposes and use cases for the devices.

[fiath]

Jens, I'm sorry that you feel you need to keep up these accusations, now here, after your posts in other forums.

I want to let it be known that we have very much gone out of our way to avoid publically detailing why we could not use the Catweasel, as we got asked all the time why we couldn't. I think the section on our website we had about it was extremely nice about it actually! However, since you still persist with this, even after we have tried hard to ignore you.

I seem to remember we told you why we could not use it years ago. Instead of taking our comments as suggestions for improvements as they were intended, you started being rather rude, and sadly never implemented the features we wanted. So unfortunately we could never use Catweasel - and we would have done if we could! (why would we not?). It's not exactly our problem that we finally went and made something suitable for our needs... It's not like we ever wanted to have to do that.

We always avoid talking about the competition, I don't really understand where all this "calling Catweasel inferior" comes from. Sure we think our solution is brilliant, and we want to promote it - why do you think this is about you? Maybe you are confusing what we write, with what other people have written. People are bound to want to compare the two.

I'm glad your sales are picking up anyway.

[IFW]

I am afraid we have different markets and interests - with some overlap.
Also we are not the ones posting FUD, nonsense and downright ridiculous claims about competing products - especially, as our aim is very different with KF.
Our microcontroller based solution is well, technically, microcontroller based, as much as say mobile phones are...
It's fast, super reliable, can write... and...
As for disk formats... er... you are barking up the wrong tree, remember that is our interest.

As for advertising one quick example from CW, but there are countless other examples - I'll let people see for themselves.
http://www.amiga.org/forums/showthread.php?t=51582
"I have been instructed to inform you that this project endangers one of the Amiga community's few remaining hardware manufacturers. Please for the sake of the community as a whole discontinue this project. "

We never, ever said anything about CW.
Why? Because we wanted to keep your business interest healthy and consider doing otherwise very unprofessional.
It's that simple for us.

[Schoenfeld]

Originally Posted by IFW

35ns is fine, too bad there is only a 7 (8?) bit counter for that.

Another false claim. You obviously don't have access to the Catweasel MK4 programming documentations. Yes, I keep them restricted, but this also puts me into the position of knowing that you cannot make this claim (why else do you think I'm working with such a big FPGA on the board? The counter had this limit on the ISA version only).

Now for some theory:

Disk drives have been working with a 4MHz local clock for ages (I have 8 inch drives here that use this clock). This clock is used for driving the motor and aligning flux transition pulses. 4MHz gives a cycle time of 250ns. If you want to sample the outcoming data without a loss, you could either do brute-force and oversample "all you have", or you could use some brains and view that signal as band-limited. This lets you use the Nyquist-Shannon sampling theorem, and that says that anything beyond 8MHz (125ns) essentially gives more data, but does not carry any more information.

That's why anyone with 125ns clock or better can claim to read "any disk format there is, no matter what". The drives are the limiting factor, not the controller.

Now once and for all: Quit making false claims. You can't break the laws of nature. You can't even bend them. However, you're breaking EU laws by misleading customers.

Jens

[Schoenfeld]

Originally Posted by IFW

"I have been instructed to inform you that this project endangers one of the Amiga community's few remaining hardware manufacturers. Please for the sake of the community as a whole discontinue this project. "

You're putting this post in a context that lets people think I instructed someone. I did not. However, I told your German guys that I don't like the way you're putting your press release, as the first version read like "we're the first to read&write Apple 400k/800k disks". This is simply not true, as the Catweasel has done this since 1996.

Your press release was not changed until I had to quote laws. So now you want to put me in a bad position by defending myself against false claims?

You're not exactly looking good by trying to attach the above quote to my name. Your knowledge of signal theory (or lack thereof) doesn't make you look much better. Yes, I turned down your suggestions, because they don't make sense. You just don't get more information by increasing the sample rate.

Jens