DOS, Disks, Density and Sector counts... education question.

[flashjazzcat]

Speaking of mind tricks... your 1xSIO speed question was ignored. I guess you should read the XBIOS documentation.

[Gunstar]

Speaking of mind tricks... your 1xSIO speed question was ignored. I guess you should read the XBIOS documentation.

Documentation...documentation...where did I put that documentation?

[Kyle22]

IM(not so)HO, In the year 2018, why can't people write things that work across multiple solutions?

 

For Example: If you write a game that fits on a single disk, have it available in the directory as file(s). That makes it much easier for those with high speed devices to use it.

Don't use illegal opcodes. There are many of us now who have C02, 802 and 816 chips. If the game is multi-disk, make it hard disk friendly, or at least compatible with a large ATR.

 

Come on guys. Let's make things user-friendly.

 

[+Nezgar]

Though on the flipside, early 154x drives were so bad that they only used the first 35 tracks.

The SA390 drive mechanisms used in the early Apple Disk ]['s were physically incapable of stepping past 35 tracks due to a hardware design limitation. A small modification to the Apple DOS supposedly let it use 36 tracks, but 3rd party mechanisms could do 40 tracks, again needing non-standard DOS modifications.

 

And for the Commodore 1541, according to wikipedia, "although the 1541 DOS only uses 35 tracks, the drive mechanism itself is a 40-track unit."

 

And according to c64-wiki: "Access to tracks up to 40 or 41 is physically possible, but the standard DOS doesn't use them for data storage. Some copy protection schemes and non-standard DOS versions do use them."

 

https://www.c64-wiki.com/wiki/Commodore_1541

 

I think it was a carryover for compatibility with previous drives for the PET like the 4040 that used same SA390 mechs that had the hardware 35 track limitation.

 

So, based on that I don't think those tracks were not used due to a data reliability concern, although it is true that data there is more likely to develop errors due to the increased flux density.

[thorfdbg]

Yes, the 1050 writes at double density, only that with a smaller sector size. Why Atari did that, I don't know for sure.

Even though the question is old, here is the answer - it is quite trivial: The 1050 does not have enough internal RAM to buffer a 256 byte sector before sending it to the host. In fact, the 1050 has 256 bytes(!) of RAM, half of which is used as stack and for zero page registers, and the other half is used as sector buffer. Operating without any RAM is quite hard (though some third party products were able to write 256 byte sectors without using any external RAM whatsoever - a real art!)

[+kheller2]

Even though the question is old, here is the answer - it is quite trivial: The 1050 does not have enough internal RAM to buffer a 256 byte sector before sending it to the host. In fact, the 1050 has 256 bytes(!) of RAM, half of which is used as stack and for zero page registers, and the other half is used as sector buffer. Operating without any RAM is quite hard (though some third party products were able to write 256 byte sectors without using any external RAM whatsoever - a real art!)

 

 

This almost makes me think a reason why they didn't release the 1050 cable of 256b DD was because: 1, the extra RAM chip was probably expensive, and 2, they didn't have a dos that could work with it. Oh wait... DOS2.0d could have been fixed. Hmmm..

[+DrVenkman]

 

 

This almost makes me think a reason why they didn't release the 1050 cable of 256b DD was because: 1, the extra RAM chip was probably expensive, and 2, they didn't have a dos that could work with it. Oh wait... DOS2.0d could have been fixed. Hmmm..

 

The extra chip was probably less than $0.50 in quantity even back then. But remember, this is the same idiot-led company that developed not only one but TWO completely unique consumer Trak-Ball game controllers compatible with the 2600, A8 line and 7800, but released only ONE game ever for any platform that could even use it! It also had an excellent chroma-luma video signal in every A8 machine since the 800 but omitted a single damn trace on the PCB to the DIN jack on the 1200XL and 800XL [for reasons].

 

NOTHING that company ever did after maybe 1982 made a lot of cohesive sense when viewed on a timeframe longer than about 6 months, it seems to me.

[thorfdbg]

The extra chip was probably less than $0.50 in quantity even back then.

 

That is not quite the point... The RAM in the 1050 does not come from RAM chips, but from two 6532 RIOTs with 128 bytes RAM each. For a 256 sector buffer, one would not only require one additional chip, but also a complete re-layout of the PCB.

[+Nezgar]

Even though the question is old, here is the answer - it is quite trivial: The 1050 does not have enough internal RAM to buffer a 256 byte sector before sending it to the host. In fact, the 1050 has 256 bytes(!) of RAM, half of which is used as stack and for zero page registers, and the other half is used as sector buffer. Operating without any RAM is quite hard (though some third party products were able to write 256 byte sectors without using any external RAM whatsoever - a real art!)

 

Behold the 1050 Turbo, the idea put into practice - ROM ONLY upgrade for the 1050, achieving true double density using only the original 256 bytes of RAM and a large 32KiB (i think?) banked ROM, and bonus features like faster-than ultraspeed 71428bps I/O (but not ultraspeed protocol compatible), protected disk copying, automatic slow-RPM control (for writing more than 18 sectors copy protection), and a parallel printer-interface! (in the original only, not in the reimaged version)

Reimaged version is available here: http://atariage.com/forums/topic/277204-some-new-not-really-hardware-made-by-tfhh/

 

This almost makes me think a reason why they didn't release the 1050 cable of 256b DD was because: 1, the extra RAM chip was probably expensive, and 2, they didn't have a dos that could work with it. Oh wait... DOS2.0d could have been fixed. Hmmm..

 

Yeah.. Atari already made DOS 2.0D by June 1980.. And even DOS 2.0S worked in double density if copied to and booted from a DD disk. The main thing missing was an option to switch densities, or automatically switch density. PERCOM's patched DOS 2.0P added the density change option, and SmartDOS added automatic switching.

[+DrVenkman]

 

That is not quite the point... The RAM in the 1050 does not come from RAM chips, but from two 6532 RIOTs with 128 bytes RAM each. For a 256 sector buffer, one would not only require one additional chip, but also a complete re-layout of the PCB.

 

No, that's EXACTLY the point. The 1050 was a clean-sheet design and required a new PCB ANYWAY. In fact, there's more than one PCB design for 1050 drives. The marginal cost of adding a single chip to the net production cost of the already-new design would have been utterly lost in the noise, especially given how many RIOT chips Atari was already buying literally by the millions already just for 2600 production.

Edited April 11, 2019 by DrVenkman

[Kyle22]

This is what happens when you change from people who care about and love their work vs Corporate idiots that only care about the ROOT OF ALL EVIL.

 

Put an extra 128 byte RAM in there. You cheap ass. Where you from, India? How CHEAP can you be? That is the question.

 

Ruthless 'businessmen' trying to shave every penny off of the cost. NO concern for product quality or Public reaction.

 

Atari made QUALITY before they bent over to the FCC and Warner. Later on they allowed the Tramiels to destroy the company.

 

Shameful.

 

I am hoping that the new VCS is gonna be a spectacular seller. I want this thing to put M$ and $ony under.

 

[ricortes]

I thought i looked it up, a 6116 2K RAM was about the same price as the 6810 128 byte one when the 1050 came out. They probably just had a lot of 6810 on the shelves and decided to just use them. I mean when the ST came out they went with what were obsolete single sided 3.5" floppies! Ditto for the choice of using a 6507 instead of a 6502, lots of 6507 around from the 2600 production.

 

But I do like nutty conspiracies! If we had a drive with 2.13k RAM, all it would have taken was to program the drive to take another command that would translate to load RAM and RUN. From there we could have done a copy protection map upload to the drive and had it return the correct errors for certain sector requests for example.

[+Nezgar]

Further to design / management shortsightedness.. In the ANTIC Interview 220 with Scott Scheiman, who was one of the designers/programmers of the 850, he mentions he originally wanted a UARTs and 25 pin serial ports, but that was shot down in favour of 2 of their plentiful VCS 6532 RIOT chips, so the software spent most of it's time in a polling/setting loop to get the serial data through... He also suspected they had a higher power requirement, but I'm not sure about that...

[thorfdbg]

UART or not... this would not have made much of a difference. The major limitation of the 850 is that you cannot have incoming serial traffic and outgoing SIO traffic at the same time, i.e. one cannot store received serial data to disk. It would have needed additional RAM on the 850, and usage of the "interrupt" line of the SIO bus to signal that this buffer is ready to receive by the host. That would have been a somewhat more complicated design, though: Additional on-device memory (probably 128 bytes, better 256 bytes), and a smarter communication protocol.