Loedown

Greetings,
This thread is not for the faint hearted!

I was doing some research today on floppy drives, read about their signals and alike, but there's several points which I couldn't get proper answers for, so if anyone can answer them, I would be grateful.

1. Modern FDD contain a magnetic sensor for drive revolutions and sector starting points, what signal comes from the drive to signify that the index point has been reached?

2. Data is read or written to the disk as serial data, but unlike RS232 which can have start / stop bits, what signifies to the controller that it has indeed written or read the right data, what does it use as a syncronising indication, as the drive is only as good as its last revolution / index read and the precision of the 300 RPM / 150 RPM?

3. How does the chip inside the Amiga determine if indeed it has to read a long track, surely it has a set amount of data it 'expects' from one track to another, or does something signify that it will receive additional bits, is there something written in the sector header?

4. Is track 82 the absolute most upper track ever written?

Paul

Advertisement:

dlfrsilver

track 83 is the most upper track written.

Example : Scooby doo and scrappy doo from Hi-tech. Most of their games are using track
1 to 83.

__________________
Denis

Loedown

Then I can assume you mean 0 - 83, or do you mean 0 - 82, based upon the logic of calling the first track, track 0 as track 1?

Paul

prowler

It's 0 - 82, Paul, (or 1 - 83, as Denis says above).

IFW

1, index
2, self-clocking formats, data separator, address mark detectors
3, short version: cell width changes and the hardware adjusts to cell width
4, you could write 0...83 on some models, 0...81 is more common. Anything from 80 is unsafe but generally should work until 81.

Loedown

Point 2, what determines the clock, is there a specific signal from the drive to indicate that the sector end has been released, what is the signal for data separation and how does the controller determine the address marks?

Could I have the long version of point 3, please?

I know I'm asking plenty of questions, but I believe I may be able to make a perfect copy of any floppy, regardless of the format, including all errors.

Paul

IFW

You are just about to enter Wonderland...!

1, Self-clocking formats always have a clock pulse embedded helping to adjust to the current clock rate. That and the reference clock together aligns the data.
A data separator separates clock signals and data signals from a single stream of mixed clock and data - as it is recorded on the disk.
The logical layout of data is a completely different thing.
Address marks are being detected through various methods, it depends on the design. Usually data that cannot be recorded by normal encoding rules is the address mark. Controllers have various components that can detect this and once such mark has been encountered determine which part of the data is clock and which is data in the case of MFM recording.

2, Long version: see above. The keyword is continuously adjusting the data window to the clock speed derived.
This is necessary as neither drive speed is really constant, nor cell width is for real.

3, er... good luck with that. Yes you can, but data you should write to a disk is quite often not the data you read from the disk...
You can make an analogue copier that way that works reasonably well and quality degrades depending on the location on the disk as well as the source material.
But this has been done 20 years ago...

Loedown

I had something a little more radical in mind. Using a BBD and PIC, if I can read data off a disk exactly as it's laid down, then as long as the destination drive is perfectly synced, data can be put back exactly, this way avoiding the whole weak bits / long tracks issue, if the controller simply doesn't care what format the disk geometry is apart from having designated tracks, then in theory, the whole thing should work.

Time to crack out the data analyzer and breadboard.

Paul

Advertisement:

IFW

Yes, that's what analogue copiers do without buffering the data.
Again, what you read is not always what you need to write to read back the same data...

Loedown

Greetings,

So similar to having like a sync separator in TV transmission, you need to know what to strip out to write the data back correctly?

From what you've told me, IFW, I will still go ahead and start experimenting, as this is one of the biggest bugbears that ancient computer operators face and to be more honest, I have more interest in most, copying proprietry disks, ask me about Roland MC-300 system / library disks.

Paul

IFW

You need to know how to change the data so that after writing it would read back as the expected values.

IFW

An old cyclone or similar hardware assisted copier would do the exact same thing for you.
As long as it's a first generation copy you should be fine.

Loedown

Greetings,
Now this becomes really perplexing, why would a disk not contain exactly what was written to it, why does it need to be modified to be understood? Does this have something to do with MFM, are drives so low level that the conversion of MFM isn't done by the FDD controller?

I would still question why an analogue approach wouldn't work, as even if the data coming off the reading drive is encoded, as long as it's written back to the second drive EXACTLY as it was, then why doesn't this method work in all cases?

I know I'm a pain in the arse

Paul

IFW

Please do some research on magnetic recording and pre-compensation. Basically, certain sequence of cells are known to take a different position from the intended one. Now if you read that back, it can get to the point that it gets different values from the ones you'd read from the original source.
You need to compensate for that.

Loedown

I will, magnetic media was but a short course many years ago, I just recall some of the terms and have a vague idea about how it works.

From C64 days, I recall that certain tracks had certain sector counts, but in the case of more modern drives, they used a fixed set of sectors, perhaps not so much the Amiga, most likely the reference to long tracks.

IFW, you've been a great source of information, thankyou.

Paul

Advertisement: