What is it?Create GCR-images of a floppy by reading the data (almost) directly from the surface itself.
C64-emulators use D64 files as disk images of original C64-floppies most of the time. Such an image only contains the data of a floppy. A real floppy does not only contain data but also things as blockheaders, dataheaders and SYNC-characters. To make it even more difficult, the data itself is GCR-coded. But sometimes you need ALL the bytes from a disk. Most times in case of a disk with an copy protection.
Creating such a 1-to-1 disk image can done by connecting a 1541 directly to a PC using a X1541-cable or, much faster, a XP1541-cable. But it still involves some very intelligent ie. difficult programming of the 1541. Maybe such a program already exists but then I still wonder if it is capable of handling all the devilish copy protections that I know off.
So the idea raised to use a way of reading the disk nobody has thought about before (AFAIK): reading the data directly from port A of 6522#2 using the LPT-port of a PC!
HistoryThe idea raised while working at PC-Flop . This project has as goal to let a PC emulating the actual floppy drive: the mechanical drive and some electronics is removed and replaced by a cable connected to the LPT-port of the PC. Instead of storing the data on a floppy disk, the data is stored by the PC on the hard disk drive. This is done in such a way that certain features used by copy protections are covered.
The normal procedure of copying files would be loading files from disk using an original 1541 and storing them on the PC-Flop-drive. The needed time depends on the PRG used to perform the job.
If you read the document about transferring data from/to a floppy disk, you'll see that all the data passes port A of the 6522. So my idea was to read the data directly from the pins of port A by the PC.
TheoryMy very first idea was only using a cable. But during the process of writing this document, I discovered that that most likely was impossible because of two problems:
- The distance between two bits varies from three to four uSec. This is also the time the data on port A is valid. In this time the PC has to detect that the data is valid and to read it then. But there is no guarantee that this is the valid byte. So I needed to latch the data like the 6522 does. A 74ALS574 latches 8 bits whenever a positive edge occurs on its clock input. The "byte ready" signal is a negative edge so we need to invert this first.
- Then I found out that SYNC-characters don't generate a "Byte ready" signal. As long as there are 10 ones in a row, UC3 won't count as it is constantly in a pre-load state. At this moment I only have a software solution.
The changes I propose have no influence on the normal functioning of the 1541.
Realization (partly theory for the moment!)First we need an inverter. We could add another IC but studying the schematic of the 1540 reveals that UC7, a 74LS02 NOR, has a free gate: #1, pin 1, 2 and 3. Disconnect pin 1 and 2 from ground and we have our inverter.
The 1541 (#1540049-02 rev. B) has two free inverter gates: UA1 #4 (44LS14) and UC5 #3 (74LS04). I don't know if the inputs have been connected to GND.
At this moment I cannot tell you if the boards of the 1541C and 1541-II have any free inverter gates. If not, then we have to add a 74LS04 and to use one of its gates.
Whatever inverter gate we use, its input has to be connected to the "Byte ready" signal at pin 40 of 6522#2.
We add the 74ALS574 to the board by piggybacking it on top of 6522#2 by connecting pin 2..9 of the 574 with pin 2..9 of 6522#2. Pin 20 is connected to +5V and pin 1 and 10 are connected to ground. Pin 11 has to be connected to the output of the inverter mentioned above. Pin 12 to 19 are connected to the cable leading to the LPT-port.
The following pins of 6522#2 are connected to the cable as well:
6522 LPT Name Meaning pin name pin ---- ------- --- ---------- --- CA1 BRDY 40 Acknowl. 10 (I) PB7 SYNC 17 Busy 11 (I) PB0 Stp1 10 Select 13 (I) PB1 Stp0 11 Paper out 12 (I)
The softwareWe need software for the C= host as well as for the PC.
The PRG for the C= can be a Basic program which sends a small ML-program to the 1541 and commands the drive to execute it. The ML start the drive motor, positions the head above track 18 and then above track 1, waits for two seconds and then steps to each (half-) track with intervals of a second. When at track 41 (???) it positions the head above track 18 and stops the drive motor.
The PC waits until the head starts to move and then stops moving for at least one second. This is the signal for the PC to save the data read at port A into a 16 KB-buffer the moment the "Byte ready" signal occurs.
And now the SW-solution to determine how long the gap with SYNC-chars is. The time between the moment the SYNC signal becomes active and the first "Byte ready" signal is measured. We know the time between two bits so the number of bits can be calculated. We have to subtract the value eight for the last byte and the number of ones the last valid byte was ending with. The resulting number is divided by eight and we have the number of bytes.
Realization (the practice)Unfortunately I haven't good news until so far. I started with a simple BASIC-PRG on my C64 that does nothing else poking values into the memory map of the 1541 using the "M-R"- and "M-W"-command. This way enabled me to start/stop the drivemotor, to change the track and to change the "Density Select". A small ML-PRG did nothing else then monitoring the BREADY line and putting the read value in a 16 KB-array. The problem is that I simply do not recognize the data the PC has read. I know the header of every sector contains five $55 bytes in a row. I simply don't find them :( So what is going wrong ????
(to be continued....)
You can email me here.