Second Disk drive
From Grahams Wiki
The +3 can take a second disk drive. In true Amstrad Style its Almost standard.
Actually, to be fair, apparently you can plug in a normal Shugart disk drive by swapping some pins around.
Unfortunately, some time in the 1990's some muppet decided to make PC floppy drives slightly different, however, with some modifications, and a bit of extra circuitry, they can be fixed.
Contents |
[edit] Differences
The differences between Shugart and modern FDDs are only very minor, however the main one that affects the +3, is that the +3 is expecting a disk-ready function, while a modern floppy disk drive supplies a disk-change signal. Fortunately, this is quite easy to remedy. A small change is required to the floppy disk drive however.
In addition, all modern FDDs are wired to Disk 1 rather than Disk 0, however, this can be fixed by wiring in the cable.
[edit] disk change
The disk change basically pulls the ready line low if there is a disk inserted and the disk is inserted. If this isnt connected, the +3 will either end up in an endless loop if a disk is not present, or will just insist you insert a disk.
As the line must only be pulled Low, you must use Open collector gates. Eg, the 74LS38. Dont use a normal TTL gate (EG, the 74LS00) as you will break something, probably the internal disk drive, as it tries to force it low, while your gate is trying to force it high.
You also need to make a modification to the drive to extract the signal for when the disk is inserted.
[edit] Changes required to the FDD
Most disks appear to have four switches at the front of the disk drive, Eg,
| | |[]A C[]| |[]B D[]| +------------------------+
The switches have the following jobs:
- A Disk inserted
- B Write protect
- C Disk inserted
- D High density
Some disk drives will only have one of A or C, but all will have one.
The first thing to figure out is which switch is in use. In some drives, only one of the switches will work, despite both being present. Use a multimeter.
Most will have a common pin connected to Ground. Again, use a multimeter. On my disk drive, the common pin is conveniently labelled.
Next, you need the other pin. You need to figure out if its normally open or normally-closed. This is important. You need to solder a wire to to this pin, and bring it out the back.
This is my disk drive before the modification. It is a Mitsubishi MF355F.
This is the wire. Having checked with a multimeter, it is Normally Closed, so it ends up connected to Ground when a disk is *NOT* in the disk drive.
[edit] Intermediate board
The Diskchange is really trivial. On mine, It uses half of a 74LS38.
The circuit I use is as follows:
Which ends up as:
Note that if the disk-inserted switch was low if the disk was inserted, I would use a third gate to invert the signal before it gets to the gate, similar to way the Disk select is inverted.
One thing to bear in mind is that i have heard rumours of problems caused by this sort of signal on other computers while writing, because the ready signal goes low before the disk gets up to speed. (Reading is not a problem because the computer will retry).
I personally havent experienced this with the +3 and this floppy drive (Or the other two i have experimented with), as +3DOS implements a 1 second spin up delay in software, which is more than is required.
If this did end up being aproblem, a way of doing it would be to also AND in the motor-on signal that has had some Hysterisis added to it to provide a delay, Eg, use one of the spare gates to discharge a capacitor-resistor, and plonk this output into the second spare gate to invert it again, then pass this in to the same gate as the disk select signal. Tuning the Capacitor-resistor network, is an excercise for the user :-), however a 1 second delay would be a good target to aim for.
Because the gates are open-collector, A low to High transition will force the capacitor to discharge immediately (Thus the drive is immedately not ready) however, the High to low transition will allow the capacitor to be chaged by the resistor (Which will take longer)
Here is the completed board. As you can see, really complex.
[edit] Final assembly
I made a case out of an aluminium sheet. This also meant I could bolt the regulator to it. Note that the case must be electrically isolated from the regulator tab.
You can see the regulator board at the bottom of picture. Basically The regulator has a normal metal heatsink attached, and I bolted this to the case with some heat transfer compound. The regulator is a 7805 and two-capacitor jobby. as lifted straight from the datasheet.
Note that the floppy drive only requires 5v. Almost no 3.5" floppy drives require +12v, and if it does, you probably have an original Shugart-compatible anyway, and shouldn't be using the hack. The regulator provides power to the disk drive, and the 74LS38 on the main board.
Finally, this is what it looks like in all its glory.
