The DDI-2

[TFM]

What?!? Mine has devalued by €20 in under a year, it's as bad as a new car I tells ya 

Ah! You got one, and still bitch about 64 KB! Shame on you! 

[TFM]

it seems more cheap to just buy a 6128 when you have a 464...

This is the solution I suggest to everyone!

[TFM]

Back to topic...




Will the four drives solution work?




Will the DDI-2 use alternate addressing?

[beaker]

Ah! You got one, and still bitch about 64 KB! Shame on you! 

Actually TFM, that's only half the story. I've actually also got one of Yarek's 512kb RAM modules as well  Bryce modified for me last year and changed the header and made me a cable so I can use it with the LowerROM and Megaflash and I am guessing the DDI-2 when it's released....

[attach=2]

I was just trying to make the argument for those who may not have these parts 

[Cholo]

Is Symbiface still available?, where from?, how much?

Yup, Prodatron over at http://www.symbos.de does a batch SF2's from time to time in the past.

[Bryce]

@Beaker: I've been looking for a KC Compact for years. Supposedly all my expansions work on it too, but I'd like one to do my own tests to be 100% sure.

@TFM: I'm still looking at the four drives solution. Now that the expansion has two floppy headers it's a high possibility that I will implement it (just for you ). The alternative addressing is highly unlikely because it looks like it would mean adding another IC to the board, which just isn't possible or wanted at the moment.

Bryce.

[Bryce]

Status Update: I've managed to implement both the Ready switch and a manual/automatic side switch (in a safe manner) and I've even managed to get this to route on a 100x80mm PCB. The board is packed to its limit and the layout is getting ridiculously complex (see below), so I doubt it is possible to implement anything more. Both the 4 drives and alternative addressing would require an additional IC that I just don't have the space for.

A question occured to me while battling with the routing: On an original DDI-1, the hardware is designed to make use of the side signal, even though the FD-1 doesn't use it, nor is it supported in AMSDOS. So instead of having manual side switches on cables or on this interface, why hasn't someone just added a "|Side" command to AMSDOS, so that 3.5in drives are fully usable in AMSDOS?

Bryce.

[beaker]

Does it have a little LED to show the unit's getting power from the machine?  I always find them reassuring for some reason 

[Bryce]

No it's LEDless at the moment. I did consider adding a power and/or activity LED, but that can come later. I was also considering adding a "peripheral header", but that's looking doubtful now due to lack of space. The periheral header would have allowed you to plug in an addition PCB with track, sector and side display. Just a gimmick, but I like things that flash pointless information

I'm going to freeze the design for now and build a prototype to confirm the design so far.

Bryce.

[Bryce]

Ok, just for you beaker, I've added a power LED. The layout below looks a bit different because the ground plane is now visible too.

Bryce.

[beaker]

Yay, I didn't get the 64k RAM that I didn't need but I got an LED 

[arnoldemu]

A question occured to me while battling with the routing: On an original DDI-1, the hardware is designed to make use of the side signal, even though the FD-1 doesn't use it, nor is it supported in AMSDOS. So instead of having manual side switches on cables or on this interface, why hasn't someone just added a "|Side" command to AMSDOS, so that 3.5in drives are fully usable in AMSDOS?

It's done in hardware so that it works with all software.

Imagine you had a 3.5" drive B, and on 1 side you had copied some software that used a special loader.
Loader thinks it's using side 0.

Then on the other you did the same.

With the switch you can choose which side to load from, the software doesn't know. It still thinks it's on side 0, but you've hardware switched it to side 1.

With software it'll only work under amsdos and not for special loaders.

Same reason that the drive switch is hardware, you do this for software that only likes to load from drive A, use the switch and now the 3.5" drive is drive A and the 3" drive is drive B. Software happy

[Bryce]

Yes, but the special loaders on original software don't set the side do they? If you had a 3.5in disk full of software. Lets say we have "Gauntlet" saved on Side 0 and "Strike Force Harrier" on side 1. Just finished playing Gauntlet, hit the F1 key (which the modified AMSDOS uses to swap sides) and type Run "Disk" or whatever and load my software. Wouldn't that work? The software still doesn't know which side I'm using. Even modern loader like Batman forever would work because you have a switch that both the user or software can set.

The ABBA switch is a different matter, I understand why that should be implemented in hardware.

Bryce.

[MaV]

Please use these switches:


or better yet:


And make them hum and crackle with the occasional flash arcing across the poles - especially during thunderstorms outside.

I want my CPCs to finally come alive!!! Muahahaha!

[arnoldemu]

Yes, but the special loaders on original software don't set the side do they? If you had a 3.5in disk full of software. Lets say we have "Gauntlet" saved on Side 0 and "Strike Force Harrier" on side 1. Just finished playing Gauntlet, hit the F1 key (which the modified AMSDOS uses to swap sides) and type Run "Disk" or whatever and load my software. Wouldn't that work? The software still doesn't know which side I'm using. Even modern loader like Batman forever would work because you have a switch that both the user or software can set.

The ABBA switch is a different matter, I understand why that should be implemented in hardware.

Bryce.

Yes they do set the side. They access the FDC directly and send it read data commands.
First byte is command, second byte is drive and side select. 2 bits for drive, 1 bit for side.

The | command would only work if ultimately it was sending an OUT to set the hardware to force the side.

[Bryce]

Ok, understood. But how many special loaders were there that set the side? I didn't know there were any until you mentioned it.

@Mav: The top switch looks good for the ABBA switch, the PCB can be hidden nicely under the switch Not sure about the sparks though. I'd have to add a DC/DC convertor to get the voltage up high enough.

Bryce.

[arnoldemu]

Ok, understood. But how many special loaders were there that set the side? I didn't know there were any until you mentioned it.

@Mav: The top switch looks good for the ABBA switch, the PCB can be hidden nicely under the switch Not sure about the sparks though. I'd have to add a DC/DC convertor to get the voltage up high enough.

Bryce.

Any loaders that use direct fdc disc access will set the side.

AMSDOS also sets the side, but here we could hack it so we could force the side we wanted, but this would only work for loaders that use AMSDOS, and not direct to the fdc.

[TotO]

Brice, why not using the internal 4MHz clock instead of adding one?
The CPC FDC765 datasheet show a 8MHz source too, but the CPC use less.

[gerald]

Brice, why not using the internal 4MHz clock instead of adding one?
The CPC FDC765 datasheet show a 8MHz source too, but the CPC use less.

Interresing point
  In the DDI1, a 8MHz clock is generated from the 4MHz to clock the data separator which then feed the FDC with a 4MHz clock.
  In the 6128, the FDC is clocked with the system 4MHz.

[TFM]

why hasn't someone just added a "|Side" command to AMSDOS, so that 3.5in drives are fully usable in AMSDOS?

Because back the day Amstrad sold their 3" drives [nb]The reason for using 3" was that Amstrad got a big batch of 3" drives very cheap. Nobody did use 3.5" at all back in 1983/4.[/nb] and nobody thought about adding a 2-sided drive, that idea was born later.

[Bryce]

Brice, why not using the internal 4MHz clock instead of adding one?
The CPC FDC765 datasheet show a 8MHz source too, but the CPC use less.

I was waiting for someone to notice that There's two reasons for that. No.1 was so that I could reduce the amount of ICs and hence the size of the PCB. At the moment I have used a standard thru-pin part, but I intend going to an SMD clock later which I can easily hide under the FDC. No2 is because the 8Mhz doesn't need to be synchronised with the CPCs clock, but because it is, it's one of the reasons why the CPC is almost impossible to overclock - When you increase the internal clock of the CPC, the disks are no longer readable. So by keeping the clocks separated from each other, we are one step closer to an overclocked CPC

@Gerard: The DDI-1 used a 9229 separator that needs 8Mhz, but the 6128 uses a 9216 separator which can take 4Mhz, but it means they also needed to add a 74HC161. Most likely a cost decision. I'm not a big fan of frequency multipliers as they tend to be unreliable and can produce a really unclean signal. I'm not aware of any stability problems with the DDI-1, but maybe this influenced their decision too?

Bryce.

[TotO]

Yes, I said that about the footprint first.


When you over-clock a CPC, the others problems are:
- memory speed
- crtc display
- sound
...

[Bryce]

I only see the CRTC as a real problem to overclocking, the other problems aren't as serious or could be worked around. but what good is an overclocked CPC, if you can't load anything from disk to appreciate it?

Bryce.

[TotO]

[delete me]

[TFM]

Overclocking? I did a lot of (real CPC life) experiments, the critical component seems to be the GA. It can usually take 6 MHz instead of 4, but never 8. CRTC did well, as all other components. But I did only test on a single machine.