General Stuff(tm)

[mrbombermillzy]

Looks like the show went well. :)

[stephen_usher]

Got home from my Dad's place in Cornwall and picked up the new DCC socket adapter boards.

They're far smaller and add the 5 ohm resistance (via two 10 ohm resistors to spread the current) and a capacitor to help dampen spikes on the motor line.

IMG_2845.jpeg

[stephen_usher]

My goodness, SOD-323 diodes are tiny. A bit of a SOD to get in the right place on an 0805 set of pads to solder. :-)

[rubber_jonnie]

My goodness, SOD-323 diodes are tiny. A bit of a SOD to get in the right place on an 0805 set of pads to solder. :-)

Oooh, they are weeny!

[stephen_usher]

Now I've practically finished the loco DCC conversion I can get back to thinking about the Apricot hard drive emulation stuff using the Raspberry Pi Pico.

I've come to the conclusion that I can simplify things massively if I move most of the logic and the buffer memory into the Pico. Gone are the problems of dual ported SRAM for the registers and gone is the static 8K buffer SRAM too. It also simplifies the address decode as well.

I still need a GAL to do the initial address decode for selection and to control the I/O buffers but everything else can be done simply by the Pico.

The GAL merely needs to check to see if A5-A7 are high and the I/O select line is low (i.e. 0xe0 + I/O transfer). This can then bring an internal /SELECTED line low.

To interface with the Pico, I can make the /SELECTED line pulled up to 3.3V. Similarly the input buffer for address lines A1-A4 and A8 plus the /WR line would need to output 3.3V. I can't remember if the normal 74LS logic buffers actively pull the lines high or need pull-ups. If the latter then it's simple. Similarly the bi-directional buffer for the lower half of the address bus, D0-D7, for the data transfer.

The Apricot has an I/O delay line which I can automatically pull low until the Pico has set things up, so this will allow the system to be held until any data has been read/written. The Pico will probably be faster anyway.

I can then put all the rest of the logic into the Pico with one core managing the raw I/O side and the other doing all the drive logic and talking to the SD card etc. which isn't timing critical.

[stephen_usher]

So, the chip count should go down to a 16V8 GAL, a 74LVC244, a 74LVC245 and the Pico. The Pico can do the 5V to 3.3V voltage conversion to power the LVC chips.

P.S. It looks like I have all the chips in stock except the 74LVC244. (GAL, Pico, 74AHCT245).

[stephen_usher]

Actually, as I found out today, I had the 74HCT244s (will run at 3.3V) in stock but not the 74VLC245s. D-oh!

So, now I have 20 74LVC244s, which are better for 3.3V buffering anyway, so it's not a total loss.

I only found this out whilst trying to put the prototype hardware together this morning.

IMG_2859.jpeg

Oh well. I'll just have to order them later. I can still do a lot of the work without them. I just can't interface with the Apricot as the data bus needs a level adjustment.

[stephen_usher]

Found that I did actually have some SN74AHCT245N chips in stock! Woot!

So, the prototype hardware is coming together...

IMG_2860.jpeg

Can you spot the 8 bit data bus?

I got the SD card working in Micropython this afternoon. Would help if I had the RX/TX lines the right way around. :-)

Now the real work begins, looking at the pins and doing stuff. Emulating the static RAM buffer+counter, doing the head select decode and then pretending to be a WD1010A ST504 hard disk controller chip.

[mrbombermillzy]

Can you spot the 8 bit data bus?

In authentic 8 bit micro early 80's marketing colours, no less. :D

[stephen_usher]

And so, the first draft schematic for the HD emulator board is ready...

Apricot-HD-Schem-Mk1.jpg

I've designed it so that it can be a daughter board on my RAM board or a stand-alone board.