Page 2 of 39

Re: New project: An ACT Apricot F1 "barn find"

Posted: 21 Apr 2021 13:11
by rubber_jonnie
PhilC wrote: 21 Apr 2021 13:05 Even £20 is too expensive for me but then 8086 stuff just isn't my thing.
All of my PC retro stuff I've acquired for free mostly over the years, not really big into 8086 either, but have a 386 laptop that sorely tested my repairing skills and a 486 DX2 66 Dolch LPAC luggable, plus various Pentium laptops.

I much prefer the 8 bit randomness that was happening during my childhood and early teens when there seemed to be a new machine released every week. Some succeeded, some failed, but it was an exciting time.

Re: New project: An ACT Apricot F1 "barn find"

Posted: 21 Apr 2021 13:46
by stephen_usher
Yeah, I know what people mean by Intel not being their thing. Generally I'm similar but really that's more the PC architecture. You can view the Apricots as more generic 16 bit machines which happen to have an Intel processor and can run a custom version of MSDOS in the same way that many Amstrads were generic 8 bit Zilog machines which ran a custom version of CP/M.

The initial Apricot Xi and PC didn't even have the BIOS on ROM, they just had a diag ROM which then loaded the BIOS in from floppy.

As for the 8086, that's a true 16 bit processor (still interleaved address/data bus though to keep the pin count down). I have no idea why IBM chose the 8088 for the PC-XT as that, like the 68008, is hobbled massively by having a multiplexed 8 bit bus.

Re: New project: An ACT Apricot F1 "barn find"

Posted: 21 Apr 2021 22:36
by stephen_usher
Well, I think that I've found the keyboard and mouse IR data transmission information in a text file of the QDAE (Quick and Dirty Apricot Emulator) so if I get the machine to boot as far as the BIOS then I have a chance of using something to generate the correct infrared packets. Maybe use a Raspberry Pi, read the keyboard and mouse input from the USB devices and send the correct data to the Apricot using an IR LED. Of course then timing is an issue as it's synchronous serial and needs clock pulses. I've no idea how to do this, especially making sure that they're 25ms precisely.

Oh, and I've also found on-line floppy disk images. I could create a converter cable between the Sony 26 way IDC cable and the newer standard used on later floppies and connect a Gotek.

Re: New project: An ACT Apricot F1 "barn find"

Posted: 21 Apr 2021 23:35
by rubber_jonnie
stephen_usher wrote: 21 Apr 2021 22:36 Well, I think that I've found the keyboard and mouse IR data transmission information in a text file of the QDAE (Quick and Dirty Apricot Emulator) so if I get the machine to boot as far as the BIOS then I have a chance of using something to generate the correct infrared packets. Maybe use a Raspberry Pi, read the keyboard and mouse input from the USB devices and send the correct data to the Apricot using an IR LED. Of course then timing is an issue as it's synchronous serial and needs clock pulses. I've no idea how to do this, especially making sure that they're 25ms precisely.

Oh, and I've also found on-line floppy disk images. I could create a converter cable between the Sony 26 way IDC cable and the newer standard used on later floppies and connect a Gotek.
If you can get this going, hopefully I can get the Beeb that is somewhat less crusty going again.

Re: New project: An ACT Apricot F1 "barn find"

Posted: 22 Apr 2021 14:28
by stephen_usher
I've found more information on the IR interface.

It says that it's a monosync synchronous serial connection at ~3.85Kb/s but from what I can see it's actually an auto-bauding asynchronous connection, using a 556 to square up and sync the pulses coming in from the IR interface. The initial byte of 10100101 (0x5e) sets up the synchronisation and the next three bytes have to come in continuously on the correct timing.

As for the "hamming codes", this code snippet from the QDAE emulator describes the encoding:

Code: Select all

void Z80_SIO::write_ir(int ch, u_int16 irword)
{
        u_int8 nibble[3];
        int n;

        nibble[0] = (irword >> 8) & 0x0F;
        nibble[1] = irword & 0x0F;
        nibble[2] = (irword >> 4) & 0x0F;

        for (n = 0; n < 3; n++)
        {
                write_rx(ch, st_ir[nibble[n]]);
        }
}
So, replicating the IR output circuitry from the keyboard schematic and attaching that to the UART on a Raspberry Pi and setting that to 3.85Kbaud (if I can) should allow me to send keyboard codes to the system.

Re: New project: An ACT Apricot F1 "barn find"

Posted: 22 Apr 2021 15:13
by stephen_usher
OK. It looks like the Raspberry Pi's SPI interface will work, using a clock divider of 64936. Output that to the IR and throw out the four bytes.

Re: New project: An ACT Apricot F1 "barn find"

Posted: 22 Apr 2021 15:52
by rubber_jonnie
stephen_usher wrote: 22 Apr 2021 15:13 OK. It looks like the Raspberry Pi's SPI interface will work, using a clock divider of 64936. Output that to the IR and throw out the four bytes.
Sounds very cool :)

How is the rest of the machine doing at this point?

Re: New project: An ACT Apricot F1 "barn find"

Posted: 22 Apr 2021 16:52
by stephen_usher
It's currently in transit with Parcels-2-Go... Which is why I can do nothing but research.

Re: New project: An ACT Apricot F1 "barn find"

Posted: 24 Apr 2021 13:36
by stephen_usher
I managed to intercept the Hermes driver at work and I've got the machine.

The motherboard doesn't look in too bad a condition. There are some slightly corroded legs on the ROM and one on the CPU but otherwise it looks clean. It was obviously repaired in 1990 and one chip at the front of the board is socketed and there's flux residue underneath.

IMG_2244.jpg

The floppy drive, however, looks destroyed.

IMG_2245.jpg

As for the case, it looks mostly repairable. The front plate is fine with only one crack which is hidden behind the black strip.

IMG_2246.jpg
IMG_2247.jpg
IMG_2248.jpg

I've no more time to look at it as I'm leaving for a fortnight this evening.

Re: New project: An ACT Apricot F1 "barn find"

Posted: 24 Apr 2021 13:54
by stephen_usher
OK. Had time to take a look inside the power supply. Only one RIFA smoke generator X2 capacitor, the other X2 is a wax covered one and they're generally bullet proof. The RIFA is well bulging and would definitely explode if powered.

Everything else in there looks fine at first glance.