exxos's Atari, Amiga & retro forum

I have a working Apricot Xi motherboard!

In lieu of the Z80-SIO/0 I modified an SIO/1, pin 25 -> socket 26, pin 26 connected to pin 27 -> socket 27.

I also installed a proper version of the peripheral bus pull-up hack.

The result, a happy machine:


As a test I disconnected the pull-up and it still booted, but goodness knows how given what the data bus looks like! Have you ever seen a data line this awful?

Wow that's quite the trace...

Well, there's obviously more wrong with this machine as I've put it back together and it doesn't see the hard drive interface card. However, if I install a RAM upgrade it sees that plugged into the same slot, which suggests that the I/O select logic has issues.

stephen_usher wrote: 23 Jun 2024 10:04 Well, there's obviously more wrong with this machine as I've put it back together and it doesn't see the hard drive interface card. However, if I install a RAM upgrade it sees that plugged into the same slot, which suggests that the I/O select logic has issues.

So what's the difference between the RAM expansion and HDD interface? Any schematics that might shed some light on it?

Well, the RAM expansion will use the RAM select lines but the disk expansion will use the I/O select lines to access it, basically a few I/O addresses. If those lines aren't working then the interface won't respond.

As for schematics, the field service guide has a complete set at the back.

stephen_usher wrote: 23 Jun 2024 10:11 Well, the RAM expansion will use the RAM select lines but the disk expansion will use the I/O select lines to access it, basically a few I/O addresses. If those lines aren't working then the interface won't respond.

As for schematics, the field service guide has a complete set at the back.

Sounds like you are along the path of finding out what's wrong :)

Well, all the signals *look* OK but the card isn't selecting itself (I have two spare cards).

The ROM tests to see if the hard disk interface is there by trying to write to and then read from the buffer static RAM. It does this by writing to a specific I/O address to reset the address counter and then writes to another address with incrementing values 8192 times. It then resets the counter and reads 4097 times from that address and compares the value it gets back. If the value hasn't wrapped around, and is correct, then it assumes an 8K buffer. If it's wrapped around then it's 4K buffer, if the value is nonsensical then the card isn't present.

I don't see the counters which are used to increment the buffer address being accessed at all, which suggests the PAL isn't finding the expected address on the bus. Or both cards are faulty in the same way.

OK, maybe the Xi detects the hard drive interface differently from the F series. On a different machine, which boot(ed) from the hard drive it didn't actually check the buffer RAM and only accessed it when reading the disk.

Unfortunately it seems that the drive I've been using has stopped responding correctly so I'm going to have to leave it there.

Anyway, decided to try to boot into DOS using a floppy image, so connected up the Gotek.

It starts to boot but then halts before printing any banner. This is an image I know works on the Xi so something else is majorly wrong on the motherboard. I think it's time to give up on this "scrap" machine and really make it scrap.

I can move all the working bits into the other "scrap" machine and make a single working one. All I've wasted are some sockets and a spare Z80-SIO/1 which I had no use for anyway.

Too warm to do much at the moment so this evening tried out the analogue port on the Electron with Plus 1+ unit.

ADVAL(0), ADVAL(1) and ADVAL(2) all give the correct values when the BBC joystick is moved around and the fire button pressed but disappointingly Elite doesn't see the stick. I wonder if it's looking for a different type of interface on the Electron.