exxos's Atari, Amiga & retro forum

stephen_usher wrote: 19 Jun 2024 18:06 The ElkSD Plus 1 cartridge arrived today.

Other than the Byzantine requirement to have all the floppy images in a container file named BEEB.MMD which is the first file on the drive and internally indexed using a database file created and managed using Microsoft Access 2000 (nothing newer or older) it works well. All you need to quickly load games.

Yeah, the same is a problem with the Beeb equivalent, however I think it may have been updated recently so that the MMD file is no longer required. There is a tool, the name of which escapes right now, which allows for editing the images contained in it too.

I did get my head around the BEEB.MMD and updated it manually a few times with new games/demos and got it uploaded to Stardot.
I think the last I modified was v0.3.

Yes I did read about some different software that does not need the complex disk injection into the MMD menu. This was a while ago so not sure of the current status of it.
But I decided to go down the Disk Upgrade ROMs on my Model B and I am currently using a Gotek.

Yeah, there are a few utilities, most run under Windows only though.

I think the MMFS comes from a time before there were readily available FAT32 libraries for developers to use and it also is simple enough to fit in a 16K ROM. It was developed in the early noughties after all.

Put the sockets for the RAM in the Apricot Xi and replaced the RAM entirely... No change. So it's not the RAM.

Followed the boot process further, so started looking at the I/O data bus. It looks like it's being dragged down.

Temporarily fitting a 10K pull-up on the bus often allows the POST to finish but with a corrupt screen and an error message

Seeing that the only chips on that bus which output onto the data bus are the floppy controller and the Z80-SIO-0 my suspicions are with that chip as I can remove the floppy controller and the symptoms don't change. Piggybacking a slightly broken SIO on top sometimes gets the boot further as well.

So, at the weekend I'll desolder the SIO and fit a socket, then see what happens to the bus with the chip removed.

Swapped out the 41256 chips in the upper bank I used for testing and replaced the originals, then did a better pull-up bodge on the I/O data bus.


The result was that the system got past the diagnostics but with an error and it couldn't initialise the keyboard.


This suggests that it is indeed the Z80-SIO-0 which is the bad chip. Probably fried by a static spike on the serial port or something.

The biggest problem is the earliest delivery time is about 2 weeks.

stephen_usher wrote: 21 Jun 2024 19:11 The biggest problem is the earliest delivery time is about 2 weeks.

For the Z80-SIO? That's kind of annoying :(

rubber_jonnie wrote: 21 Jun 2024 19:27

stephen_usher wrote: 21 Jun 2024 19:11 The biggest problem is the earliest delivery time is about 2 weeks.

For the Z80-SIO? That's kind of annoying :(

Yep, that's Little Diode on eBay.

The other "UK" sellers, with Chinese sounding names, were nearer three weeks, so that'll be shipped from China then.

stephen_usher wrote: 21 Jun 2024 21:24

rubber_jonnie wrote: 21 Jun 2024 19:27

For the Z80-SIO? That's kind of annoying :(

Yep, that's Little Diode on eBay.

The other "UK" sellers, with Chinese sounding names, were nearer three weeks, so that'll be shipped from China then.

Ahh jeez that's rubbish.

OK, so it wasn't the Z80-SIO pulling the bus down, but it's not working correctly either.

Thankfully no bent-over pins on this chip so it was relatively easy to remove and replace with a socket.

I have a spare Z80-SIO/1 which I can hack to make pin compatible with bodge wires soldered to pin stubs.