Hacking a DIY Qubide

[Pr0f]

I needed to test a few ideas out and so took a scalpel and soldering iron to a diy qubide pcb.

The ideas:

Max out the RAM - 896K, or 864K with upper ROM enabled.
Qubata ROM
Toolkits
ROM split high and low - with the ability to deactivate the top 32K of the 64K ROM for that little extra memory
RAM shadowing - so the upper 64K of the internal RAM is always disabled, and the lower 64K is shadowed, so that only writes go to the internal memory, reads are always from the faster RAM on the Qubide - the lower 64K shadowing can be disabled by a jumper.

ROM occupies 10000 and 14000 and F8000 and FC000 respectively

It looks a bit Frankenstein, but all the test objectives were met - so that's the main thing.

The design leaves ROM port free

[Pr0f]

Happy to share the logic for the 3rd GAL if anyone wants it - but the standard DIY Qubide PCB needs some rework to take the 3rd GAL chip instead of the 74LS688 that normally sites there.

[Derek_Stewart]

Hi,

I am interested in thies, I have 2 completed Qubides and a PCB waining for construction.

I also have an original, QUBIDE without any ram.

[martyn_hill]

Hey Prof!

I have a QubIDE DIY still to make up and also need some ROM space and your Frankenstein sounds perfect!

Would be delighted to mod my board with your designs and PAL eqn/jed

Martyn.

[Pr0f]

So the logic for GAL3 (which now sits in the place originally holding the address comparator (688 chip), some notes on hacking the board - there is a lot of cutting around the pins of the old 688, as they are grounded on both edges of the pin pad, and sometimes on both sides. I tried to reuse existing tracks where I could, but logic limitations on the GAL require some fiddling about - pin 1 of the socket for 688 is normally grounded - but this is required to decode the ROM for the Qubata, so must be cut, I trimmed the socket pin itself.

[Derek_Stewart]

Hi,

Could the PCB be redesigned, with the mods, I think I have the schematic and PCB files, in Eagle, but can be converted into KiCad.

[Chris_68]

Hi

Thank you very much for sharing! Serious hacking in Xmas time.

I actually ordered PCBs for QUBIDE but the design by Ilion. I will study what differs and if I can modify one of the PCBs.

Happy holidays

[Peter]

Just out of curiosity: Why did you pick Qubide rather than QL-SD as starting point for your mass storage hack? Is it the higher write speed, because the extension bus is not "read-only"?

As food for thought: The QL-SD design is open source. Even if you don't want to re-synthesize the PLD, you can use the existing JEDEC binary. If you generate your own chip select, the PLD is not even bound to a specific address. Just the driver needs the base address changed and a recompile. On the positive side, QL-SD would allow a much smaller PCB, the more widely established QLWA filesystem, and hotplugging. As for PCB layout, routing between driver EPROM and PLD is particularly easy, since I had optimized the PLD pinout for it.

If write speed is a decisive issue, it would also not be rocket science to create a "direct write" version of QL-SD. But that's beyond a quick hack.

[Pr0f]

Hi,

Could the PCB be redesigned, with the mods, I think I have the schematic and PCB files, in Eagle, but can be converted into KiCad.

I think it could be redesigned - and if that happens - it would be possible to use a larger GAL and provide additional jumper options to control RAM allocation.

I wanted to test the RAM shadowing idea that Dave / Nasta proposed some time back - so happy to confirm the scheme that came up with does indeed work

It might be a nice idea to provision a 'soft eprom' for C000 address or another ROM location - by allowing writes to go to RAM, and then have an I/O address decoded to toggle a flip flop to gate read access and block write access to the 'soft eprom' - this would allow the RAM on the card to emulate a QL ROM that could be preloaded with the image (loading the bytes), then flicking readonly access and reset of QL - the image would now be scanned as any normal ROM.

The card could be kept smaller by using SMD versions of the RAM chips - I got away with piggy backing two of them on the Qubide card, and using a slimline socket, but it only just clears the QL case

I'm also well aware that the addresses I chose are also those used by Trump and Gold cards, but the argument there was that this big RAM card was not going to be used with either of those cards. I may work on a version with an HD/ED floppy support and SMD components except the ROM and GAL's. The main reason for this mini project was to sanity test some ideas for my DIY QL project.

[Pr0f]

Just out of curiosity: Why did you pick Qubide rather than QL-SD as starting point for your mass storage hack? Is it the higher write speed, because the extension bus is not "read-only"?

As food for thought: The QL-SD design is open source. Even if you don't want to re-synthesize the PLD, you can use the existing JEDEC binary. If you generate your own chip select, the PLD is not even bound to a specific address. Just the driver needs the base address changed and a recompile. On the positive side, QL-SD would allow a much smaller PCB, the more widely established QLWA filesystem, and hotplugging. As for PCB layout, routing between driver EPROM and PLD is particularly easy, since I had optimized the PLD pinout for it.

If write speed is a decisive issue, it would also not be rocket science to create a "direct write" version of QL-SD. But that's beyond a quick hack.

Hi Peter - it was purely a testbed for some ideas for my larger DIY QL project - and I had a spare Qubide PCB knocking around.

I did have an idea of combining the QL-SD / Minerva I2C board and QIMI onto a similar QIMI style PCB as an internal upgrade - but your QIMSI made that project largely redundant