I haven't been able to find a QL shaped Enter Key top. I though I might just have the Vertical 2U Enter Key, put the " \ | " key next to the Enter key and maybe a 1U "Delete" in the top corner - I just trying to source a double-pole cherry key switch so I can take Ctrl and left arrow to it, or perhaps a horizontal 2U Enter key with a "Delete" above that. I don't thing a 3D printed one would look very professional. I'll give it a go on my new Anycubic Kobra S1 printer and see how it come out - I don't think anyone has come up with any double shot QL keys so I would have to write "Enter" on the top in crayon. The idea is you can just take the real QL keytops of the original keyboard and put them on the new mechanical keyboard. It does involve rather a careful operation to cut out the old keyboard from the QL upper case - so it's not for the faint hearted. I just didn't have a keyboard so I bought a set of MX keytops from Aliexpress for about £7.
Hi all you clever QL people. Is there anyone out there who could help me with my continuing problem with my QL Pistorm reading QL RAM. RAM Writes work perfectly, so do ROM reads but I simply cannot fathom why RAM reads are getting corrupted during video updates. On the logic analyser trace below the top line is taken from VDA (pin38) of the ZX8301. The First 2 read transactions are 0x0F and 0x10 the next 4 reads should be 0x11, 0x12, 0x13 and 0x 14 but I am reading 0xFF, 0xFB, 0xFF and 0xFF. There appears to be some sort of glitch on the data bus right around where the data is latched. this happens every time the video is updated. Any thoughts would be welcome. Thanks - Will
Last edited by Will James on Fri Jul 04, 2025 5:25 pm, edited 1 time in total.
I wonder if that's the transiever (245) chip that's used to allow access to the video RAM - you might be attmepting to read the databus just at the wrong time...
A lot of the accelerator cards like Gold / Super Gold card only read from their own copy of the RAM - but write to both the QL's video RAM and their own - this way their own copy of RAM in the video area can be accessed much faster for reads - there is only a speed penalty on writing.
Yes - that is exactly what I will be doing, but I need to be able to read accurately from the IO area which is also cursed with video interrupts the same a RAM...
I think you are right about the transceiver (245) though, I might see what happens if I try to latch right as !DTACK de-asserts.
Eurika! It's only taken me 6 months of working on this but finally, after completely rewriting the Amiga PiStorm CPLD code, I have reliable RAM and IO reads...
Now to get the Masushi emulator running... I've already written some of it for the QL.
I am very happy now. It's was a vey tough one for me but I have learned so much on the way... Verilog, Quartus, KiCAD, everything about the ZX83 and QL architecture
Once Masushi is running I will rewrite the Microdrive handling so it uses a proper timer rather than bit banging so they will continue to work
with the emulator however fast it runs. Then depends what people want... I thought perhaps USB FDD, HDD, networking, HDMI video with new HD screen mode.