FPGA replacment for the ZX8301 ULA and more
Re: FPGA replacment for the ZX8301 ULA and more
I did some optimizations in speed (10Mhz 68008 cpu)
Leon
Projects: https://hackaday.io/projects/hacker/357657
Projects: https://hackaday.io/projects/hacker/357657
Re: FPGA replacment for the ZX8301 ULA and more
Wow - more than double a stock QL - and still 8 bit wide - so this is all achieved through faster memory access and a faster CPU - 133% faster.
Re: FPGA replacment for the ZX8301 ULA and more
And with a little bigger fpga I can make it even a bit faster if I implement a true dual bus video memory and it will then be easier to make more graphics modes and also to implement a vga output. I'll try it, I hope soon.
Intel/Altera include small memory amount in their entry level fpgas and when you go for a bigger fpga for more sram with the same 144 pin package then you get less i/o pins so this is a problem. Maybe I'll have to try a BGA package in the future, Peter says it's easy to solder.
Anyway I think I found a good compromise between ram and io pins in a 144 pin package that will have all the features but will reduce the colors from 64 to 32+palette choice but will be a little faster and have a vga output and make easier to implement more graphics modes.
I'll add extra sound channels, a msec timer and potentially useful things like reading the current vertical line and I think that will be a good point to finalize the first version of the project.
Intel/Altera include small memory amount in their entry level fpgas and when you go for a bigger fpga for more sram with the same 144 pin package then you get less i/o pins so this is a problem. Maybe I'll have to try a BGA package in the future, Peter says it's easy to solder.
Anyway I think I found a good compromise between ram and io pins in a 144 pin package that will have all the features but will reduce the colors from 64 to 32+palette choice but will be a little faster and have a vga output and make easier to implement more graphics modes.
I'll add extra sound channels, a msec timer and potentially useful things like reading the current vertical line and I think that will be a good point to finalize the first version of the project.
Leon
Projects: https://hackaday.io/projects/hacker/357657
Projects: https://hackaday.io/projects/hacker/357657
Re: FPGA replacment for the ZX8301 ULA and more
I forgot to mention that I included a AM29F040 512K flash ic that currently holds the TK2 and QLSD driver, it can be mapped over the 2MB ram and reprogrammed easily with something like the TL866+plcc32 adapter.
Could I use this rom to load a version of smsq?
Could I use this rom to load a version of smsq?
Leon
Projects: https://hackaday.io/projects/hacker/357657
Projects: https://hackaday.io/projects/hacker/357657
Re: FPGA replacment for the ZX8301 ULA and more
Well the disk test makes the index a bit faster, I think it's more close to 2X, 100% faster that's not bad.Pr0f wrote: Tue Nov 14, 2023 8:18 am Wow - more than double a stock QL - and still 8 bit wide - so this is all achieved through faster memory access and a faster CPU - 133% faster.
Leon
Projects: https://hackaday.io/projects/hacker/357657
Projects: https://hackaday.io/projects/hacker/357657
Re: FPGA replacment for the ZX8301 ULA and more
No I do not say that. I can do it at home for prototyping, but it is difficult.lliont wrote: Tue Nov 14, 2023 8:32 am Maybe I'll have to try a BGA package in the future, Peter says it's easy to solder.
It requires experiments, manual skills and an infrared rework station.
Also my PCBs have special preparations for manual BGA placement.
All I can say is that the BGA packages which I use are relatively easy for machine soldering.
Re: FPGA replacment for the ZX8301 ULA and more
Sorry Peter I misunderstood from this post viewtopic.php?p=50846#p50846, and because for me the 144 is not that hard I assumed it will be relatively "easy" to solder the BGA too.
here is a picture of my board:
here is a picture of my board:
Leon
Projects: https://hackaday.io/projects/hacker/357657
Projects: https://hackaday.io/projects/hacker/357657
Re: FPGA replacment for the ZX8301 ULA and more
I was just looking at the CPU clock speed - so 10/7.5 x 100 - so effectively 1.33 times quicker on CPU alone - which means the greater amount of the speed increase has come about through your memory access improvements.lliont wrote: Tue Nov 14, 2023 9:35 amWell the disk test makes the index a bit faster, I think it's more close to 2X, 100% faster that's not bad.Pr0f wrote: Tue Nov 14, 2023 8:18 am Wow - more than double a stock QL - and still 8 bit wide - so this is all achieved through faster memory access and a faster CPU - 133% faster.
-
- Font of All Knowledge
- Posts: 4767
- Joined: Mon Dec 20, 2010 11:40 am
- Location: Sunny Runcorn, Cheshire, UK
Re: FPGA replacment for the ZX8301 ULA and more
Hi,
You new new board looks great, if it is as good as the previous boards, I will want one.
To load SMSQ/E a Supergold card would be needed, but the Supergold card disables the 68000 so this looks like a new version of SMSQ/E, which might be harder than manual BGA soldering.
You new new board looks great, if it is as good as the previous boards, I will want one.
To load SMSQ/E a Supergold card would be needed, but the Supergold card disables the 68000 so this looks like a new version of SMSQ/E, which might be harder than manual BGA soldering.
Regards,
Derek
Derek
Re: FPGA replacment for the ZX8301 ULA and more
LOL, so QDOS and Minerva it will be with easy switching between them.Derek_Stewart wrote: Tue Nov 14, 2023 10:29 am To load SMSQ/E a Supergold card would be needed, but the Supergold card disables the 68000 so this looks like a new version of SMSQ/E, which might be harder than manual BGA soldering.
Leon
Projects: https://hackaday.io/projects/hacker/357657
Projects: https://hackaday.io/projects/hacker/357657