Tang Nano 20K -- an even better 65?02 platform?
Re: Tang Nano 20K -- an even better 65?02 platform?
figForth now works. I moved the IO page to FE00-FEFF to leave more contiguous space for applications.
Re: Tang Nano 20K -- an even better 65?02 platform?
Binary load format changed!
Now a 2-byte header is expected:
* 1-byte load/execution page
* 1-byte page count
The binary file must be padded to a 256-byte boundary.
Now a 2-byte header is expected:
* 1-byte load/execution page
* 1-byte page count
The binary file must be padded to a 256-byte boundary.
Re: Tang Nano 20K -- an even better 65?02 platform?
75MHz -- seems to be stable, even though timing analysis report states 66.6 is currently fMax. I will work at 75MHz and back off if there is any trouble, but so far everything I've thrown at it runs without a problem. 80 loads the monitor but seems unstable on loading binaries. Even 66MHz is pretty respectable, and is a good number (my first Intel box was running at 66MHz...)
I reorganized the IO input mux as the old one was getting crusty...
I reorganized the IO input mux as the old one was getting crusty...
Re: Tang Nano 20K -- an even better 65?02 platform?
Got a working SDRAM controller (basic, just a test really).
I started with https://github.com/nand2mario/sdram-tang-nano-20k, which is a working standalone demo, but it is pretty buggy. If you started with it, watch out! The PLL is misconfigured in this project, and the state machines are a bit junky.
It took me a couple of days banging my head trying to figure out how to do this cleanly, so that back-to-back requests work. Luckily, the SDRAM seems to be identical to a Micron part,https://www.micros.com.pl/mediaserver/i ... 32143t.pdf.
Once things worked, I started pushing my luck by speeding up tRCD-tRP-tCL, and it works at 1-1-1 with a 13ns clock, implying a really fast SDRAM, better than Micron's -5 grade (15ns)... Not to mention that CL should be generally 2 or 3, but it seems to work with 1. I will have to shake it down.
I am thinking of using it as a RAMdisk, along with SD card, once I decide on how to deal with disks -- maybe one of the CPM-65s or DOS/65, or if I get too annoyed, maybe I'll do something different.
Maybe I'll set up a 2K BRAM as a dual-ported memory, mapped as 8-bits into the 6502 system, and as 32-bits from the SDRAM controller side. I can then burst-read 512-byte sectors very quickly.
I started with https://github.com/nand2mario/sdram-tang-nano-20k, which is a working standalone demo, but it is pretty buggy. If you started with it, watch out! The PLL is misconfigured in this project, and the state machines are a bit junky.
It took me a couple of days banging my head trying to figure out how to do this cleanly, so that back-to-back requests work. Luckily, the SDRAM seems to be identical to a Micron part,https://www.micros.com.pl/mediaserver/i ... 32143t.pdf.
Once things worked, I started pushing my luck by speeding up tRCD-tRP-tCL, and it works at 1-1-1 with a 13ns clock, implying a really fast SDRAM, better than Micron's -5 grade (15ns)... Not to mention that CL should be generally 2 or 3, but it seems to work with 1. I will have to shake it down.
I am thinking of using it as a RAMdisk, along with SD card, once I decide on how to deal with disks -- maybe one of the CPM-65s or DOS/65, or if I get too annoyed, maybe I'll do something different.
Maybe I'll set up a 2K BRAM as a dual-ported memory, mapped as 8-bits into the 6502 system, and as 32-bits from the SDRAM controller side. I can then burst-read 512-byte sectors very quickly.
Re: Tang Nano 20K -- an even better 65?02 platform?
SPI/SD card is now integrated into the bootROM.
figFORTH is fixed to use 1K blocks and R/W patched to use the SD card. For now I have a FAT16-formatted card with a large contiguous file starting at sector $B80. At some point I will have to make a simple file-system calculation that is more generic.
figFORTH is fixed to use 1K blocks and R/W patched to use the SD card. For now I have a FAT16-formatted card with a large contiguous file starting at sector $B80. At some point I will have to make a simple file-system calculation that is more generic.
Re: Tang Nano 20K -- an even better 65?02 platform?
A quick update:
* TaliForth is now working, complete with block storage via SDcard
* Downgraded to 70MHz due to stability issues
* TaliForth is now working, complete with block storage via SDcard
* Downgraded to 70MHz due to stability issues
Re: Tang Nano 20K -- an even better 65?02 platform?
Installing Gowin V1.9.11.01 Education (Linux x64), and of course it does not run -- some libraries missing. So,
Giving up for now after dorking around with qt and libraries and paths until sick.
If you have it working, please let me know
Also, somehow lost my Nano20K, so ordered another from Amazone, coming tomorrow.
Giving up for now after dorking around with qt and libraries and paths until sick.
If you have it working, please let me know
Also, somehow lost my Nano20K, so ordered another from Amazone, coming tomorrow.