Page 9 of 12
Re: Smonson's slow H4 build
Posted: 13 Jun 2020 07:26
by Smonson
I finally got hold of a real UV-C erasing bulb, after buying one fake and getting another one cancelled due to postal service being withdrawn from the seller in China.
So I was able to burn TOS 1.04 and 2.06 onto the two halves of my ROM chip, but only 1.04 works for me on the H4. I read the contents of the chip back and compared the data and it's a match to the TOS 2.06 binary that I downloaded from
http://www.avtandil.narod.ru/tose.html (tos206uk.zip).
Is there anything you need to know for this version of TOS to boot on the H4? I burned it at address 262,144 (0x40000) and it's not byte-swapped (it's the same byte order as 104, which works). I just get a completely black screen on startup.
Re: Smonson's slow H4 build
Posted: 13 Jun 2020 07:57
by derkom
TOS 2.06 will not boot on the H4 without a 2.06 decoder for example on the
beta IDE board.
Re: Smonson's slow H4 build
Posted: 13 Jun 2020 08:51
by Smonson
Aha! That's exactly what I needed to know, thanks!
Re: Smonson's slow H4 build
Posted: 13 Jun 2020 10:01
by exxos
I have done some dedicated 206 decoder boards for such things as well :)
Re: Smonson's slow H4 build
Posted: 13 Jun 2020 14:38
by Smonson
It makes sense in retrospect, I should have realised ahead of time. Since the STFM places TOS at 0xfc0000 and 0xffxxxx is, I think, all reserved for memory-mapped peripherals. I can easily live without TOS 2.06 for the time being.
Re: Smonson's slow H4 build
Posted: 13 Jun 2020 15:02
by PaulJ
@Smonson , you should consider giving @cmorley's eprom emulator a try. It still requires a decode that provides the $0, $fc0000, and $e00000 as the eprom but its easier to use, holds 8 256k images and you don't need to pull the eprom out, burn it, and re-install.. then repeat if something went wrong. I just leave the USB cable connected so I can blast a new image and switch to it in seconds. I program sections that are not the selected image while the Atari is on and just press the reset. When you program one of 8 slots you can choose partial burn, byte swap, and a number of other things. Its a lot more convienient than using eproms especially if you working on a TOS code.
Re: Smonson's slow H4 build
Posted: 13 Jun 2020 15:09
by Smonson
PaulJ wrote: 13 Jun 2020 15:02
@Smonson , you should consider giving @cmorley's eprom emulator a try.
Thanks, I'll keep it in mind. I prefer to make my own stuff though. It seems like overkill for what I need anyway.
Re: Smonson's slow H4 build
Posted: 13 Jun 2020 18:24
by troed
You can create your own TOS E00000 decoder with a GAL16V8 and some old schematics IIRC. I think I have those somewhere.
TOS_206.png
/Troed
Re: Smonson's slow H4 build
Posted: 14 Jun 2020 08:54
by mfro
Smonson wrote: 13 May 2020 11:05... Unfortunately Quartus 12 no longer runs on the latest Ubuntu and I don't have enough laptop disk space to install another OS so that's a bummer, I will be constantly walking between my PC and the other room to try stuff. A very annoying decision on Intel's part for taking Cyclone-II support out of Quartus at version 12.1.
My understanding is that 13.0pl1 is the latest version that supports CII?
I'm not on the absolute latest (using Debian) but for now, I always got Quartus to work with some fiddling. Last time I installed Ubuntu 12 in a KVM, installed Quartus there, verified which shared libraries it drew (using ldd) and copied all of them into the Quartus bin directory of the native machine which worked fine.
Re: Smonson's slow H4 build
Posted: 14 Jun 2020 09:30
by Smonson
Unfortunately 12.0sp2 is the last version for cyclone-2. It's a 32-bit application and that's the reason why the new Ubuntu can't run it. I can still run it on Windows actually - I didn't even think about that when I first mentioned it - probably because using Windows for software development is even more inconvenient than walking from room to room ;)
Eventually I'll get a bigger SSD for my laptop, then I can run it in a VM.