MiSTer: MiST on Terasic DE10-nano board.

[Sorgelig]

To improve the core, you need to be familiar with original hardware(Archie). I never had or used this machine.
I didn't look into this core, so i didn't port it (yet).

Maybe it's worth to ask the core developer if he is willing to buy DE10-nano board (which costs only half of MiST price) and join the development.
I cannot handle all cores alone.

[Zarchos]

To improve the core, you need to be familiar with original hardware(Archie). I never had or used this machine.
I didn't look into this core, so i didn't port it (yet).

Maybe it's worth to ask the core developer if he is willing to buy DE10-nano board (which costs only half of MiST price) and join the development.
I cannot handle all cores alone.

Yes I understand all this and don't misunderstand me : I praise and worship your work and efforts.
Do you sell the board btw ? (Not kit but ready to use).
I'd buy 2, 1 for me and 1 for the Archie core maintainer.
I'd give it to him as an incentive and a 'thank you'.

[Sorgelig]

I don't sell anything, sorry.
You can ask the dev to estimate the resources required for Archie core. Since FPGA on DE10-nano is much bigger than on MiST, may be a naked DE10-nano will be enough for this core and then you can buy it directly from Terasic.

[Zarchos]

ok. I guess I'll easily find this Mr Terasic.

[Sorgelig]

It's not an Mr. It's a company manufacturing and selling DE10-nano: http://www.terasic.com.tw/cgi-bin/page/ ... 67&No=1046

[Sorgelig]

I've checked original CPC6128 plus. Base specs list 128KB of RAM, so it can be fully implemented in FPGA RAM without any daughter board. Even additional around 256KB of extended RAM/ROM can be implemented. So if 512KB of expanded memory is not very important then naked DE10-nano should be sufficient.

The core code of FPGAAmstrad is too cluttered and too exaggerated (if it possible to say so) - thus only original author can make a good port.

[NML32]

Something has changed with MiSTer FPGAGen. I loaded up Aladdin (U) [!].bin and I noticed the running animation isn't working. Aladdin just slides across the screen. I tested the same game on the MiST and the running animation works.
It appears to be broken after core FpgaGen_20170707.rbf

[Sorgelig]

It looks like the same metastability problem. I will recompile it.

[alexh]

How many clocks does the core have? A few async fifos here and there between domains?

[Sorgelig]

How many clocks does the core have? A few async fifos here and there between domains?

it's open source: https://github.com/MiSTer-devel/FpgaGen_MiSTer
Do you want to improve the core?

[alexh]

I don't really have the time but I could try an automated CDC analysis

[Sorgelig]

In FPGAGen there is no running animation in Alladin because of 6-button pad. Set "3 buttons only" -> Yes. It will fix the problem.

[Sorgelig]

In FPGAGen there is no running animation in Alladin because of 6-button pad. Set "3 buttons only" -> Yes. It will fix the problem.

I've pulled almost all my hair to understand where is the problem because after several tests it didn't look like metastability problem. So, probably This is one of games where 6-button pad has issues.

[Sorgelig]

@Zarchos,

heh, it looks like i've messed Archimedes and Amstrad cores Too busy recently... My comments was about Amstrad.
Archie core is well organized. I will see if i can port it.

[NML32]

In FPGAGen there is no running animation in Alladin because of 6-button pad. Set "3 buttons only" -> Yes. It will fix the problem.

Thanks, that fixed the problem and I made 3 buttons my default setting.

[NML32]

Does anyone else's MiSTer menu open randomly without the snowy background?

MiSTer_Menu.jpg

[Sorgelig]

I forgot when last time i saw this. Always with noise for me. Does it happen with all Menu releases?

[NML32]

I forgot when last time I saw this. Always with noise for me. Does it happen with all Menu releases?

I'll try some of the previous builds since I can't remember when I first noticed it.

Update: I think it started with menu_20170712.rbf
I went back to menu_20170707.rbf and rebooted several times and no issues. I went back to menu_20170712.rbf and the snow changed to the white background after a few reboots.

[NegSol]

As announced I soldered the SD-RAM daughter board today. Thank you Sorgelig for the great documentation and photos. Being a Noob when it comes to SMD soldering, I had some trouble, but with some persistence and an USB inspection microscope I got it working . I went for the horizontal setup.
I ran all memory test core Sorgelig kindly provided. The results where as expected. 60 MHz, 120 MHZ, 130 MHz and 140MHz worked without any red marks. 150 MHz I get the first 32MB test run marked as red and all following marked green. 167 MHz is an absolute not got - this was expected. Actually I was surprise it worked at all

Anyway here are some Vids:

https://youtu.be/XbhbmwGJuIE Mem Tests
https://youtu.be/XNCtpCTDdtk Board and Minimig Core
https://youtu.be/wzhTBHAM_6E Minimig Core - 9 Fingers
https://youtu.be/0zaarnVOrwE NES Core (20170712)

Ok - So I must say I really like the MISTer more and more thank you Sorgelig for the porting of the cores!

Here are some thing I notices (really early observations -all v20170712):

NES core works absolutely great - just like on MIST with all nice enhancements of the MISTer (e.g. joystick button assignments)
SamCoupe core works absolutely great - just like on MIST with all nice enhancements of the MISTer (e.g. joystick button assignments)

Minimig AGA core works great -Hard disks work and Whdload also work but there are some issues - here are the issue I noticed:
-the disk drives do not read any adfs (so this is a bummer)
-the configurations are not saved
-when I load old configurations from my MIST the configurations appear to get loaded, but the OSD stop working and will not be displayed
Workbench 3.1 whenever I open a window there are graphical glitches that follow the mouse pointer

here is a video showing the issues:
https://youtu.be/hhjHMt0XUHo Minimig core 20170712 issues


Still very impressive to see AGA working on the MISTer in this "early" stage.

[Sorgelig]

Congrats! You are the first (who is not me) who've soldered SDRAM board Did you solder AS4C chip i've recommended? Can you make photos of your SDRAM board?
I guess, you've didn't solder I/O board?

About Minimig: don't use settings from MiST - they are not compatible! I've re-checked ADF - they are working for me. May be you have problem with ADF because of settings from MiST.

[Sorgelig]

About cursor issues: it's known bug of Minimig on progressive modes. You need to choose highres mouse cursor to fix the glitch.

[Sorgelig]

On video it's hard to understand but it looks like you didn't trim the pins of connector on SDRAM board. Look at this pic: https://github.com/MiSTer-devel/Hardwar ... _horiz.jpg
After trimming the pins, you board will pass not only 150MHz but also 167MHz test

[NegSol]

Yes, I planned to trim the pins, but did not think it would make a big difference. I suppose it makes a difference on higher frequencies. I will trim the pins today and will give you feedback.
The mouse cursor glitches looked familiar to me. I used the same harddisk file I use for my MIST. There is a lot of experimenting ahead. I like it .
With the disks drives I am a bit clueless. Maybe the configuration. But I tested initially without and MISTer did not save any configuration file. Not even the default. Configuration were stored during core reset but did not survive a cold boot. Anyway I will investigate. Could be the default kick.rom I picked (2.0 or the like) maybe I will behave differently when I pick 3.1.

Thanks again.

[NegSol]

So - I tested the memory after trimming the pins. No change. 140 Mhz goes on forever without any errors. 150 Mhz has errors every 30 secs. 167 Mhz no go

Still very happy with my first attempt at this and smd soldering. Also 140 Mhz should be plenty, right?

Here are some pic of the hardware:

https://drive.google.com/file/d/0ByqDwa ... sp=sharing (board before trim)
https://drive.google.com/file/d/0ByqDwa ... sp=sharing (board ram chip 1)
https://drive.google.com/file/d/0ByqDwa ... sp=sharing (board ram chip 2)
https://drive.google.com/file/d/0ByqDwa ... sp=sharing (board other side)
https://drive.google.com/file/d/0ByqDwa ... sp=sharing (board mounted)

[Sorgelig]

140MHz is enough, sure.
I know where is the difference on your board: This type of 40pin connector should be soldered fully sit, and that 3pin connector need to be soldered without plastic part (can remove plastic later and trim pins a little). I will add more notes in PCB repository.
Anyway, your result is good already. No need to re-solder to avoid PCB/connector stress.