NewBrain on FPGA running and alive.

[rondc]

Up

1

Down

::

With the help of Chris Desp’s NB-Emu and Mame/Mess code, plus the schematics and documentation we have collected from NewBrain, Shaeon [a member of RW-fpga dev Team] embarked on the adventure and we finally have a working NewBrain core.

The core was developed on the Calypso board, later ported to Poseidon, SiDi, and Senhor boards. Except for Calypso, which is a cyc-1000, the rest are Altera Cyclone IV and V.

It’s currently running on four different FPGAs and loading audio. It has a TZX, BAS, and BIN file loader. It can be configured with up to 2 MB of RAM, the COP420 can be used in emulated or real mode, and it already has almost everything needed to get the expansion and disk drives up and running.

You can watch the video here: https://www.twitch.tv/videos/2879520912
From minute 56.

Will update with all information, git, binaries and all of this.
Thank You
Regards

 

Attachments:

1. 

   snake.jpg
   

Topic

[CDesp]

Up

0

Down

::

Great implementation.

Tape files with binary follow ups should automatically change to the next file after each file loads. Like Pacman.bas, it needs to load the pac.screen.bin and the pac.bytes.bin. So it should probably go to a seperate folder and somehow put the files in order, or use a container like .tap.

There are also .bas files that are disk files (cp/m) like demo30.bas. this one loads(merge) other bas files depending on the menu you choose, so it needs the cp/m, either the small one or the full with the Expansion. In my fpga i use a new 8K rom on bank 4 that implements a new device 12,13,14,15 (like cp/m) that loads from the sdcard.

Newbrain uses all devices the same way as streams so you can replace any device with another.

If you need help with the cp/m let me know.

[rondc]

Up

0

Down

::

Hi CDesp
Now Disks are working, one step more done. I hope it gets finished very soon.-

 

 

Attachments:

1. 

   nb-dsk-core.jpg
   

[rondc]

Up

0

Down

::

Here is the latest video, with disk controller and running CP/M

https://www.twitch.tv/videos/2882097737

From the beginning. We’re making the final corrections, debugging, and the source will be available on Git shortly.

I’ve only been able to find a few CP/M disk images in various formats; is anything preserved in disk format?

[CDesp]

Up

1

Down

::

You can try the files that newbrain emulator creates when you select a directory, it should work.

Run newbrain emulator with CPM, select a directory, then dont close the emulator, goto the disc directory and there should be a 800kb disk file that you can copy.

[rondc]

Up

0

Down

::

Hello. I’ll give it a try. Thx
more improvements

 

 

Attachments:

1. 

   nbmainmenu.png
   

[CJ Palmero]

Up

0

Down

::

I have finally made the core public here:

https://github.com/shaeon/newbrain

Any advice would be welcome.

Regards.

[CDesp]

Up

0

Down

::

Nice work.

Newbrain uses busreq/busack to paint the screen so the real speed is a lot less than 4MHz.

That is the reason if you play chess with the fast setting on it disables the screen to go full speed.

So for some games like the crossing  the speed should be reduced otherwise it is almost unplayable.

[rondc]

Up

0

Down

::

Finally shaeon has published the core at git:

https://github.com/shaeon/newbrain

For the fpga boards:
MiST
MiSTica
SiDi
Poseidon

and Senhor ( mister clone )

[newbrainer]

Up

0

Down

::

(Reposting this message in the forum, in case someone haven’t read the news)

I would appreciate a lot if someone having an actual NewBrain computer to run the following basic program:

300 CALL 62383
400 K=0
500 K=K+1
530 A=K^2
540 B=LOG(K)
550 C=SIN(K)
600 IF K<100 THEN 500
700 CALL 62399,MS
750 PRINT “ELAPSED SECONDS: “,MS/50
800 END

Ideally, I would like from you to run it at least 4 times and post your results here.
That will help us to adjust the core in order to be more accurate!

Thanks in advance!

[CDesp]

Up

0

Down

::

VFD on
Run 1 elapsed 6.98
Run 2 elapsed 7.06
Run 3 elapsed 7.14
Run 4 elapsed 7.2
Run 5 elapsed 6.4 (after put 31 that clears the screen)
So it depends of how many chars the screen is showing. That is the busrq/busack. if a line has a1 char then the video circuit enables the cpu after that char, it does not wait for the line to end.
with put 31:run elapsed 6.38 every time no chars on screen
with vfd off the same results so the only the chars on the screen affects the result.
with list (so the listing is displayed) i get 7.3 secs

[newbrainer]

Up

0

Down

::

Thank you very much Chris!

Your feedback is exactly what I wanted!

 

[newbrainer]

Up

0

Down

::

That’s the outcome:

NewBrain FPGA core

😉

[Author]

Replies