exxos's Atari, Amiga & retro forum

@JezC let's hope not.

PhilC wrote: 08 May 2025 21:44 @JezC let's hope not.

:crossed1:

So a bit of news, I checked the Mach32 in a PC and it does work, so not sure why it doesn't want to play on the Raven.

Also not sure why the Raven only detects 256K on the Et4000. I'll have a look at the ISA schematic and see if something is amis to cause the blank lines on the ET4000 and the Mach32 to not work.

It could be power related, so I have ordered a Corsair PSU to arrive tomorrow to rule it out.

Congrats on nearly there! :)

It kind of looks like it’s missing the lower byte on those 16bit accesses to framebuffer. Which, I suppose could mean any number of things so far.

Would be interesting to see what result you get from pressing the NMI switch and running these two commands from the terminal:

vga init
vga test

(You’re supposed to get a 320x200x8bpp test image)

Interesting about the memory report. My card also report wrong; 512kb instead of 1MB. Perhaps because of the minimal x86 emulation running vgabios.
My card reported wrong on the MS-DOS PC that the ebay seller had too though.

@agranlund vga init displays the VGA graphics card details, the same as it would in dos.
vga test then displays an RGB test screen as below:

I've ordered a couple of different graphics cards to try out, including a Trident 8900, another ET4000 and a WDC card. I'd like to learn.how to make these work once my Raven is fully working.

Looks very promising - nice work :)

dml wrote: 09 May 2025 20:44 Looks very promising - nice work :)

Thanks. The hardest parts are the small surface mount chips. I'd recommend doing all the surface mount stuff first and would do that way if I was building another Raven.

PhilC wrote: 09 May 2025 16:34 vga init displays the VGA graphics card details, the same as it would in dos.
vga test then displays an RGB test screen as below:

I'm strongly suspecting something related to word access since that worked and the TOS desktop isn't quite.
(also, Mach32 relies on that which could explain why that card is not working for you)

I would give this a try to collect further info:

Start with initing the test image: Then try reading the first word of the last red line which should return 3f3f. if it doesn't then let's try each byte, which should return 3f and 3f. Now test word write:
This should produce a blue + green pixel at the beginning of the last red line. And read it back to verify. first by word and then by byte. If you didn't get the blue+green pixel, or the word readback value was wrong then let's double check by writing each byte individually then try word + byte reads. If everything checks out you may as well verify long write+reads while at it :)

@agranlund thanks, will give that a go and report back.