Raven. A homemade Atari-like computer

[frank.lukas]

... have no idea if NOVA supports those other two.

Nova driver support 98% of all ISA ET4000AX 1MB cards ...


I use the Targa Win + with the T8 Nova Driver in the Nova VME, STGA, Nova MegaST and VOFA Adapter without problems.

[danboid]

I'd like to see a FPGA based ISA card that has HDMI and VGA outputs on it, like a video card equivalent on the picogus.

If I suggest that idea enough key forums, maybe someone will do it?

[peters]

Dreamblaster arrived and was plugged in to the Milan SB16, no output yet.

I'll post what I discover on AF:

https://www.atari-forum.com/viewtopic.p ... 7&start=25

[agranlund]

I don’t see the light blink on the cf adapter soi’m not sure it’s actually doing something. It’s a cf 4gb with a 512mb partition fat.

Some weird things in the log:

Code: Select all

IDE i/f 0 device 0 detected
IDE i/f 0 device 1 detected
IDE i/f 0 device 0 is type 2
IDE i/f 0 device 1 is type 1

I would not expect device 1 detected and then presented as type1.
Mine is essentially this:

Code: Select all

IDE i/f 0 device 0 detected
IDE i/f 0 device 0 is type 2
IDE i/f 0 device 1 is type 0

And then we get a "Timeout in wait_for_not_BSY(0xfff00000,20)" when accessing device0 which is the one that should exist.
We do get something in terms of "unit 18 is managed by EmuTOS internal drivers: Disk" but when comparing against my own log output I am not getting those "timeout in wait" for that device, and I am also getting a whole lot of information about the MSDOS-FAT16 partition it finds on my card.
Perhaps it's all being freaked out by that bogus device it detect. Or perhaps not.

My strongest hunch at the moment is if it's related to 16bit access through the 68150 and those two bus-buffers right next to it.

The YM test seemed to work and that is essentially testing 8bit writes.
calibrate_delay() shows similar values to my log output so let's assume MFP timer is working fine too. Besides, you are getting rs232 output from it correctly.

IDE access would be 16bit, and the gfxcard will be set to 16bit access but perhaps not the initial detection.

Can there be a short, or no-connection, on some pin on those buffers?


I would have perhaps had the Narnia PLD on the suspect list since it helps with ISA stuff, but since there is also the disk thing and it is helped by the Iksi one I'd put that lower on the suspect list. Iksi also helps with YM and MFP which appears to be working fine.

And just to get it verified, I switched out my 40mhz 68150 for a 33mhz version and at least in my machine, it still brings me to the desktop being overclocked to 48mhz - I did not test much else after that.

All videocards work in a 486dx2. Just tried them to be sure.

Aah, dx2.. I had one of those dx2-66Mhz Compaq all-in-ones, with an external and very slow CD-ROM drive. So many fond memories of that machine. Played Diablo1 on it so much and thought it was perfectly fine until I saw how it was meant to look like on a friends dx4-100mhz :lol:

[Oldskool]

Not sure.
My eyesight is not 100% anymore but I don’t think I have shorts. My big soldering iron does not fit in that gap so maybe need to ask a friend to do a reflow however all legs seem “solid”.
Concerning Narnia and Iksi. Used new ic’s and the tl866 (did not change the defaults in the programmers software).

IMG_1771.jpeg

But indeed good that you found a difference. Tried various cf cards but output over serial is the same.

[agranlund]

My eyesight is not 100% anymore but I don’t think I have shorts. My big soldering iron does not fit in that gap so maybe need to ask a friend to do a reflow however all legs seem “solid”.

Without a microscope it's probably easier to "beep" the 68150 side on the plcc socket for the 68150, for verifying shorts on the databus - all those buffer outputs go there anyway. I wouldn't expect issues on the 68060 side since ROM and RAM appears to be working.
There is the /OE signal too I suppose and if it is shorted or not connected on one of the buffers then that particular byte of the word would break.

There is the databus between 68150 and all the peripherals too but it's only connected to through-hole stuff so I'd recon it's a lot harder to short anything on that one?


Give this one a test and see what happens, with EmuTOS drivers, I tried to make it do 8bit I/O to the gfxcard - not sure if it's going to help or if I actually succeeded. I do get picture here though :)

Also got rid of most EmuTOS debug output except for VGA and IDE stuff..

rom_tos_dbg2.bin.zip

[stephen_usher]

What would be ideal is an ISA card which has a GEM VDI driver on-board in ROM and exposes dual-ported RAM, with one area overlaying the ST graphics addresses in bus snooping mode for compatibility. :-)

[agranlund]

Give this one a test and see what happens, with EmuTOS drivers, I tried to make it do 8bit I/O to the gfxcard - not sure if it's going to help or if I actually succeeded. I do get picture here though :)

Or this with 8bit ram access too. It's not going to like it and the desktop wont look correct but as a test.

rom_tos_dbg3.bin.zip

[stephen_usher]

Oh, I forgot that the ISA bus is only 8 bit.

[agranlund]

Oh, I forgot that the ISA bus is only 8 bit.

No it's 16bit, but yeah that's still only half of what you'd ideally want :)