ST536 STE EDITION

[exxos]

Sorry @exxos it's taken me so long but finally tonight I managed to swap out the Xilinx. Gonna check it for shorts tomorrow and potentially program it.

:thumbup:

[exxos]

Killed my CF card :roll: think I plugged it in one pin to high :roll: PC wont even find it now :roll:

Madness price of some even used ones. Guess I could try some noname china brands and see if they work..

[coonsgm]

I'm pulling my hair out here reflowing solder points checking for any pin issues. Everything seems good...checked for shorts, checked for non-connected pins going around the PLD about 3 times pin by pin.

I think I'm going to take a break and come back to it in a few days.

[exxos]

I'm pulling my hair out here reflowing solder points checking for any pin issues. Everything seems good...checked for shorts, checked for non-connected pins going around the PLD about 3 times pin by pin.

You need to use YAARTTT as you should be able to find out if there is a bad bit or not.

If you have STOS it's easy to do tests there..

Code: Select all

loke ($2000000),12345678
print leek($2000000)

https://exxosforum.co.uk/atari/STOSFLOPPY/STOS1904.zip

[exxos]

@coonsgm just thought, have you hot the 33pf cap on the CPU ?

viewtopic.php?p=128013&hilit=Bad+clock#p128013

[coonsgm]

I'll try and take a look tonight!

[exxos]

Is this code doing what I think it's doing ?

Code: Select all

address <= shadow_enable ? shadow_row : { A[23:11] };

address = A[23:11] when shadow_enable = 0
address = shadow_row when shadow_enable =1

??

EDIT

AND:

Code: Select all

wire [12:0] shadow_address = {5'b00111, A[18:11]}; // Maps to 0x4F00000-0x4F7FFFF

??

[exxos]

I don't get it :pullhair: :pullhair:

TOSCOPY.PRG is run, it copys TOS206 space to $4F00000 (copy verified in STOS)
Then the PRG accesses 4F80000, which triggers copy_done to go high (also verified)

In main top I had to patch the RAM decode with & (TOSCE | ~copy_done | AS30);
which should make ram_decode =0 on TOSCE =0, copy_done =1, AS30 = 0

Code: Select all

assign TOSCE =  ( A[29:20] != 10'hE & A[29:20] != 10'h3FE & A[29:3] != 27'h0 ) | ~RW30  | (AS30 & AS) ; 
wire ram_disable = 1'b0; // 1 = disable TTRAM
wire ram_area = ((A[29:28] != 2'b00) | (A[27:24] == 4'b0000)) & (TOSCE | ~copy_done | AS30);
assign ram_decode = ram_disable ? 1'b1 : ram_area;

ram_decode is 0 on match

As when copy_done is high, and TOSCE = 0 , we are accessing ROM (which is now in TTram)
ram_decode going low also sets the CPU to 50mhz.

then in SDRAM module.

Code: Select all

// Address translation: Redirect TOS206 addresses to 0x4F00000-0x4F7FFFF
wire shadow_enable = copy_done & ~AS30 & ~TOSCE; // Enable for TOS206 addresses
wire [29:0] A_INT = shadow_enable ? {7'b0011110, A_IN[22:0]} : A_IN; // Translate to 0x4F00000

Basically "done"...

BUT it locks up when copy_done goes high, so I assume the address translation isn't working, or the CPU isn't in 50mhz mode during the shadow rom access. Or the idea simply doesn't work, or i've done something dumb.

I have the 16bit ROM disabled when copy_done = 1. The 536 databuffer is turned off. I even ignore STE ROM DTACK. I even stop the STE from seeing ROM access.. I even stop my own ROM DTACK from going going low (as it would likely conflict with the SDRAM logic STERM otherwise).

:shrug: :pullhair: :pullhair: :stars: :stars: :headbang:
:dizzy:

[PhilC]

@exxos I should have a bit of spare time over the next few days now I've got the Raven working.

Re the rom copy, can you not do this with the xilinx?

[exxos]

Re the rom copy, can you not do this with the xilinx?

Copying that address range would require a megaload of counters ... It's best to let the CPU do it. But the copying works fine anyway.. it's when it switches from ROM to TTram it locks up.

I am out of ideas...I'll probably have to resort to adding a small SRAM and just bank switch the thing.