rpineau wrote:I just finished the first card. I'm in the process of testing all the connection for defect (short circuit, broken track, ...) before plugging it on the Falcon.
I'll start the 2nd card once the first one is tested.
rpineau wrote:My guess is the Falcon is not going to start because of bad instruction in the TOS. But I'll put my logic analyzer on the address pin to see if the 68060 fetches the first few instruction.
I'll try to post a picture later today.
rpineau wrote:Yes, that was my guess too regarding the caches.
AS for booting to the desktop, the 68040 has all the 68030 instruction, whereas the 68060 is missing some (software implemented in the Motorola's 68060_lib) which is why I think it might crash before reaching the desktop, but testing will tell.
If this is the case we might need to patch the TOS and program an EPROM for the testing. Something that change the PC pointer at the beginning of the TOS to the extra code (there is space at the end of the TOS) and then jump back to the regular TOS init.
I have 2x 27C4096 EPROM (100ns) so I can totally program one with the extra code (if we end up doing this I'll send you one with the card).
Once this is working we can program some 27C4096 OTP with the patched TOS for the people who make their own adapter (the OTP version are easy to find and only cost $7 and are available in 90ns and 55ns (same price).
Code: Select all
DIVU.L <ea>,Dr:Dq 64/32 ⇒ 32r,32q
DIVS.L <ea>,Dr:Dq 64/32 ⇒ 32r,32q
MULU.L <ea>,Dr:Dq 32*32 ⇒ 64
MULS.L <ea>,Dr:Dq 32*32 ⇒ 64
MOVEP Dx,(d16,Ay) size = W or L
MOVEP (d16,Ay),Dx size = W or L
CHK2 <ea>,Rn size = B, W, or L
CMP2 <ea>,Rn size = B, W, or L
CAS2 Dc1:Dc2,Du1:Du2,(Rn1):(Rn2) size = W or L
CAS Dc,Du,<ea> size = W or L, misaligned <ea>
rpineau wrote:I agree with you that the MUL and DIV are the dangerous ones so we'll see how far it goes (if I haven't made any hardware mistakes).
rpineau wrote:Sorry about the lack of news.
I've been busy with work and my other hobby (astronomy).
I should have time this weekend to finish to test all the connection and I should be able to do the first test on the Falcon itself.
rpineau wrote:The next step is to put back the AB060 (no longer and AB040 ) in the Falcon and see if it starts. I'll probably won't have time to test that before next weekend.
rpineau wrote:Here are some picture of the adapter and the AB060 (you'll notice the hack to put the 74FCT157 SOIC-16 on the board... as I made it for a SO-16 version) :
rpineau wrote:So I have some better news ... the 68060 starts.
rpineau wrote:Looks like my reset on the Falcon is flaky and doesn't always go back up to +5v and get stuck at 0v.
rpineau wrote:On the attached screenshot we can see the 68060 fetching data as it cycle through some of the addresses :
On strange thing is the DTACK no doing a lot in there (this one if from the Falcon connector so it's DTACK going to the F030 mother board. I assume the TA* is directly generated from the MACH210 to the 68040/68060 but I would have assume the Combel would generate DTACK* for each ROM access by the CPU.
rpineau wrote:SO at this point it looks like the card works to some level and there might be some instruction crashing the 68060 in the early stage of the TOS (I'll disassemble the TOS to check, or at least look at it in mon30).
I think I need to do some fixes on my Falcon mother board before I can move forward with this.
rpineau wrote:@Doug : I'll contact you directly to ship you this board as it seems to be safe to use with a AB040 and a Falcon. Given your soldering skills if we find some hardware issue I have no doubt you can fix them .
Code: Select all
E00000: bra.s $00E00030
E00002: subi.b #$E0,d4
E00006: ori.b #$E0,$00(a0,d0)
E0000C: dc $0000
E0000E: or.l d7,(a2)+
E00010: dc $00E0
E00012: ori.b #$E4,$24(a0,a1*4)
E00018: movep $1993(a0),d1
E0001C: dc $00FF
E0001E: dc $1A68
E00020: dc $0000
E00022: dc $56BA
E00024: dc $0000
E00026: move.b d7,$00(a0,d0)
E0002A: bgt.s $00E00010
E0002C: ori.b #$00,d0
E00030: move #$2700,sr
E00034: move $FFFF8006.w,d0
E0003A: move $FFFF8006.w,d0
E0003E: move #$0007,$FFFF8940.w
E00044: cmpi.l #$FA52235F,$00FA0000
E0004E: bne.s $00E0005A
E00050: lea $00E0005A(pc),a6
E00054: jmp $00FA0004
E0005A: lea $00E00064(pc),a6
E0005E: jmp $00E7B000
Users browsing this forum: joska and 4 guests