TF536 + 68000 relocator and ROM board project.

Other boosters or variants.
User avatar
stephen_usher
Posts: 5580
Joined: Mon Nov 13, 2017 7:19 pm
Location: Oxford, UK.
Contact:

Re: TF536 + 68000 relocator and ROM board project.

Post by stephen_usher »

Oh yay. It's started bombing as soon as there's a floppy access now, even in the diag ROM.

Swapped back to the Mk.I just to make sure that the base machine's OK and everything's fine. Phew!
Intro retro computers since before they were retro...
ZX81->Spectrum->Memotech MTX->Sinclair QL->520STM->BBC Micro->TT030->PCs & Sun Workstations.
Added code to the MiNT kernel (still there the last time I checked) + put together MiNTOS.
Collection now with added Macs, Amigas, Suns and Acorns.
User avatar
PhilC
Moderator
Moderator
Posts: 6016
Joined: Fri Mar 23, 2018 8:22 pm

Re: TF536 + 68000 relocator and ROM board project.

Post by PhilC »

@stephen_usher might be time to get the scope out perhaps?
If it ain't broke, test it to Destruction.
User avatar
stephen_usher
Posts: 5580
Joined: Mon Nov 13, 2017 7:19 pm
Location: Oxford, UK.
Contact:

Re: TF536 + 68000 relocator and ROM board project.

Post by stephen_usher »

I have been looking and I can't see anything obvious. Maybe the address/data lines are a tiny bit more noisy but that's about it. It's probably tipped something just over the edge when accessing the DMA chip, which is what the floppy hangs off. It also doesn't see the UltraSATAN on the ACSI bus (it doesn't even poll it). This is all with the 68000 only on the board.

I've by-passed the /HALT GAL so that's not the issue and changing the address decode GAL code to the original doesn't help either.

However, it does pass all the DIagROM tests which don't require the cable harness, except for the floppy, which crashes on first access.
Intro retro computers since before they were retro...
ZX81->Spectrum->Memotech MTX->Sinclair QL->520STM->BBC Micro->TT030->PCs & Sun Workstations.
Added code to the MiNT kernel (still there the last time I checked) + put together MiNTOS.
Collection now with added Macs, Amigas, Suns and Acorns.
User avatar
stephen_usher
Posts: 5580
Joined: Mon Nov 13, 2017 7:19 pm
Location: Oxford, UK.
Contact:

Re: TF536 + 68000 relocator and ROM board project.

Post by stephen_usher »

More specifically it's when it reads from the floppy.

I may try putting some pull-ups on the data bus as that looks the most scrappy.

Having said that I'm now getting memory errors again, and bits being set on the screen which look a bit like address offset errors.
Intro retro computers since before they were retro...
ZX81->Spectrum->Memotech MTX->Sinclair QL->520STM->BBC Micro->TT030->PCs & Sun Workstations.
Added code to the MiNT kernel (still there the last time I checked) + put together MiNTOS.
Collection now with added Macs, Amigas, Suns and Acorns.
User avatar
stephen_usher
Posts: 5580
Joined: Mon Nov 13, 2017 7:19 pm
Location: Oxford, UK.
Contact:

Re: TF536 + 68000 relocator and ROM board project.

Post by stephen_usher »

I've soldering in the second 68000 and am running tests on that now. It's a lot better now, but there is some memory corruption still.

The floppy works. It's definitely something taking the machine beyond it's operational range.

Next will be pull-ups, when I can remember where I put the bag of matrix resistors.

P.S. Running YARTT seems to show the errors being all 1s where they should be zeros, which would suggest the line isn't high enough when the memory or the processor think that the data's stable. This would suggest that the data lines need pulling up more than they are.
Intro retro computers since before they were retro...
ZX81->Spectrum->Memotech MTX->Sinclair QL->520STM->BBC Micro->TT030->PCs & Sun Workstations.
Added code to the MiNT kernel (still there the last time I checked) + put together MiNTOS.
Collection now with added Macs, Amigas, Suns and Acorns.
User avatar
stephen_usher
Posts: 5580
Joined: Mon Nov 13, 2017 7:19 pm
Location: Oxford, UK.
Contact:

Re: TF536 + 68000 relocator and ROM board project.

Post by stephen_usher »

Putting the two leads of the scope at either ends of the data lines to compare the traces doesn't show much difference. At the far end the high values are ver-very slightly attenuated and that's it.

The main difference, again, is the amount of noise on the power and ground. I'm guessing that the ground is causing the power fluctuations through the coupling capacitors.

I've only one ground trace which goes anywhere near anything active and that goes the long way around one of the processors and next to the MMU.
Intro retro computers since before they were retro...
ZX81->Spectrum->Memotech MTX->Sinclair QL->520STM->BBC Micro->TT030->PCs & Sun Workstations.
Added code to the MiNT kernel (still there the last time I checked) + put together MiNTOS.
Collection now with added Macs, Amigas, Suns and Acorns.
User avatar
stephen_usher
Posts: 5580
Joined: Mon Nov 13, 2017 7:19 pm
Location: Oxford, UK.
Contact:

Re: TF536 + 68000 relocator and ROM board project.

Post by stephen_usher »

It's definitely a ground noise problem.

I cut the ground traces to the CPUs and wired up a direct wire from the pins going into the main board to the "68000" socket. The noise was a lot less and no more memory errors.

There's a long trace which goes around the top of the "68000" socket which seems to be picking up huge amounts of noise so I'm going to cut that one and wire up an alternative route and see how much that cuts the noise further.
Intro retro computers since before they were retro...
ZX81->Spectrum->Memotech MTX->Sinclair QL->520STM->BBC Micro->TT030->PCs & Sun Workstations.
Added code to the MiNT kernel (still there the last time I checked) + put together MiNTOS.
Collection now with added Macs, Amigas, Suns and Acorns.
dalek
Posts: 224
Joined: Thu Nov 08, 2018 11:03 am
Location: NSW Australia

Re: TF536 + 68000 relocator and ROM board project.

Post by dalek »

Your board doesn't seem to have a ground fill?
User avatar
stephen_usher
Posts: 5580
Joined: Mon Nov 13, 2017 7:19 pm
Location: Oxford, UK.
Contact:

Re: TF536 + 68000 relocator and ROM board project.

Post by stephen_usher »

I know. That was intentional as the last board which did have it was picking up a huge amount of noise from it, and it made it impossible to cut the ground connections for experimentation. Also, where most of the traces go there's no room for the fill. It would only really be able to be done under the CPU sockets themselves.

I've been cutting ground traces and re-routing wires to see if I can cut down on the noise. It helped a small amount. However, adding a second set of 0.01uF capacitors between the second ground and power pins on the CPUs has cut the problematic peaks from both CPU sockets.

Unfortunately, I still can't get a CPU stable in the "Accelerator" socket.

I could really do with some inductors to dampen the other high frequency noise though.

P.S. I'm guessing the trace which is going underneath the CPU die is picking up the noise.
Intro retro computers since before they were retro...
ZX81->Spectrum->Memotech MTX->Sinclair QL->520STM->BBC Micro->TT030->PCs & Sun Workstations.
Added code to the MiNT kernel (still there the last time I checked) + put together MiNTOS.
Collection now with added Macs, Amigas, Suns and Acorns.
User avatar
stephen_usher
Posts: 5580
Joined: Mon Nov 13, 2017 7:19 pm
Location: Oxford, UK.
Contact:

Re: TF536 + 68000 relocator and ROM board project.

Post by stephen_usher »

I'm now feeding the CPU sockets directly with big, thick wires on the ground and they are still generating huge spikes, but only whilst the CPU is running.

Hmm... If I 'scope /BGACK on the second CPU it won't boot. I'll have to investigate.
Intro retro computers since before they were retro...
ZX81->Spectrum->Memotech MTX->Sinclair QL->520STM->BBC Micro->TT030->PCs & Sun Workstations.
Added code to the MiNT kernel (still there the last time I checked) + put together MiNTOS.
Collection now with added Macs, Amigas, Suns and Acorns.
Post Reply

Return to “EVERYTHING ELSE”