exxos's Atari, Amiga & retro forum

I know some people moaned about no though-expansion on DFB1X. But what would actually plug into it anyway ? Is it worth re-visiting the idea ? I assume not ?

The port would be a "clone" of the Falcon expansion bus.Likely would have to sit "above" the current falcon connectors.. it would be a PITA to route it in, hence asking before I even considering bothering..

The Nova Falcon adapter to be able to use a graphics card like the Mach 32/64 would be one usecase (it apparently works with the DFB1: viewtopic.php?t=5538 )

Or the FalconSpeed PC emulator card.

If it could mirror the form and functionality of the one on the CT6x it could facilitate the Super Videl / 2 or the Mega Videl which is being developed now

alexh wrote: 23 Sep 2026 20:34 If it could mirror the form and functionality of the one on the CT6x it could facilitate the Super Videl / 2 or the Mega Videl which is being developed now

Isn't the CT60 expansion bus totally different than the Falcon expansion bus ?

I can only physically fit a duplicate of the Falcon expansion on the card, and that's going to be a PITA to route. Its possible the bottom smaller connector could be moved or duplicated again to give a bit more room, but that would need a custom card layout, so not really worth the trouble of having a 3rd layout.

don_apple wrote: 23 Sep 2026 20:09 The Nova Falcon adapter to be able to use a graphics card like the Mach 32/64 would be one usecase (it apparently works with the DFB1: viewtopic.php?t=5538 )

Or the FalconSpeed PC emulator card.

Interesting, looks like it would actually fit even if it was moved upwards slightly..

alexh wrote: 23 Sep 2026 20:34 If it could mirror the form and functionality of the one on the CT6x it could facilitate the Super Videl / 2 or the Mega Videl which is being developed now

I second this idea...however adding the original Falcon expansion socket too would be good. Is both feasible?!

As alexh said the Mega Videl is being developed (by foft) who said about supporting other accelerators:

DFB1 etc: In theory. Some small rtl changes + it needs to be 3.3v or add level shifters, since the fpga is not 5v tolerant. Lets focus on one thing at a time though.

viking272 wrote: 23 Sep 2026 21:15 I second this idea...however adding the original Falcon expansion socket too would be good. Is both feasible?!

There is not physically room to do that without totally redesigning the whole card unfortunately. You would need a card the size of the CT60 to get it all to fit. I don't even know why the larger connectors even exist in the first place as opposed to just copying the Falcon bus ?

Getting another Falcon expansion port on there is already near impossible.. That is why I was asking if it would actually be useful to add a expansion port on and if anything can actually use it in the first place..

SV clearly isn't going to fit the requirements here, but I'd assume it it need the 060 anyway. Running a 030 on SV probably would not be that great.. But I don't know anything about that stuff and never used it..

I don't know anything about the Nova stuff either but obviously it has been mashed in as pointed out already. Is there a schematic for what that adapter board does ?

I suspect even just the pin headers alone would stop the keyboard going back in. There's no room to board north without hitting the power supply. Even thinking about the routeing makes me itch.

The piggy back that Frank did is, I think, the best option. You have to accept it's not going to fit in the case at that point, so just go up.

As to what can use it. Only things that don't need DMA. Nothing can sit on top of the DFB1 and become bus master. The high priority bus grant pins are shorted together and the low priorty one is used by DFB1.

So: things like graphics cards and FastRAM boards, sure. Perhaps an ethernet adapter, if it doesn't do DMA. Perhaps a coprocessor that has its own RAM and is polled (last one is purely theoretical AFAIK)?

FalconSpeed is out as it runs as a bus master.

I have idly thought about what would be needed to add support for daisy chaining the bus grant, but it's all theoretical ATM. The approach I settled on in my mind was actually not to modify DFB1 at all and instead consider a new relocator and expander board:

Basically, accept that if you want to fit two expansions you're going to have to sacrifice the PSU, as per CT60, and so have a board that plugs into the expansion and provides two new expansion connectors in the PSU space (perhaps rotated 90 degrees?). This interposer/adapter would require logic to pass through the bus requests from each board and allocate the bus grant to the correct requestor.

...in theory... ;)

BW

Yep that's the problem but like was posted in another thread, the port locations isn't a factor at this point.. The DFB1X barely fits as it is.


Is why I asked newly if there was a schematic or something for that adapter board so I could actually see, but I don't think it would actually hit the power supplied by the looks of that image alone.. But of course the height is another factor.. But both are basically relevant with what's hanging off it all anyway.

Routing wise yes it's rather a significant chunk of a redesign.. I did actually start it about a year ago and got part way through it before I started to think is it really worth the hassle adding it.

exxos wrote: 24 Sep 2026 11:51 Routing wise yes it's rather a significant chunk of a redesign.. I did actually start it about a year ago and got part way through it before I started to think is it really worth the hassle adding it.

I remember having left-right signal lines packed in as tightly as possible on both signal layers north of the top adapter to fit the footprint. I held my nose and did it. :P

EDIT: ah, no, I did manage to keep them *mostly* one layer, so there is actually more theoretical room than I remembered. Those vias on the address lines, though... :lol:

BW