Mongrel edition motherboard thoughts
Posted: Tue Oct 23, 2018 2:16 pm
I have a lot of SMT GLUE & MMU currently.. I want to try and make a adapter PCB so they can be used in place of PLCC chips ultimately.. Because probably won't be particularly cheap, but as there is no real alternative currently then it is the only way to supply replacement chips..
But I was thinking that while the motherboard will mostly have FPGA chips replacing them, it is likely going to be a long way off being completed.
So my thought was I could do a inbetween motherboard, which would basically be similar to the board how it is now, but using the SMT GLUE & MMU.. In fact a lot of parts would be SMT on this board, so likely I could actually get these boards assembled to a point myself. Likely I could add in built on DRAM (4MB) would be SMT also.
The thought mostly here is that as like Troed was doing with the DoubleST, that the remake board should be a lot more stable with such a accelerator. Plus the MMU & GLUE are the latest production batch of those chips so likely to be more overclockable than the previous PLCC types.
The main problem with such a overclocked board, is the amount of work it actually needs doing to a original ST where it generally isn't terribly stable overall. But on this new motherboard, it should be more stable and should not really take much more than a small PLD to do the logic for it.
Those who have no idea what I'm talking about, the ST would run at 200% speeds.. ST-RAM runs at 200% as the MMU is overclocked. One side-effect is that the shifter doesn't seem to play nice in mono mode anymore according to Troed's tests so far.
I am not planning on creating such a board any time soon, it is a lot of work, but I see the FPGA stuff taking a lot longer to sort out the way things are going. There is still a lot of mods and things to do on the current board so it will likely be another year even before the current design is finalised finished.
I still have a lot of booster work to finish off yet, so really I need to complete that work before investing a lot of time in the FPGA stuff.
So the mongrel board, would basically be a more SMT based machine with integrated booster and RAM. As there will be a PLD onboard, it can also have TOS206 decoding built in. This chip I would call the "Super GLUE"
Ultimately the FPGA GLUE would have all the modifications built into one chip, but again I can just see many problems lining up delaying the next gen board literally by years.
So I think in order to keep work for me down to something reasonable, I think this mongrel board would be the next logical step. I would add in a 68000 DIP socket to take in my SEC booster once completed, and we could probably add in some extra stuff along the way has already talked about..
This does not necessarily mean the current Alpha board would become obsolete. Simply because there would have to be sockets or holes to be able to change the MMU/GLUE etc once the FPGA chip replacements are done. Of course the mongrel edition being more SMT based, would mean the chips would not be changeable.
So I am just thinking out loud more than anything. Of course any modifications or upgrades done on the alpha would go on to the next revision board whatever happens anyway.
But I was thinking that while the motherboard will mostly have FPGA chips replacing them, it is likely going to be a long way off being completed.
So my thought was I could do a inbetween motherboard, which would basically be similar to the board how it is now, but using the SMT GLUE & MMU.. In fact a lot of parts would be SMT on this board, so likely I could actually get these boards assembled to a point myself. Likely I could add in built on DRAM (4MB) would be SMT also.
The thought mostly here is that as like Troed was doing with the DoubleST, that the remake board should be a lot more stable with such a accelerator. Plus the MMU & GLUE are the latest production batch of those chips so likely to be more overclockable than the previous PLCC types.
The main problem with such a overclocked board, is the amount of work it actually needs doing to a original ST where it generally isn't terribly stable overall. But on this new motherboard, it should be more stable and should not really take much more than a small PLD to do the logic for it.
Those who have no idea what I'm talking about, the ST would run at 200% speeds.. ST-RAM runs at 200% as the MMU is overclocked. One side-effect is that the shifter doesn't seem to play nice in mono mode anymore according to Troed's tests so far.
I am not planning on creating such a board any time soon, it is a lot of work, but I see the FPGA stuff taking a lot longer to sort out the way things are going. There is still a lot of mods and things to do on the current board so it will likely be another year even before the current design is finalised finished.
I still have a lot of booster work to finish off yet, so really I need to complete that work before investing a lot of time in the FPGA stuff.
So the mongrel board, would basically be a more SMT based machine with integrated booster and RAM. As there will be a PLD onboard, it can also have TOS206 decoding built in. This chip I would call the "Super GLUE"
Ultimately the FPGA GLUE would have all the modifications built into one chip, but again I can just see many problems lining up delaying the next gen board literally by years.
So I think in order to keep work for me down to something reasonable, I think this mongrel board would be the next logical step. I would add in a 68000 DIP socket to take in my SEC booster once completed, and we could probably add in some extra stuff along the way has already talked about..
This does not necessarily mean the current Alpha board would become obsolete. Simply because there would have to be sockets or holes to be able to change the MMU/GLUE etc once the FPGA chip replacements are done. Of course the mongrel edition being more SMT based, would mean the chips would not be changeable.
So I am just thinking out loud more than anything. Of course any modifications or upgrades done on the alpha would go on to the next revision board whatever happens anyway.
Plus it is a STE chip, so it would be unable to do TOS104 decoding, at least I assume so.
) was happy with 64Mhz pixel clock (doubleST mono)
Busy sorting out the next PSU batch.. Amazed I've spend a grande on transformers just