HW shortage ?

Pagina 15/16
8 | 9 | 10 | 11 | 12 | 13 | 14 | | 16

Van bartl68

Expert (86)

afbeelding van bartl68

25-09-2010, 16:17

Well all nice, will see the end product, the capacities & ofcourse what it will cost finally ;-)

Van ivke2006

Expert (84)

afbeelding van ivke2006

04-10-2010, 15:47

Hi Marlon-B,

Any progress? We did't get any update for more then one week. Wink

Btw. Very nice hardware. Hopefully you will manage to complete your 'project'.

Van marlon-B

Expert (88)

afbeelding van marlon-B

04-10-2010, 22:11

Hi Marlon-B,
Any progress? We did't get any update for more then one week. Wink

That's because I'm busy building/testing/building/testing...

The only progress-update that I have is that I've build a new basemodule-prototype with the ability to reroute the IO-only-slot (slot 4) to the main controller on the basemodule so it can directly communicate with this slot (or rather the cartridge in this slot) instead of the MSX.

I want to test if it's possible to (for example) feed a GFX9000 with data at a faster rate than the MSX can.
So the hardware-blitter can be used for that as well.


Btw. Very nice hardware. Hopefully you will manage to complete your 'project'.

thnx..

and I dont see why I wouldn't complete it. I'm having alot of fun. Just like the OLD days. Wink

Van Bastiaan

Champion (333)

afbeelding van Bastiaan

05-10-2010, 21:02


And I don't think I will get fed up very soon as I don't have to go to work or anything for at least a couple of years(i've been bought out by larger company you see)
So this is the only project I'm working on.Smile

Sounds like a great situation! And any new pics of your cool project?

Any chance of also using all the knowledge you are gathering in some "all-in-one" combination cartridges (using sub slots/only usable in main slots) besides your expander+blocks?

I wouldn't mind 1 or 2 carts with lots of functions: mass storage+dos+ram+flash/mega-RAM+scc-I+msx-music+msx-audio+.+... Smile
And I think I am not the only one! I don't need an expander, I just want lots of expansions at the same time ! Running Naked in a Field of Flowers
It may even be cheaper without all those expensive connectors and big PCB....

Van marlon-B

Expert (88)

afbeelding van marlon-B

05-10-2010, 23:41


Any chance of also using all the knowledge you are gathering in some "all-in-one" combination cartridges (using sub slots/only usable in main slots) besides your expander+blocks?

I wouldn't mind 1 or 2 carts with lots of functions: mass storage+dos+ram+flash/mega-RAM+scc-I+msx-music+msx-audio+.+... Smile
And I think I am not the only one! I don't need an expander, I just want lots of expansions at the same time ! Running Naked in a Field of Flowers
It may even be cheaper without all those expensive connectors and big PCB....

The reason I chose to do my own "add-on" design is that I needed a more efficient basis for future expansions I plan to release.
In my design memory is shared between all attached add-ons. Meaning DMA (direct memory access). For example you can issue a command to read some file from the mass-storage add-on and store it directly in the shared memory. Then you can issue a command to send it straight to a MP3-player or MOD-player or whatever.
Or stream music/video from your PC through USB.
With NO overhead for the MSX.
And adding extra memory is a piece of cake and all add-ons can use it directly.

Also I'm not bound any more to the standard dimensions of cartridge designs. Because these casings are (almost) impossible to get and if you do it is too expensive.
I can use any available and customized casings for my add-ons. For example my SD-reader is enclosed in a case of 65x45x18mm which is very very small.
Also my add-ons use a 50 pins connector with a 1.27mm pitch . Which is more current,smaller and cheaper than the 2.54mm pitch design.
Overall the add-ons will be cheaper to produce than the standard cartridge designs.

So to answer your question: NO, I will not design and build regular expansions with all-in-one features besides my slotexpander. I'm sorry Crying

All my future expansions will be of the new design.
And I will be more then happy to provide other hardware designers with specs to build their own.

What I could do is to design a separate base-module with NO (regular) slotexpander functionality just my own add-on hardware.
This will then be much cheaper and people who already have an expander dont have to buy another expander just to use the add-ons.

Van ivke2006

Expert (84)

afbeelding van ivke2006

06-10-2010, 11:13

I have a bold question too: Smile

Would it be very difficult for you to develop a 'standalone' solution, based on your current development en knowledge you have?
What I mean is basically a new MSX motherboard which is..lets say..at least MSX2 compatible but more speedy (for example Z80 @ 20mhZ, Z180@33Mhz or even a eZ80@50 Mhz :P ) and maybe your hardware-Blitter to enhance the graphical capabilities.

Maybe it's way to much to ask for, but it seems you have a lot of hardware knowledge and that's a rare feature in the current scene (No offence to sunrise and others!).
I very much like al the extensions which are currently developed but in basis you need a 20+ year home-computer to use it. It would be very exciting if there was a possibility to buy a new msx computer mainboard.

So..what do you think? ;)

Van marlon-B

Expert (88)

afbeelding van marlon-B

06-10-2010, 20:37

I have a bold question too: Smile

Would it be very difficult for you to develop a 'standalone' solution, based on your current development en knowledge you have?
What I mean is basically a new MSX motherboard which is..lets say..at least MSX2 compatible but more speedy (for example Z80 @ 20mhZ, Z180@33Mhz or even a eZ80@50 Mhz :P ) and maybe your hardware-Blitter to enhance the graphical capabilities.

Maybe it's way to much to ask for, but it seems you have a lot of hardware knowledge and that's a rare feature in the current scene (No offence to sunrise and others!).
I very much like al the extensions which are currently developed but in basis you need a 20+ year home-computer to use it. It would be very exciting if there was a possibility to buy a new msx computer mainboard.

So..what do you think? ;)

No it would not be that difficult to do this. The best and fastest way to do this would be to use a powerfull FPGA (like the 1chipmsx) and some supporting hardware.
But then I'm only talking about the hardware. You'll also need a new BIOS(or several) written from scratch to support all the new hardware and still keep the MSX feel(Basic,msxdos,roms).
And this is something that usually takes a long long time to get right. Especially for just one person.

I think if you really want to do this right it should be a team effort or rather a community effort.

starting with a wishlist.
then choosing the most optimal but still doable (not making it too expensive) specs from this list.
then divide the work between several teams. keeping very tight communication or even come together regularly.

And keep in mind I'm talking about a very long term project here (at least two years).

Van marlon-B

Expert (88)

afbeelding van marlon-B

19-10-2010, 11:54

SlotXpander progress update:

Hardware blitter works perfectly ! The blitter (now) runs at 66mhz (but I might dial it down because it's a bit overkill)
The blitter can also act like a coprocessor with 16 (memmapped) registers and handle complex calculations.
An ordinary MSX 1/2 can now OUT-RUN a Turbo-R Wink

I've also made a big design change in the base-module.
I'm now using a CPLD (after reading RetroTechie's CPLD activities I became infected,how R U progressing BTW?)
to handle all complex logic and memory management.
Board is ofcourse now much smaller. And base-module and all add-ons now run on 3.3V.
Meaning less power usage.

I'm now busy writing some software to demonstrate some of the capabilities.
And ofcourse test the HELL out of it !!
Hoping to show you some (home) movies soon...

until then....Cool

Van tcruise

Master (133)

afbeelding van tcruise

23-10-2010, 02:16

Awesome! Can't wait to see the results, and then perhaps get one of the 1st units to test! Keep up the good work.

Van marlon-B

Expert (88)

afbeelding van marlon-B

23-10-2010, 19:58

tiny *UPDATE*:

as I'm using a CPLD now as memory-access-controller it dawned to me that I could easily implement hardware debugging !
So I did a simple experiment. I added a switch and whenever this switch is activated debug mode is active.
and the memory controller returns a $18 $FE for every instruction fetch. And it worked. the msx paused.
By pressing another button it can single step by comparing the last address read for intruction fetch.
Ofcourse this only works if entire program including bios (0-4000) section runs in the ram of the base-module.

And then I even did another experiment with instruction-code replacement. I replaced LDIR-instruction with my blitter-copy this is how:

whenever the mem controller encounters:
ld hl,xxxx -> $21 xx xx it saves xx xx
ld de,yyyy -> $11 yy yy it saves yy yy
ld bc,zzzz -> $01 zz zz it saves zz zz
LDIR -> $ED B0 -> it returns nop nop inserts wait and then activates a blitter copy with the saved values.
the blitter can copy 10 bytes per microsecond in contrary to LDIR which does 1 byte->21Tstates=5.8 microseconds and 10bytes=57 microseconds

so the posibilities of instruction replacement are endless.

These were only experiments so I'll try and find a cleaner way to implement it all. Suggestions are ofcourse welcome.

I just had to share it with you ...Cool

Pagina 15/16
8 | 9 | 10 | 11 | 12 | 13 | 14 | | 16