(new title): Hard decision, but the answer is pretty clear.

Chat about anything CX16 related that doesn't fit elsewhere
Scott Robison
Posts: 952
Joined: Fri Mar 19, 2021 9:06 pm

(new title): Hard decision, but the answer is pretty clear.

Post by Scott Robison »


When David reviewed the Mini-PET (I think it was that) he said something to the effect of if it had been out before the dream computer video, he might not have done the dream computer video, instead just looking for ways to expand Mini PET to have video and sound. Thus the X8 does check a lot of the boxes.

SolidState
Posts: 13
Joined: Sun Aug 22, 2021 11:53 pm

(new title): Hard decision, but the answer is pretty clear.

Post by SolidState »


I'm not sure what all the trepidation is all about. Frank released the entire design files for the X8 under a 2-Clause BSD License at the end of last year. The Cloanto IP makes it a bit problematic to build and ship as a finished product though. It could be done as a "dev kit" and you get to flash the firmware. The board comes with an integrated Cortex-M0 host processor, so it's already setup for this.


x8.jpg
Wavicle
Posts: 284
Joined: Sun Feb 21, 2021 2:40 am

(new title): Hard decision, but the answer is pretty clear.

Post by Wavicle »



4 minutes ago, SolidState said:




I'm not sure what all the trepidation is all about. Frank released the entire design files for the X8 under a 2-Clause BSD License at the end of last year. The Cloanto IP makes it a bit problematic to build and ship as a finished product though. It could be done as a "dev kit" and you get to flash the firmware. The board comes with an integrated Cortex-M0 host processor, so it's already setup for this.



x8.jpg



Do you have a link for the design files? A web search and reverse image search are coming up empty.

x16tial
Posts: 177
Joined: Sun Feb 07, 2021 8:23 pm

(new title): Hard decision, but the answer is pretty clear.

Post by x16tial »


Wifi too...  smh

SolidState
Posts: 13
Joined: Sun Aug 22, 2021 11:53 pm

(new title): Hard decision, but the answer is pretty clear.

Post by SolidState »



37 minutes ago, Wavicle said:




Do you have a link for the design files? A web search and reverse image search are coming up empty.



There was a discussion about the X8 files on this thread. It looks like the repos were made private, but I think there are some forks out there (the image was a render I did just now from KiCAD). 

SolidState
Posts: 13
Joined: Sun Aug 22, 2021 11:53 pm

(new title): Hard decision, but the answer is pretty clear.

Post by SolidState »



33 minutes ago, x16tial said:




Wifi too...  smh



The board has a Wifi UART. It provides a basic AP that you connect to and then telnet to the board to establish a serial connection. You could also do TFTP to transfer files to/from the memory.

BruceMcF
Posts: 1336
Joined: Fri Jul 03, 2020 4:27 am

(new title): Hard decision, but the answer is pretty clear.

Post by BruceMcF »



7 hours ago, x16tial said:





  • X16 runs at 8Mhz -- 12 Mhz is better


  • X16 has a TON more RAM -- this doesn't matter with fast SD speeds, that same 8k can be swapped from SDCard rather than switching banks


  • X16 has twice the video RAM -- this might be the worst part, it means only 1bpp for 640x480, but tricks with sprites, charsets, etc would probably compensate


  • X16 has 4 expansion slots - provide an i/o bus on the X8?


  • X16 has a Yamaha sound chip - add on expansion with said i/o bus?


  • X16 has an IEC disk drive port - this is really kinda silly.  nice nostalgia, but let's be real, who is going to use this?


  • X16 is infinitely more "hackable" - make the X8 more hackable, provide that i/o bus.  and how about tools to make reprogramming the fpga easy?


  • X16 has SNES ports - not needed with the usb port


  • X16 has PS/2 - usb is better




As attached as we all are to the X16, the X8 is just the better idea.  Even the name X16 is a holdover from when the 65C816 was going to be its processor, I believe.



In my opinion we should all bite the bullet, and go with the X8 as the platform.

Releasing both I think is a mistake.  @The 8-Bit Guy, you do this full time, so developing for multiple platforms seems like no big deal.  Most of the rest of us don't have that luxury.

 



If there was an adult in the room, she would tell you that you are assuming that is is possible to get something for nothing.

Remember, the X8 is the "proof of concept" answer to the question, how MUCH of the CX16 design could you fit into the same FPGA that the CX16 uses for its Vera Video/Audio/SPI chip.

The answer was useful for giving an idea whether the CX16e could work with only one SRAM chip ... and also to give an idea how much more powerful of an FPGA would be needed for the CX16e.

But above, what you are talking about IS the X16e, not the LX8, and there is a reason David estimates the CX16e would cost twice as much to bring to market as the LX8.

The LX8 is a proof of concept of putting as much as possible into an FPGA-only system, and only using the internal RAM works for the LX8 precisely because it allows it to avoid bringing out the system bus.

So "provide an io bus to the X8" means, "build a CX16e rather than an LX8"

So "add on expansion with said I/O bus" means "build a CX16e rather than an LX8"

So "make the X8 more hackable, provide that I/O bus" means "build a CX16e rather than an LX8"

We know that the Vera FPGA is pin constrained, because lack of pins was cited as the reason for dropping the serial port function when the register addressing was increased from 8 addressable registers to 32 addressable registers ... which is, after all, exactly two more pins.

So there is no way they can add USB and add 16-18 pins for the I/O bus (16 if an external decoder is used for the five slot selects, 18 is each slot select gets its own pin).

The two reasons a CX16e would be about twice the cost of an LX8 are that it would involve a surface mount RAM, because you can't GET that much RAM as an embedded FPGA module, and that it would involve a more powerful FPGA, with more logic cores and more pins, because the FPGA they are using for the Vera just doesn't have either the logic cores or the pins to handle full simulation of the CX16p ... and that more powerful FPGA would cost more money.

So the LX8 as you are imagining it is not the same LX8 that is basically available to bring to market now, unless they decide to do a few tweaks to it.

Snickers11001001
Posts: 140
Joined: Wed Jan 20, 2021 6:43 pm

(new title): Hard decision, but the answer is pretty clear.

Post by Snickers11001001 »



2 hours ago, SolidState said:




There was a discussion about the X8 files on this thread. It looks like the repos were made private, but I think there are some forks out there (the image was a render I did just now from KiCAD). 



Thanks new guy! 

Wavicle
Posts: 284
Joined: Sun Feb 21, 2021 2:40 am

(new title): Hard decision, but the answer is pretty clear.

Post by Wavicle »



3 hours ago, SolidState said:




There was a discussion about the X8 files on this thread. It looks like the repos were made private, but I think there are some forks out there (the image was a render I did just now from KiCAD). 



Ah, okay. Those were only the board design files as far as I know. I was hoping the RTL for VERA had been published since it is the only bespoke "ASIC" in the X16.

BruceMcF
Posts: 1336
Joined: Fri Jul 03, 2020 4:27 am

(new title): Hard decision, but the answer is pretty clear.

Post by BruceMcF »



4 hours ago, SolidState said:




I'm not sure what all the trepidation is all about. Frank released the entire design files for the X8 under a 2-Clause BSD License at the end of last year. The Cloanto IP makes it a bit problematic to build and ship as a finished product though. It could be done as a "dev kit" and you get to flash the firmware. The board comes with an integrated Cortex-M0 host processor, so it's already setup for this.



x8.jpg



Yes, the Cloanto IP is what makes it something the CX16 team could "just do now".

Post Reply