Jump to content

BruceMcF

Members
  • Content Count

    312
  • Joined

  • Last visited

  • Days Won

    8

BruceMcF last won the day on September 12 2020

BruceMcF had the most liked content!

Community Reputation

182 Excellent

2 Followers

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. It IS an ordinal scale. 2/10 is saying that is there is only one step between this board and not meeting the criteria in any way whatsoever, and seven steps between this board and meeting the criteria perfectly. I don't have any experience buying from these channels, but since the source will be Chinese, they are indeed likely to prescreen the chips to sort between those they sell to their ongoing customers and those they sell to hucksters selling to people gambling on components on eBay. If someone with experience buying from the first channel reports that they have good experience, I don't have any reason to assume that that will change for these orders.
  2. They don't meet it perfectly, but then again, Vera FPGA and YM2151 grey market, 65C02 stock, VIAs stock, glue logic stock, RAM stock, FlashROM stock ... ... 2/10 is a silly score to give on that basis. That's at least a C, so set it at 7/10.
  3. I have heard there are no plans for keyboards other than the US, I would assume for economies of scale reasons. However, it's a user definable character set with up to 1024 characters, so ISO-8859-1 (Latin1) would be straightforward to support. The legacy PETSCII character set isn't so much "US" as a "Norristown, Pennsylvania, US" character set, but it's not strictly limited to PETSCII.
  4. That's the stage 3 that might include CPU and VIA cores (which are available under license from WDC) and all of the glue logic together with Vera in a single beefier FPGA, but the stage 2 is surface mount with cost reductions, which could include a CPLD/FPGA for the glue logic or a beefier FPGA to include it with Vera. An FPGA might handle the YM2151 in either stage3 or both stage2 and 3, with the real stock for the through pin reference design.
  5. Quite right about the Basic/KERNAL ROMS not being cleanly divided between Basic and KERNAL code. Note that much of the graphic commands seem to be calls into the GEOS Kernel, but there's still a certain amount of processing each Basic Keyword has to do to parse it's arguments.
  6. Certainly. You could do an SPI interface and plug in a RPi0 supported by a Kernel vector handler on device 9. Make ATN into /SEL, DATA into MOSI and /SRQIN into MISO. As long as you refrain from using tho other IEC device numbers, you should be fine.
  7. The word on having bit banged serial was when the UART on the Vera was dropped because they needed the pins for more register addressing. If the board is now booting up, hopefully they can start work on that so we'll know what speed we are talking about. But for testing device drivers and such in hardware, 9600bps 1.2KB/s, so 16KB in around 13 seconds, and it might be 19200bps, so 16K in under 7 seconds. The Centronics parallel is just the pinout given of the User Port on the second prototype board ... I wouldn't automatically expect Kernel code to support that, but the VIA pins connected would mostly support an EPP parallel port. But at 8MHz, even if it was 50 clocks per byte, that's 160KB/s..
  8. Certainly. Horses for courses. Someone with a UART on an expansion card may have the User Port free. Someone with a card with 2 VIAs on it can have a full parallel port, a two serial shift register serial port that should conservatively support 38.4kbps, and still have the user port free plus something like 16 free GPIO. Someone connecting to an EPP device on the User Port without any expansion cards might be looking for the fastest & cheapest way to use the IEC port for the job. One thing I have zero interest in trying to recreate is my daisywheel printer that sounded like a submachine gun with a silencer, but the EPP box on the User Port, I do have some interest in that. So flexibility in options is appealing.
  9. I was taking that into account. The whole 8bit retro project building market is the size of market I was talking about. Where economies of scale kick in is at the assembled board level. You have VGA or composite video, simple sound, and mass storage in one board added with a block pin headed on your main board. Plus a pile of example 6502 code and, eventually there will be some available Z80 code as well.
  10. Well, I am guessing that if the C64 could bit bang serial to 2400baud, the CX16 will certainly be able to hit 9600baud, maybe 19200baud. There are lots of things wher 19200 is plenty fast enough, but I had a home internet connection on a second landline in the late 90s, and there are that situations where 56kbps is annoyingly slow. Backing up the SD card is one of those, and if I have to swap out the SD card to back up while developing, I'm going to end up doing fewer backups than I should. Actually, for back up, even a standard Centronics parallel port would be fine, as the data is going one way, so some input status lines to handshake would be enough. That eliminates worry about the USB to Parallel port cable drivers working correctly in EPP mode. As far as policing the tone of posts, I reviewed the thread and don't see the pushy post. Maybe you could quote the post you are reading as pushy. All I see is people talking through potential optional add ons for a project they are interested in. Indeed, I don't know who there is TO push, given that nobody is saying the project team should be doing any of this. Even the posts that were not aware that a bit banged serial port is planned were not clamoring for the project team to add X or Y chip yo fill the gap.
  11. Yes, running the User port lines out to a breadboard to talk to microcontrollers or microcontroller petipherals. Bringing up a tethered Forth in a small microcontroller is one example. And before people jump in and say, "no, the way you do it is this way with these software tools", I'm going to point back to where I said, " not everyone will want to" ... a lot of people are happy with the tottering piles of tools on development environments that run on dlls from other development environments talking to proprietary drivers, but those who are not thrilled by that approach exist. Indeed, it's almost guaranteed that if there is a mainstream approach, there will be pockets of people who are not enamored with the mainstream.
  12. I don't think that there is an implication that EVERYONE need to or want to transfer files in and/or out of their CX16 at speed faster than the built-in serial port allows, but coming up with use cases for SOME people to want to do so is not hard. (1) Because for many people, the easiest backup for the SD-card will be to write it out to some folder in a PC's mass storage (2) It doesn't REQUIRE cross development, but many people will do cross development. (3) Some people have things they'd like to do that exceed their budget for bespoke hardware and see a way for an existing PC to be used as a resource to fill the gap. (4) The amount of GPIO and freedom to program to the bare metal makes the CX16 attractive as a bench computer for some kind of work, and you'd like your bench computer to be able to talk to your laptop.
  13. Also using an FPGA because that is the current technology for market niches of this size.
  14. If you want a more convenient form factor for sneakernetting, there's always compact flash, a 2GB CF card is around $12 on Amazon, and there are USB CF readers available. I'd assume that a Device #9 file server over the serial port will already be sorted out by launch, so the issue is if that is not fast enough. An EPP Parallel port interface to a USB / DB25 converter with full LPT and COM functions would be plenty fast enough if bit banged RS-232C serial is not, and if the other capabilities available from higher speed serial are desired, some form of dedicated UART would also fill the bill. At the cost of a USB/DB25 converter, or the cost of a UART card, an accelerated-SD2IEC solution would be more for those who already have or want an SD2IEC for a VIC-20 or C64 in their setup, so the SD2IEC is not really an extra overhead expense.
  15. If it's to communicate with a PC, I don't understand the "SD2" part ... a USB to parallel converter plus a parallel to IEC cable would do. That would come out to a long way less than $70. And if it's to sneaker net to the PC, then a second SD card slot on an expansion card could be done that was substantially cheaper than an SD2IEC.
×
×
  • Create New...

Important Information

Please review our Terms of Use