| |
Paladin
Registered: Oct 2020
Posts: 43 |
VIC-II Kawari
Randy dropped some of these on eBay in the wee hours (1AM Eastern Time) of July 29th. Not long after that I got COVID pretty bad so have just now got the chance to slap the 'Mini' version of his chip in an (250407) NTSC machine. I had to grab the demo disks and the config disk from Randy's Github page https://github.com/randyrossi/vicii-kawari and dropped them on my Ultimate-II+ ... After setting the jumpers on the Kawari to use the on-board oscillator for PAL and the motherboard oscillator for NTSC I fired the C64 up and loaded the config utility and set the VIC to a 6569R3 and saved the config. Anytime you switch the chip you need to cold boot the machine for the changes to take effect.
I spent the afternoon going through some PAL demos and so far I haven't run into any issues and am extremely happy.
Randy ordered 200 more of the Kawari-Mini boards so be on the lookout for those. I believe he is also looking to work with someone over in Europe to help get these things out there on that side of the world. As far as I know, no POV (Plain Old VIC) versions were sold but these chips should be more affordable though they will lack the on-board oscillators for NTSC/PAL switching as well as the new extensions.
Did anybody else out there grab this or the 'Large" model with the HDMI out? Does anybody out there have any demos etc that they would like tested or software that "might" give the Kawari issues?
Cheers and let me know if you have any questions or suggestions |
|
... 1 post hidden. Click here to view all posts....
|
| |
Paladin
Registered: Oct 2020
Posts: 43 |
Thanks for the reply machete. Interesting find, I'll give that one a test and see how it goes as it might be worth opening an issue on Github.
My testbed is an NTSC 250407, all original chips except the VIC-II of course. The Kawari is set to 6569R3 using its built in oscillator.
Randy has also opened up a Discord channel and everybody is welcome to join. He's answering questions and has a poll setup at the moment about the 'Large' HDMI version and whether people think he should drop that version due to some issues.
https://discord.gg/4TUd5mWP |
| |
Paladin
Registered: Oct 2020
Posts: 43 |
Latest status update
VIC-II Kawari Updates - Sep 6, 2022
Stock Update
I ordered 200 mini boards and they should be available next month. I'm trying to make the distribution more fair to different regions this time. I will make 50 boards available each week for 4 weeks and send the notification email at different times. This is also for me to have a reasonable number of packages to prepare/ship in one week. The next batch will be sold through an online store front. If sales stays strong, I will try to keep them stocked until demand starts to drop off.
Discord Server
If you have a device (or not) and have questions or want to chat about the project, feel free to join my VIC-II Kawari discord server.
https://discord.gg/4TUd5mWP
Large Board Update
FPGAs for the large board design are available again. However, I discovered many HDMI displays will prevent the Kawari from booting while they are powered on. The FPGA core can be partially powered by the display device pullups on the TMDS lines. Some monitors don't 'let go' even after no video signal is detected. Others will after a short timeout.
Normally, devices have HDMI buffer ICs that prevent this but that would add more cost/space and possibly open up licensing issues. I might order large boards without the components that are necessary to output the DVI signal as a 'base' unit and have the HDMI connector as an add-on option (with a clear explanation of the issue and how to work around it). However, I have no good solution to the boot issue other than to turn off the display or use an HDMI switch. Given the issue, I'm wondering if I should just abandon the DVI option.
Is DVI worth it? Should I abandon the DVI option? Or just manufacture devices with the limitations? (Feel free to join discord and express your opinion in the #dvi-option-poll channel)
MK2 Reloaded Compatibility
A local C64 user kindly lent me a MK2 Reloaded board and I was able to get both the large and mini Kawari boards to work. It turns out the MK2 Reloaded boards ignore CAS/RAS timing since it uses static ram. Instead of emulating DRAM, it expects address lines to be valid at specific times. An alternative firmware will be available for MK2 boards. The alternative firmware will still work on real C64 boards but it can cause glitches so firmware will be separate depending on use case. Both firmwares will run on either board so you can switch from one to the other. Unfortunately, I have not been able to make one firmware that will completely satisfy both real and MK2 boards at the same time. If you have an MK2 board and want to test the firmware, contact me.
SaRuMan Compatibility
A SaRuMan + 250407 (an odd combination) did not work with v1.5 of the firmware. However, I have an alternate firmware for this that appears to address the ram glitches. This is likely the same issue Jan Beta ran into with his SaRuMan + 240466 combo. I'm waiting to hear back from him on whether this fixes his issue as well. This will be available as a firmware update soon. If you have this issue and want to try the fixed firmware, contact me.
Large Board Analog RGB NTSC Color Fix
Version 1.5 (or lower) of the firmware had incorrect NTSC timing parameters that was causing a dark screen or low/no blue color over analog RGB (s-video/composite was not affected). V1.8 fixes this and is available for download/flash from the github site under disks/flash dir. If you had plans on building a RGB connector, you may want to update your firmware.
C64-C Short Boards
As stated in the eBay listing and docs, I'm not recommending the Kawari to replace 8562/8565's in C64-C short boards. Even though it (mostly) functions, closing the machine and timing glitches are a problem. It may be possible the board can receive a dedicated 8562/8565 firmware in the future but I have no plans on doing so. This project remains a 6567/6569 replacement project for now. |
| |
tlr
Registered: Sep 2003
Posts: 1717 |
Quoting Paladin
Large Board Update
FPGAs for the large board design are available again. However, I discovered many HDMI displays will prevent the Kawari from booting while they are powered on. The FPGA core can be partially powered by the display device pullups on the TMDS lines. Some monitors don't 'let go' even after no video signal is detected. Others will after a short timeout.
Normally, devices have HDMI buffer ICs that prevent this but that would add more cost/space and possibly open up licensing issues. I might order large boards without the components that are necessary to output the DVI signal as a 'base' unit and have the HDMI connector as an add-on option (with a clear explanation of the issue and how to work around it). However, I have no good solution to the boot issue other than to turn off the display or use an HDMI switch. Given the issue, I'm wondering if I should just abandon the DVI option.
TMDS is supposedly CML terminated with 50ohm to 3v3.
How are you driving these, LVDS? If so, maybe you could try AC-coupling your outputs to get rid of your issue?
Haven’t checked the TMDS line coding. If there is too much DC it won’t work, but it’s worth a try IMO. |
| |
Paladin
Registered: Oct 2020
Posts: 43 |
@tlr
I passed your question and possible fix on to the developer of the project on his Discord. I'll let you know what he says. |
| |
Paladin
Registered: Oct 2020
Posts: 43 |
@tlr
Here is the response from Randy
"On the Spartan6 board, it is TMDS directly from the fpga to the connector. But that design will never get an update because the Spartan6's are unavailable. On my new design using the Trion, it is LVDS from the fpga to a LVDS to CML shifter. Despite there being a buffer between the fpga and the connector, the back power issue is still present. I tried AC coupling LVDS directly to the connector (no shifters) in my first attempt to use LVDS but I could not go above 25Mhz pixel clock with that solution ( I need ~32mhz). I can try it I guess." |
| |
machete
Registered: Aug 2005
Posts: 6 |
Hey Paladin, I will post my findings on Discord with screenshots. Did you discover other issues? |
| |
tlr
Registered: Sep 2003
Posts: 1717 |
Quoting Paladin@tlr
Here is the response from Randy
"On the Spartan6 board, it is TMDS directly from the fpga to the connector. But that design will never get an update because the Spartan6's are unavailable. On my new design using the Trion, it is LVDS from the fpga to a LVDS to CML shifter. Despite there being a buffer between the fpga and the connector, the back power issue is still present. I tried AC coupling LVDS directly to the connector (no shifters) in my first attempt to use LVDS but I could not go above 25Mhz pixel clock with that solution ( I need ~32mhz). I can try it I guess."
Missed the response here. Interesting, so he tested that with the current board but didn't get the bandwidth. Then with a new design with a CML buffer it still back powers, strange!
Is the back powering really happening through the CML buffer? What buffer is this? |
| |
Paladin
Registered: Oct 2020
Posts: 43 |
Quote: Quoting Paladin@tlr
Here is the response from Randy
"On the Spartan6 board, it is TMDS directly from the fpga to the connector. But that design will never get an update because the Spartan6's are unavailable. On my new design using the Trion, it is LVDS from the fpga to a LVDS to CML shifter. Despite there being a buffer between the fpga and the connector, the back power issue is still present. I tried AC coupling LVDS directly to the connector (no shifters) in my first attempt to use LVDS but I could not go above 25Mhz pixel clock with that solution ( I need ~32mhz). I can try it I guess."
Missed the response here. Interesting, so he tested that with the current board but didn't get the bandwidth. Then with a new design with a CML buffer it still back powers, strange!
Is the back powering really happening through the CML buffer? What buffer is this?
Randy responded back...
"rmr — Today at 9:07 AM
Yes, I was surprised too. The part is SN65CML100D." |
| |
Paladin
Registered: Oct 2020
Posts: 43 |
Quote: Hey Paladin, I will post my findings on Discord with screenshots. Did you discover other issues?
I tested the "Higher Level" by Lower Level demo and I'm not seeing those graphics issues like in the pictures you posted in the Discord channel. Pretty strange... So far I haven't noticed any oddities on my side at all.
I see you put the chip in a 250425. What does the rest of that setup look like? All original chips on the 250425 except the VIC-II?
The only thing I think I haven't mentioned yet for my setup is that my video goes out the C64 8 pin DIN to an S-Video plug and into a Retrotink-5x and then out the Retrotink-5x via HDMI to an LCD TV. |
| |
machete
Registered: Aug 2005
Posts: 6 |
I have all original chips on the 250425 except the VIC-II.
My output is 8 pin DIN directly into S-Video of an old LCD TV (Sharp Aquos 15"). |
| Previous - 1 | 2 - Next |
