Mk2 Compatibility with Kawari

Caution: Non registered users only see threads and messages in the currently selected language, which is determined by their browser. Please create an account and log in to see all content by default. This is a limitation of the forum software.


Also users that are not logged in can not create new threads. This is a, unfortunately needed, counter measure against spam. Please create an account and log in to start new threads.

Don't Panic. Please wash hands.
    Code
    1. I am the happy owner of a Mk2, and I have no problem to report here :)
    2. At the same time, I am interested in another very attractive project, a new version of the VIC-II, named "Kawari".
    4. I don't know if I'm allowed to post a link about this product, so I won't do it, but google can easily help you to find more about this.
    6. The ship seems compatible, according to the site's data, with almost all C64 motherboard models.
    7. However, I'm wondering about its possible compatibility with the MK2, before making a purchase, and frying one or both of the components :D
    9. Do you have some infos about that ? I "new" motherboard need a "new" Vic-II - That would be great !

    Hello Erestor,


    If you google it its stated that 250407, 326298, 250425 & KU-14194HB* (all “breadbin” and some C64C revisions) are working.

    But if you google on "kawari with MKII" Then you find it can be done and seperated firmware will be available.

    The only problem i see there is that the auto detection of the VICII will not work correctly, which may or may not cause problems.


    I don't know of any customers that have a "Kawari" though, so can't tell. Generally we can only guarantee (and recommend) original MOS ICs to work for that matter (but that doesn't mean others will not work - we just have no way to test all those replacements)

  • Tobias

    Changed the title of the thread from “Mk2 Compatibility” to “Mk2 Compatibility with Kawari”.

    It's safe to try, though. The MK2 will revert to a safe 5V for the VIC chip for unknown values, but I believe the design doesn't even use the 12V line of the "old" VIC chips, so there's no danger of applying over-voltage.


    Jens

    It works - at least mostly. Have power been off for a long time it sometimes fail to start but then a quick power cycle will do the trick.
    Haven't played around with it too much so I'm unsure whether it's the J-CIA's fault and/or the Kawari as I only tried it for fun - Kawari and J-CIA is used in another machine (where they work perfectly together). Something is clearly not always booting as fast as the MK2 expects components to be ready. If you want I can try with regular CIA to see if its the Kawari that is the issue.
    Other than that there's not too much room in a C-case. At least with the Large there's no room for a 2nd SID unless stacking sockets - but then there's positively no room in a C-case (breadbin perhaps). There's something with the video as well. There's more noise than in a regular machine using copperdragons RF modulator replacement but it's sure better and sharper than a real VIC anyway. Haven't tried with the S-video fix though (can if you want me to) to see if that makes any difference. Otherwise image looks fine - at least when using RetroScaler 2x as converter - granted it is quite forgiving and even way off signals it'll convert quite nice.

    Can't remember what MK2 detects VIC chip as but not too important as the Kawari only connects to the 5v supply. At least it doesn't seem to cause diode to blink error (very annoying). Firmware on Kawari must be 1.14 or higher in order to work as that is when S-RAM was supported. If your Kawari is not on a new enough version you need to upgrade it in a machine that's using DRAM (regular ram, not those modern ram replacement modules).


    But yeah. There's quite a few minor tweaks I'd like to the firmware. Now with all those replica ICs it would be nice to be able to slightly increase time before it starts to be able for replica ICs to fully program their FPGA's. Another thing on my wish list is using the restore key to be able to switch between dual mono (both chips playing the same - D400 address or just even regular mono) and stereo. You can set it up do a various things but the most obvious to normally use mono but switch to stereo on demand seems not possible. So so far the best compromise is an external physical switch to switch between mono an stereo as most content would otherwise be left channel only.

    Quote from Kaninbjerget

    Something is clearly not always booting as fast as the MK2 expects components to be ready. If you want I can try with regular CIA to see if its the Kawari that is the issue.

    If it's just a matter of startup-time, we could change the MCU code to give the system more of that with a firmware update that can be applied through the USB port of the MK2 board. We may be able to add audio options then as well, but I don't want to promise anything right now. We're very short on labour force right now, with half of my employees on sick leave.


    Quote from Kaninbjerget

    If your Kawari is not on a new enough version you need to upgrade it in a machine that's using DRAM (regular ram, not those modern ram replacement modules).

    Would be interesting to know if this works on a C64 Reloaded MK1. While it uses D-Ram, it has some address-storage in place to fix a VIC bug.


    Jens

    Quote from Jens

    If it's just a matter of startup-time, we could change the MCU code to give the system more of that with a firmware update that can be applied through the USB port of the MK2 board. We may be able to add audio options then as well, but I don't want to promise anything right now. We're very short on labour force right now, with half of my employees on sick leave.

    That'd be really nice :love: All the best wishes to your employees. Have been down a few days myself as well. Can be I'll experiment a little with it this evening to see what patterns of behaviour I can find.

    Quote from Jens

    Would be interesting to know if this works on a C64 Reloaded MK1. While it uses D-Ram, it has some address-storage in place to fix a VIC bug.

    Since I don't have an MK1 I can't tell. I did try with firmware 1.13 (I believe it was) when I got it in MK2 and it didn't work - garbled video. If I remember correctly it's not really the type of ram that is important but merely that modern ram is noticeably faster. But unless getting a used Kawari all should come with 1.14 or newer so shouldn't be an issue. I haven't tried updating the Kawari in MK2 though - but the last time I updated was with ram replacement module.

    Did a little experimenting. I remembered a few things wrong.

    Kawari (Large) - almost - worked on first try. Deres was an odity of random characters on screen flicking on the screen. Not much lets say 5 very short flashes of a random character anywhere on the text area in basic. A regular reset (using power button flip up) and it went away. My guess it that it's not starting quite fast enough for the MK2. So the issue of just black screen must be the J-CIA which my guess is they're even a little slower - those I've not played with this round.

    The error diode blinks - unfortunately - that I remembered wrong. I wonder what pins to brigdge with a resistor in order for the MK2 to think there's a 6500 series in there... Pal it gets right. Here's the debug info:
    VIC-II: Unknown (3A 3A 0000 PAL)
    For the SID it's pin 25 and 26. Some suggest 1k make MK2 detect ARMSID, SwinSID, BackSID and others as an 8580 to avoid the annoying blinking. Should I want to use a SID replacement I'd sure play with the value - guess as high as possible the better. However for the video there's at least one thing more. Since it defaults to a 85xx VIC the video is on the bright side (Kawari puts out 65xx spec video).
    If some can suggest which pins to resistor on to make it think it detected 65xx VIC I can try it.

    For the slightly noisy video - that is only when using the large DIN socket and not using the S-video fix. Using S-video fix or the mini DIN S-video output the video is very nice. Perhaps a little more jailbaring when looking closely than my real machine (with copperdragons RF replacement) but nothing that should bother anybody and again still way better than a real VIC. With my real C64's I use a DIN adapter box for both audio and video so naturally I just plugged that in the MK in the first place but as the difference between S-video fix bypass/enable is to the subtle side on the real machines I just left jumper in it (no fix) - but on MK2 the difference is much more apparent.

    So if anything other than boot time to better suit Kawari I don't know what would be a good approach. I understand that many are not familiar with electronics and they'll quickly ruin genuine chips - either by accident or downright on purpose (but hey I've had two 85xx that detected wrong as 65xx and they've run for many hours without failing though they got very hot - how are they doing now?). But at least if a resistor fix for identification could be found that'd be very nice and at least for those savvy it's an acceptable solution - no big deal soldering a resistor on the underside of Kawari.

    Edit: Played a little with Kung Fu Flash as well. When using it and Kawari and it has been powered off a a long time (few minuttes) then video is garbled and neither a regular nor the unstopable reset (KFF uses the unstopable) resolves it - but a quick power cycle does. So still my hypothesis - it's not quite starting fast enough for MK2 when all capacitors (both MK2 and Kawari) is down to zero volt.

    Quote from Kaninbjerget

    Pal it gets right. Here's the debug info:
    VIC-II: Unknown (3A 3A 0000 PAL)

    Does it also get NTSC right? Can you configure it to be NTSC and re-try?


    Quote from Kaninbjerget

    (but hey I've had two 85xx that detected wrong as 65xx

    Did this happen on one of our boards, and do we have a support case for that?


    Quote from Kaninbjerget

    When using it and Kawari and it has been powered off a a long time (few minuttes) then video is garbled and neither a regular nor the unstopable reset (KFF uses the unstopable) resolves it

    The VIC chip does not get a reset signal, so no matter what reset you're sending, it's not going to see it.


    Jens

    Quote from Jens

    Does it also get NTSC right? Can you configure it to be NTSC and re-try?

    I can try it in the evening. Never bothered with NTSC myself though.

    Quote from Jens

    Did this happen on one of our boards, and do we have a support case for that?

    Yes. I long going case which started with this thread back in April
    Stereo request feature + VIC side question

    Sent you an email monday evening.

    Quote from Jens

    The VIC chip does not get a reset signal, so no matter what reset you're sending, it's not going to see it.

    That you're probably right about :)

    Quote from Jens

    OK, that's not a wrong SID recognition, but the still-not-investigated-VIC detection going wrong. I thought I had missed something.

    We all mix up things from time to time :)
    It was just to examplify that for the SID you can trick it to think a non real SID like ARMSID, BackSID, SwinSID etc. is an genuine 8580 by bridging a 1k resistor on pin 25 and 26 thereby not being annoyed by a blinking LED and in chipset info it'll say 8580 detected. However I've not played with it myself as I'm using real SID and if I decide to I'd try 10k or even higher values first.
    It was a clumsy way to ask if there's a way to trick it into thinking it got a 6569 VIC. Without a schematic it's a bit hard to guess which pins to trick values for and then - what value range should be within.


    I tried switching Kawari around to use motherboard oscillator for NTSC and onboard for PAL. It resulted in garbled video. However going into factory menu and forcing NTSC timing a perfect NTSC picture appeared.

    My suggestions for supporting Kawari. Let autodetection decide what voltage to supply to VIC (doesn't matter for Kawari as it's only using the always 5v - pin 13 if I remember correctly. Then add option to force NTSC or PAL and force which video level (85xx vs 65xx). That way at least with correctly identified VIC nothing is destroyed (when it's detected right!) but of course the user risk risk garbled video if forcing a setting that is not appropriate.Using anything else than auto should disable warning for incorrect VIC chip (it's OK it displays unknown - just please no blinking LED).

    While at it I think many people would appreciate a setting to ignore SID errors as well - again perhaps force a setting if there's a difference in the audio path (there is a minor difference in a capacitor on real C64s). Again let auto detection decide voltage but merely just don't blink the diode for not detecting a correct SID. As time goes by many people opt for SID replacements and there's a quite a few of them. FPGASID, uSID64 and SidKick64 to add 3 which I know in addition to those already mentioned - but the first 3 are by far the most popular.

    Quote from Kaninbjerget

    It was just to examplify that for the SID you can trick it to think a non real SID like ARMSID, BackSID, SwinSID etc. is an genuine 8580 by bridging a 1k resistor on pin 25 and 26 thereby not being annoyed by a blinking LED and in chipset info it'll say 8580 detected.

    That suggestion originally came from me, somewhere in this forum. However, that's not an option on the VIC, as the detection doesn't work with a normally-unused pin, but with the black level output of the non-initialized VIC chip. So if the Kawari developers want to approach us from the other side, that "startup value" could be adapted.


    For the LED, I could think of a setting "disable error-blinking". I honestly don't remember how much space there is left in the MCU flash to add things, but we'll look into it. Peter's next visit is not too far away, and we should put that on higher prio next time.


    Jens