---
Update---
According to Jens there are less than 20 of these faulty carts out there, so please read this post carefully if you believe yours is one of them.
Quick facts:
* This particular bug affects the clockport on MMC64 so that a RRnet adaptor connected directly to the MMC64 wont work.
* This bug is only present on less than 20 of the very first red V1.1 carts produced, according to Jens Schoenfeld.
* There seems to exist another hardware problem which affects only certain systems, mainly C128. Read the post for details
---End-of-Update---
Eventually some light is being shed over this fuzzy subject as
Individual Computers has now acknowledged a bug in the first CPLD version of MMC64. This problem can be fixed by reprogramming the MMC64. This can apparently be done for free (shipping excluded) by sending your MMC64 to Jens Schoenfeld or by meeting up with him at the upcoming
Breakpoint party. According to Jens there also seems to be a timing/signal hardware problem with MMC64 affecting certain systems.
Before you all start sending your MMC64 to Jens, do verify that you have a faulty cartridge by reading this post carefully and testing with the latest version of
Contiki: 1.2-devel1. This only affects RRnet connected directly to the MMC64 clockport.
Read my recent email correspondance with Jens to get the details:
RaveGuru wrote:
> I can't get the RR-net adaptor working
>properly on the MMC64. I'm initilizing the MMC64
>by writing #$0a/#$1c to $df13. After reading the
>revision information I'm clearing bit 3 at $df11
>to put the clockport at $de00 and enable bit 1
>at $de01 to activate the clockport. Still I'm
>just getting nonsense readings from the RR-net. What am I missing?
Jens Schoenfeld wrote:
There might be a problem with your type of
computer. Do you have a C128 or a C64? Many C128
computers have serious timing problems that I
haven't been able to solve with the MMC64 yet.
The Retro Replay works fine because it uses a
much older (and slower) logic chip. The ultrafast
logic chip on the MMC64 is "too fast" for a few
C128 models. Sometimes it helps to re-install the
metal shield in the C128 (many people remove that), this reduces GND noise.
Further, the clockport is re-mapped to $de20 if a
Retro Replay is detected in the pass-through port of the MMC64.
RaveGuru wrote:
>Ok. This is interesting. But I'm using a C64
>(new motherboard) with the metal shield intact. Any suggestions?
Jens Schoenfeld wrote:
None at the moment, and I have to wait until I
can consuilt Oliver on this matter. I have to
"shield" him a bit from customers, because he has
to do some exams at the university. I don't want
to bother him with these things while he's in exam preparations.
Meanwhile, can you check if the RR-Net works with
Contiki in the hardware configuration you are
having the problems with? This would outline the
problem to being hardware- or software related.
RaveGuru wrote:
>Hi again! I've now tried with Contiki 1.2-devel1 and it does not work.
>I'm not seeing a single ethernet frame coming from the RR-net as long as it's
>hooked up to the MMC64. I swapped over to the RR and retried the same
>procedure, and it worked instantly (as usual), swapped the RRnet back to the
>MMC64 and repeated 3 times to be sure.
>
>Test system: C64 model G, RRnet Rev.D, MMC64 Rev.B
Jens Schoenfeld wrote:
That MMC64 - is it blue or red? (that is: How old is it, when did you buy it?)
RaveGuru wrote:
>Hi!
>It's red, says Rev B, V1.1 and (c) 2005 on the downside.
Jens Schoenfeld wrote:
When did you buy it - just shortly, or more than a year ago?
RaveGuru wrote:
>It was bought in March last year, almost when it first was released.
Jens Schoenfeld wrote:
OK, then there's a chance that you have the very
first CPLD version. That can be updated, and I'd
do it for free if you send the card at your cost to me.
Another chance for a hardware update would be the
Breakpoint party on the easter weekend - will you
be there? I can bring equipment for the update with me.
Login
FAQ
Search
