Things that aren't quite right, and things you should be aware of.
The controller board (and occassionally the peripheral boards) sometimes
reset when a big relay or motor is switched. It only seems to happen
when the CREMs are switched, the purchase state relay is switched, or
the Coke cam motors turn. On my stack, it would sometimes reset when a
nearby motor (which wasn't connected to the controller in any way) would
switch. I'm really baffled. We need to try putting a grounded metal box
(or aluminum foil) around it to see if it's EMI.
There is no way for the controller to detect if a soda stack is empty.
I think that the best way to go about doing this is to mount our own little
switches under the big "column empty detection" switches.
Dimes are not accepted by the coin changer unless it is tilted at a jaunty
angle. Tilting it requires repositioning the screws.
The Money module always says that the nickel tube is full. Something must
be wrong with the connection on either the coin changer board or the
money I/O board.
Parts of the LED occasionally go "dim". Pushing your thumb on the chip
in question tends to fix the problem, albeit temporarily. I think the
If you use the Program module to program a kernel that is so bad that you
can't issue commands or upload resources, then you won't be able to
reprogram it locally and you'll have to go down to Moore. (There is
currently no HC11 bootloader, though maybe there should be. Or maybe
something in the HC11 EEPROM that can program the external EEPROM.) In
any case, always assemble your new kernel as a module (use
kernelmoddasm) and test it out in RAM before actually programming
it into EEPROM.
Sometimes the controller would spontaneously spew out a ton of garbage
messages (some of which included previous real messages). This may have
only happened when an inductive load was switched, and be related to the
internal RAM resetting (but not the processor itself). CokeOS now does
RAM and message queue integrity checks, so this may help shed light on the