> > I thought of throwing that idea out there, but having lived through the great > RAM famine the thought of a spare > RAM chip seemed kind of silly. Incidentally, the RAMs are all soldered on my > Camintonn boards.
All the RAMs (but not the buffers) on the NatSemi boards are in sockets. And I am not sure this is a Good Thing, they are not good quality sockets at all. They're the very cheap ones that make contact with one side of the IC pin only. But replacing the lot with turned pin is something I will only do if I start having RAM problems on those boards. I've done a little buzzing out with an ohmmeter : The green LED + a 270R resistor are across the 5V line (which is presumably the battery backed memory supply line. if you have that optio). I assume this indicates the board is powered up. The moving contact of the switch is grounded. One of the other contacts of the switch goes to the yellow LED + 270R to +5V. So as suggested, the yellow LED indicates the state of the switch. The other contact goes into the logic. I've not traced it all out yet, but it would appear to disable many of the buffers when it is low (i.e. when the yellow LED is off). So the idea that the yellow LED is on for normal operation and the switch can turn it off and disable the board makes a lot of sense. That 'extra' RAM gets power, but nothing else. As far as I can tell none of the other pins go anywhere. Isn't that a bad idea, to have a MOS device with all inputs floating? I would have thought it would have been better to tie eveything (including the +5V pin) to ground, or to power the RAM but have pull-up resistors on all the inputs. -tony
