[net.micro.6809] Help with 64K expansion

jeff1@garfield.UUCP (Jeff Sparkes) (01/10/84)

	I bought a 64K expansion kit from Jameco, but I don't fully 
understand the installation instructions.  They say:
	"Remove or disconnect the follwing capacitiors" and then proceed to
list 8 capacitors.  The capacitors in question are located between the
RAM chips.  (Specifically, they are the left capacitor of the pair.)
What I don't understand is should I just yank the capacitors out and leave a 
break in the circuit, or put some solder in between so that the circuit
is complete? 
	I know next to nothing about this stuff, so I need HELP!!!!

					64K is only chips away!
					Jeff Sparkes
				(utcsrgv||allegra||ihnp4)!garfield!jeff1

emjhm@uokvax.UUCP (01/21/84)

#R:garfield:-79100:uokvax:3500027:000:2315
uokvax!emjhm    Jan 20 00:03:00 1984

  If you have a Technical manual handy, the explanation will go a little easier.
The trick in replacing the 16K RAM's with 64K RAM's is in getting two more
address lines from a 16-pin chip.  In the 16K mode, the SAM chip generates
the 14-bit address to the RAM by placing the lower 7-bits as a column address
strobe and then the higher 7-bits as a combination row address strobe and
read/write operation.  One additional address line is needed on the 64K RAM's
such that 8-bit col and 8-bit row make up the 16-bits of address to the chips.
It so happens that the capacitators that you are told to yank out are bypass
caps for a power supply connection which was given up to the additional
address line.  The 4164's only need a single 5 V supply in lieu of the
three supplies needed for the 4116's.  The bypass caps are there to short
ac to ground(power supply transients) and also to keep the varying loads of
the individual chips when they turn on their data outputs from stealing the
bias they need to keep operating.  These caps must be removed so that the
newly assigned address line doesn't see a ground when it asserts.  What it
would actually do is delay the address assertion to the point that things
wouldn't be there at the right time.  Under no circumstances should the
caps be replaced with jumpers.  This could cause damage to the 74LS02 
which drives the extra address line for the 4164's.  It would also cause the
high order bit of the low and high address lines to always appear as a logic
ground.  The jumpers that must be moved are merely swapping the old power
for new address, moving the 5V to two different pins.  Notice that there
is still one pin extra to make way for the 256K RAM's and still be pretty
much compatible.  Another jumper is sampled by the initialization software
through the system PIA chip so that the SAM chip can be told to refresh
all 64K of memory.  Another jumper out of the SAM chip routs the new
address line through a couple of NOR gates so that the row and column address
strobes are only generated in the upper 32K only when the SAM chip is told
that the upper 32K is RAM and not ROM.  The cludge of haywires on the "E"
and older boards is the extra logic required to disable refreshing the
upper 32K of RAM when the ROM is being used.
Hope this helps.

Jim Miller