|PCB : Silicon ratio|
Message #12 Posted by Joerg Woerner on 7 Nov 2010, 10:51 a.m.,
in response to message #11 by uhmgawa
I always enjoy the recent tear-downs of various gadgets, e.g. the iPhone, reading "company xyz managed to stack an ARM9, a SDRAM and even a Flash memory in a so-and-so housing".
Well, more than 30 years ago (in 1979) the engineering school I attended purchased a bunch of the latest personal computers, the CBM 3032 series. Okay, just one CBM 3032 for the teachers, the students worked with the CBM 3016's.
One day, after school, I opened the hoods of both the 3032 (32k RAM) and 3016 (16k RAM) and recognized immediately: The fully loaded Commodore had two rows of 16k*1 dynamic RAM installed, the 3016 featured only one row and ONE ROW of neatly alligned 10 mm holes, drilled carefully into the PCB at the original places of the optional RAM's!
After studying both the layout and datasheets, I asked the other day my teacher, if I could borrow the CBM 3016 over a weekend for an "important project".
He agreed and I piggy-backed a set of MK4116 chips carefully one-to-one except the RAS and CAS pins, that I hand-wired to the left-overs of the traces ;-))
Wow - you should have seen his face after booting the CBM 3016 the following Monday to make sure the CBM is still working and he read the 32k RAM-Test message.
FYI: I earned a few hundered Deutsch Marks by updating all the CBM 3016's of this school and even another school.
OOPS: And later, collecting TI calculators, I learned that TI invented piggy-backing already in 1975 with the SR-51.
That's about the fuzz of stacked silicon ;-))