tracing out schematics

[Tony Duell]

> I've read in the past that Tony called someone a whimp cause he said he
> didn't dare to reverse engineer a 4 layer board or something like that. I

It may hgave been somewhat tongue-in-cheek :-). I've done things a _lot_ 
more complicated than a 4 layer board and lived to tell the tale. But 
then I've done rather a lot of this...

It's got to the stage here where for a simple board (say a PC I/O card 
with one known 'big chip' and some glue logic) it takes perhaps an 
afternoon to produce a complete, annotated, schematic. But against that, 
something like the HP9100B calculator took 6 _months_...

However, I would always claim a schematic can be produced (direct 
chip-on-board being a possible exception!). And therefore if the choice 
is between having to draw out a schematic or not repairing the machine, 
the former always wins!

> wonder, if _that_ is easy, how does one solve problems like traces that
> run under components? If its a simple and cheap component, it could be
> removed and replaced later.

In generaly you trace connections with a meter, not just by eye (and 
don't worry about scrapping off the soldermask to see which of the 
half-dozen traces coming out from under an IC corresponds to the one 
you're tracing!). So traces under components (or on middle layers of the 
board) don't worry me too much. But you may well have to desolder some 
components -- as I mentioned in the other message, you must remove 
anything that tests as a dead short.

-tony