Actually, I've found that, after you learn the basics of boolean
logic (and / or / invert) and play around with some of the
basic TTL or CMOS Ic's implementing those logic elements, one
of the best places to get "instruction" was to look through
the "application notes" put out by the various IC manufacturers.
Unfortunately, today's ICs are so complex, that application
notes are little more than "how to wire the super-whiz-bang-thingy
to a 8051 or pentium" so they aren't much good for learning.
Check your local used book stores for old data books - I see them
all the time around here - but we are in a college town... The
application notes are sometimes in the back of these books and
sometimes printed as separate volumes. Unfortunately (or perhaps
fortunately) those are collectable, too...
When I started to design my own stuff, one of the first things
I found essential was an oscilloscope. A good used, relatively
inexpensive, Tek or HP dual-trace, triggered-sweep, 60 MHz (or
so) scope will cost you around 100 to 200 US dollars. I even
purchased a Tek 453 on eBay (or is that eOverPay?) for about $150
and it was in near perfect condition. If you don't mind "bulky" and
can make it to a "ham fair" or "radio rally", look for some of the
older Tek "boat anchors" such as the 535 or 545 or the newer, lighter
56x series (early to mid 60s.) They usually run somewhere around 20
to 50 US dollars over here. They are slower (10 to 30 MHz is common)
but that may not be important for learning or restoring old "classics."
Unfortunately, these "classics" are quite heavy, so long-distance
shipment usually increases the cost beyond their value.
I used a 535 (and later a 545) to help keep an old home-brew computer
running at Oregon State (many years ago.) The basic machine clock
was around 22 MHz (seems fast for 1971, doesn't it?) but it was
a "serial" machine so it really had only a 100 micro-second word time.
In any case, a scope is SO MUCH better than a logic probe. But speaking
of logic probes... Although not a "boat anchor" or "classic", Radio
Shack had (do they still sell them?) a little "logic probe oscilloscope"
for about 100 US$. Cute little thing had a 20 by 30 pixel (approximate)
screen built in to a hand-help box about the size of a normal logic
probe. Basically it was a 5 MHz triggered-sweep storage scope. Comes
with a cable and software to plug into a IBM compatible for a "real" display.
Best "logic probe" I ever purchased. But I still prefer to use
"classic"
scopes to work on "classic" computers...
Gary
At 03:14 PM 10/3/98 -0700, you wrote:
hard to come by these days, but how about modern
equivalents? I think
that's the lowest level I'll personally go, although it would be really
fun to build a 4-bit computer out of transistors, anything more complex
is probably really difficult.
Why not go for a bit-serial (aka 1-bit) CPU? It's an extremely classic
design, and is certainly the way to go to minimize total transistor
count.
OK, here's a good question for everyone!
Does anyone have really good references on learning basic computer
design and implementation? To put things in perspective, right now
I'm at a "Getting Started in Electronics" and "Introduction to Digital
Electronics" Radio Shack level -- please hold back your laughter :)
I've just always been a software guy all my life. I've never taken
an EE course, but I'm definitely eager to learn, as it would have
tremendous benefits in my classic computing hobby.
At least I do have a nice Fluke 79-III multimeter and a Weller
temperature-controlled soldering iron, so I've got a good start.
And many of the posts on this list have been invaluable -- especially
the soldering/desoldering tips that were passed on ;)
If anyone has any good tips for books that are still available, I'd
be tremendously appreciative.
-Seth
--
"The crux of the biscuit Seth J. Morabito
is the apostrophe." - Frank Zappa sethm(a)loomcom.com