Hi Jeff,
On 3/20/2011 4:11 PM, Jeff Liebermann wrote:
[attributions elided]
Actually, very accessible, if they're interested. We have a local
-----------------------------^^^^^^^^^^^^^^^^^^^^^
I seem to recall making comments to that effect ;-)
startup, Elecraft, that makes ham radio kits and accessories. Most of
the qualified hams in the area are now working for them.
<
http://www.elecraft.com>
There's not much design work being done locally, but the engineers and
techs certainly know how the radios operate.
I have a similar problem. When I was still designing electronics, the
major obsession was reducing the cost. At that time, it meant
reducing the component count. If you get my attention, I can waste
enormous amounts of time, on futile exercises in component count
reduction. Same with unusual ways to implement simple things.
In the mid 70's, I had to design a replacement "front panel"
for a Nova (2 or 3... I can't recall) minicomputer. This is
the vintage equipment where you "bit-switched" (toggled) data
into memory (real core!), forced a value into the program counter
and then hit "run".
The actual front panel was bulky -- all those "bats" on the front
panel switches plus plastic molding, etc. Since the machine
only had to do certain very fixed things (I guess it qualified as
an "embedded system" -- even though the processor was the size
of a small washing machine!), it was silly to ship the bulky
panel with the unit.
The panel worked by driving values onto an "open collector" bus
and stroking control lines. So, any replacement "front panel"
needed logic to interface to a few dedicated switches (reset,
IPL, etc.), decode logic to drive the right values onto the
bus, some crude timing (remember, we are emulating things that
a human being does "by hand") and, of course, open collector
drivers to actually *drive* the bus.
I ended up with a design that used a 7 segment decoder (!)
wired in a bizarre way to (coincidentally) get the right
values onto the bus based on an examination of the switch
states connected to the BCD inputs thereof. (this predates
FPGAs, GALs and even *PLAs*!)
My boss was impressed -- but too scared to risk such a bizarre
implementation: "How will anyone ever know how it is *supposed*
to work?" (to me, that was pretty obvious: the devices work
just like any other devices... read the data sheet!)
You can get a fair idea of the number of hams by diving into the FCC
ULS database and searching for counts by your local zip codes.
<
http://wireless2.fcc.gov/UlsApp/UlsSearch/searchAmateur.jsp>
For 95005 (Ben Lomond, CA), it shows 61 licenses. Unfortunately, the
FCC updates the data perhaps every 10 years, so some hams that have
moved out or died are still listed.
Lots of ZIP codes in 9000 square miles. I'll leave that as an
exercise for the student...
Worse, most hams seem to live in post office boxes.
Yes. Though, if you're smart, you use a non-USPS "p.o.box"
We have one local supplier (other than Radio Shock). They're
expensive, but not outrageous.
<
http://www.santacruzelectronics.com>
I bought some small lytic's for ~$2/ea. C'mon, now... (then
they complain when folks buy mail-order from firms out-of-state!)
Yep. Just finding the part in my mess is difficult. However, I do
have a fairly good collection of 1980's vintage discrete parts, which
are useful for raising the dead and repairing the antiques.
I find more use for *hardware* (screws, connectors, etc.) than
"components". As I get older, time is more valuable to me than
"mundane cost saving experiences" (though I replaced a power
cord on a make-up mirror, yesterday... can you spell "mindless"?).
So, if anything "complex" dies, I typically just recycle it and
replace it (I collect "spares" of lots of things so this often
just means taking another off the shelf). To invest any serious
amount of time, it either has to be a really interesting problem
*or* something of considerable value (including "irreplaceable")
One problem with specialization is that it tends to create some rather
bizarre implementations. I once reverse engineered a microprocessor
design that could have been replaced by a quad op amp.
I had a colleague design a state machine for an LCD controller
using a 64KB EPROM (back when a 64K EPROM was *big*!). When I
dumped the EPROM, I discovered it implemented a simple logic
function of *3* variables (despite all the "inputs" that
were wired to it's address lines). I.e., it could have been
replaced by a quad _______ or, perhaps, a dual AOI.
Hint: I never build or design anything that I can't sell or turn into
a sellable product. When I do build something for myself, I usually
build 3 to 5 extras, to sell to friends and recover my expenses.
Expertise in PCB layout is a must.
I tend to take the *opposite* approach. I like building "one of
a kind" pieces. Especially as gifts. I.e., the value of the
gift is the fact that it is unique. And, sufficiently expensive
to undertake "duplication" that it would need to have a significant
number of "buyers" to make it worthwhile (i.e., it is not
economically feasible for a "competitor" to make something like
this, thereby rendering the uniqueness of the gift meaningless).
Difference Engine would be cool as it would require lots of
*mechanical* skill/assembly/fabrication -- not something you
could just run off a set of Gerbers!
Anything that I design lately that would appeal to others
I just release as "open source" (hardware and software) so
anyone who wants to invest the time/energy can copy at will.
Note that it's still possible to have fun with electronics:
"How to blow up a microwave"
<
>
I've thought about setting up a tesla coil proximate to
the hummingbird feeder to zap the *woodpeckers* when they
try to use it. But, it would be hard to control where
the discharge went... :<
(I wonder what burnt woodpecker smells like?)
