On 2013-07-29, Jim Wilkins wrote:
"DoN. Nichols" wrote in message
...
On 2013-07-28, Jim Wilkins wrote:
Windows polls them, destroying your setup.
Hmm ... no way to tell Windows to keep its hands off those
ports? Another reason not to like Windows, then. :-)
Windows periodically checks printer status with the control bits, and
has a fit when it sees some combinations of the status bits. AFAIK it
leaves the data bits alone. I wrote a program to display and control
each bit.
Can you uninstall the port from within Windows, and just let
your program handle it without interference?
Sounds like the team got the payback satisfaction, then. Were
the problems in the code a result of some limits which he put on the
team, thus making it even more appropriate that he get the
o-dark-hundred calls? :-)
A state-of-the-art semiconductor test station is a very complex and
tricky machine that the IC designers may use in ways they wouldn't
reveal to us. Like machine tools some combinations of operations will
misbehave or crash.
:-)
For example to measure the resistivity of certain IC structures you
might have to apply 100 Volts and measure the resulting current flow
in picoAmps. 100 V in the wrong place can do a bit of harm.
Understood. Many logic ICs are allergic to more than about
5.5VDC. :-) But can the people using it understand that what they did
caused the problem and look for another way to do it instead of calling
for support? (Or were they calling to find out what needed replacing
after doing that? :-)
The current sensing circuit for that was an AD515 op anp integrator
with a 10pF silver-mica feedback cap. Its residual noise level was a
few hundred femtoAmps.
Pretty tiny.
At those levels the error current from dielectric absorption can
appear huge, so we had W.L.Gore make us special Teflon-insulated reed
relays for the switching matrix. Even they weren't immune but the
absorption current dropped below a picoAmp within 5mS, which was
acceptable.
http://en.wikipedia.org/wiki/Dielectric_absorption
I remember as a kid playing with components removed from old
tube radios, I found that the wound paper-foil capacitors common at that
time (early 1950s when I was taking them apart, and figure them to be
perhaps 15-20 years old at that time), I could almost always get a click
when connecting something as small as 0.01 µf across headphones -- even
after leaving them shorted for a few hours.
And I remember reading about electrets (form a capacitor by
melting the dielectric between electrodes at voltage and letting it
solidify while the voltage was applied. This was used to polarize some
condenser microphones at one period -- before phantom power became
common.
Simply not tightening the shelding enough raised the noise
dramatically. Without the shield it was a fine intruder detector.
:-)
Surprisingly I needed only a space frame box with wide-open sides to
fully shield it for the test and calibration fixture. The circuit
rejected 60 Hz and above very well and the seemingly gaping access
holes were too small for lower frequencies.
Electrostatic fields only -- not electromagnetic?
At a previous job I had designed and built a test station that
measured the reverse leakage of TV high voltage rectifier diodes. The
circuit forced 1 microAmp at up to 40,000V. The sample diodes measured
over 25KV at that microAmp. The current regulator was a 1 MegaWatt
(pulse) radar transmitter tube with X-Ray shielding.
Hmm ... I remember at Transitron (early 1960s) one item on the
production line (long before computer controlled tests) was a reverse
voltage leakage test for high voltage diode assemblies. Think about the
size of a billy club made of a spiral of diodes potted in black epoxy,
and perhaps 12" long with an anode cap (like from a tube) on both ends
Current was measured with tube based OP-AMPs (Philbrick IIRC).
At the time, an OP-AMP was a black box to me -- no idea how to use it.
Now, I love them.
And the high voltage was developed by a voltage multipler
(perhaps 12-15 stage) using somewhat more reasonable size diode
assemblies.
The door was set up with a knob which you had to turn through
about 40 revolutions to open it up -- and the first few turns opened an
interlock switch and lowered a discharging bar into contact -- long
before you got enough turns to open the door. The threads were buttress
threads engaging spring wires, so you could just shove it closed quickly
-- a nod to production efficiency. :-)
The only use of computers then and there was a mainframe being
used to match forward thermal characteristics of stabistors (diodes
selected for their forward voltage characteristics, zener diodes, and the
base-emitter forward drop of a transistor -- all checked at -50C, +50C,
and +150C in silicone oils. They would test hundreds of each at a
time, punch cards for each, and feed them into the computers to select
the best match of the three for building "ref-amps" (a starting point for
a voltage regulator). The selection process was not perfect -- we got a
selected number from a bach back to test from the customer. They were
right at the three temperatures, but out of spec between them. :-)
Enjoy,
DoN.
--
Remove oil spill source from e-mail
Email: | (KV4PH) Voice (all times): (703) 938-4564
(too) near Washington D.C. |
http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---