|
10/05/1999
Embarrassing Moments and CCDs
Some weeks ago, I wrote that one of my professors at Pitt gave
us some great advice - always keep track of the units! We just
lost that Mars Orbiter because of a failure to do just that.
Specifically, English units data were fed to a computer
programmed for the metric system. The net result, crashing on
Mars, was what might be termed the first case of "pilot" error for
one of NASA''s spacecraft. I can imagine the state of mind of
those responsible for this goof on this most basic issue.
However, I can empathize with them, having made a
mathematical error of a different type on my Ph.D. thesis work
that nobody picked up prior to publication. An embarrassing
way to start my career!
But not my only embarrassment. On September 17, the U.S.
Postal Service issued a stamp commemorating the invention of
the integrated circuit (IC) at Texas Instruments in 1958. A
ceremony recognizing the inventor, Jack Kilby, was held in
Dallas on that day and quite a few former colleagues of Kilby
were present. Two years ago, I met Kilby at a Bell Labs (Lucent
Technologies) gala celebration of the 50th anniversary of the
invention of the transistor. Kilby is a very tall, imposing fellow
and, when I saw him, I confused him with George Dacey, a
former colleague at Bell Labs whom I hadn''t seen in many years.
George, another towering figure, also served on the board of
directors of W.R. Grace, a company whose stock I own. So, of
course, I sidled up to Kilby and said that I was glad to see him
again and that I had voted for him many times for his board post.
Kilby, known as the "humble giant" at Texas Instruments,
seemed perplexed but gracefully accepted my statement and I
walked away. Later that day, when he was introduced to give a
talk on his invention of the IC, I realized the error of my ways
and later conveyed my apologies to Kilby. Hopefully, he didn''t
regard me as a complete idiot!
At this same gala affair, I was quite pleased to see, and properly
identify, two former Bell Labs colleagues, Willard Boyle and
George Smith. These two gentlemen were also inventors, of
something so important that I told them I couldn''t understand
why they had not received the Nobel Prize for their invention,
which has now been around for 3 decades.
In 1991, my wife and I visited Tucson, Arizona after a meeting
in Phoenix. On the way to Tucson we stopped at Biosphere 2,
the glass-domed 3-acre enclosure designed to simulate a self
contained environment that might some day be realized in space,
perhaps on the moon or Mars. We watched as the biospherists
worked inside in the "ocean" and in the "desert" areas. There
was some controversy over this project, funded largely by one of
the wealthy Bass brothers, and there were unexpected problems
with buildup of carbon dioxide to unacceptable levels, partly due
to an El Nio. I believe that the original goal of housing a
number of biospherists for a period of 2 years was met or nearly
so but will try to check this out for a later column.
As usual, I digress. On this trip to Tucson we also had another
space-related experience, a trip to the top of Kitt Peak, site of
several large telescopes, also housed in domes. We took a
guided tour of a couple of the telescopes and saw many beautiful
pictures of our universe. Inside one dome, the guide pointed out
that the workers below were engaged in the installation of
charge-coupled devices or CCDs. I informed the guide that I
knew the two inventors of the CCD, namely, the aforementioned
Bill Boyle and George Smith. In fact, Boyle was my executive
director at Bell Labs for a while when I worked on light emitting
diodes.
What is a CCD and why do I feel that it deserves a Nobel Prize?
A couple weeks ago, we considered the various kinds of
microscopes for looking at features as small as an atom. The
CCD is a device that can be said to enable us to look at just the
opposite, the monstrous structures in our universe. Indeed, I
have read a technical article that had the title "CCDs make
modern astronomy possible". But that''s not all. CCDs are key
ingredients not only in telescopes but also in many more
mundane things such as cameras, fax machines, scanners and
other types of applications involving images.
The CCD as an imaging device involves an array of pixels
connected in rows. For example a 2048 x 2048 array means that
there are 2048 rows, each row a string of 2048 connected pixels.
If you''re not familiar with pixels just look at the "o" here. You''ll
see the dots, or pixels, that make up the "o". Each pixel in a
CCD is an electronic circuit element with a metal oxide
semiconductor (MOS) capacitor, a device that can store an
electrical charge. You may remember that we have discussed
MOS devices earlier in these columns and that they are also the
key devices in the RAM memory in your computer. Each pixel
also incorporates a photosensitive element, that is, a device that
detects light and has an output that feeds a charge to be stored in
the MOS capacitor. The amount of charge fed to the capacitor
depends on the number of photons striking the detector and on
their energy or wavelength. Thus, if you''re looking at a distant
galaxy, just as with a film, the intensity of the light will vary
from pixel to pixel and hence the amount of charge will also
vary. Let''s say you want to get a picture of a very dim faraway
galaxy or star. Just as with a camera and a low light level, with
your telescope you have to have a longer exposure to get an
image.
After a while you have your array of pixels with their different
charges. Now what? I neglected to tell you that at the end of
each row there is an "output gate" which can convert the amount
of charge to a voltage proportional to the charge it receives.
Suppose we have a circuit for each row that moves the charge
from each pixel to the next pixel step by step, "charge coupling".
By doing this and with appropriate electronics, the output gate
and associated circuitry can be "clocked" to record the charge in
each of the 2048 pixels in the row. The other 2047 rows can also
be clocked to allow their output gates to record their pixels''
charges as voltages. By using your computer to convert these
voltages to shades of gray or appropriate colors, the beautiful
pictures you see from the Hubble or other telescopes emerge.
That''s great you say, but why not just use good old photographic
film? It turns out that with film, for the more distant/dimmer
objects it might be necessary to expose your film for many hours,
even many nights, while carefully tracking the star''s or galaxy''s
position. With CCDs, because of their much greater sensitivity
compared to film, the time can be cut to minutes or at most a few
hours. Not only are the CCDs more sensitive, but they can be
made sensitive to not just the visible spectrum but to the X-ray,
ultraviolet and infrared regions of the spectrum, giving a much
more complete picture of what''s going on in this hyperactive
universe.
There are other, more technical advantages of using CCDs
compared to film but also some disadvantages. One is that film
has greater resolution; however, with increasingly large arrays of
pixels the resolution becomes quite good for the CCD. A
disadvantage is the amount of data the CCD delivers. For
example, our 2048 x 2048 array delivers 10 million bytes (10
megabytes) for each picture! (For those who recall our math
column on bytes and bits, 2048 x 2048 = approximately 5 million
pixels. I assume the 10 megabytes I''m quoting comes from
having to use two bytes to store each pixel''s data.) My floppy
disks only store 1.5 megabytes so I''d need 7 of them to store one
picture!
I was heartened to see that Boyle and Smith have not been
forgotten after 30 years and that Nobel Prize is still a possibility.
Just this past June they received the first "Breakthrough Award"
at the IEEE (Institute of Electrical and Electronics Engineers)
Device Research Conference. I''m not sure now whether this
device research conference was under the same auspices as the
Solid State Device Research Conference I attended in 1967.
That conference was in Santa Barbara and was the only such
conference that I''d been to. It was a memorable affair, especially
the banquet, at which one activity consisted of singing ribald
lyrics to popular tunes, apparently a tradition among the device
crowd. We more staid members of the audience made our
departures when beer-soaked napkins were starting to be thrown
across the room. Sorry, Brian, I''m not much of a beer drinker, a
gin and tonic being more up my alley.
One last note. Bill Boyle''s contribution to astronomy was not his
only space-related effort. In the 1960s, he also was one of those
who helped NASA select the site for the Apollo moon landing.
Fortunately, they must have kept track of their units in those days
and did not go sailing right on by the moon!
Allen F. Bortrum
|
