The sound was quite audible throughout the room. The age of the vacuum tube is over, he announced, the age of solid state electronics is here. Prof. Boone talked to a half dozen or so of his vacuum tube students afterward and suggested that perhaps Dr. Morton was a bit premature. But no doubt, he said, the only thing constant in technology was change and obsolescence. Around 1954-55, Pierce wrote a proposal and got a contract to build a broadband IF amplifier, 100-500 mhz if I remember, using a close
spaced grid tetrode. A tapered element bandpass L-C network
would work between the small capacitance of the plate of one tetrode and the high capacitance grid of the next tetrode. I was given a chance to redeem myself by bringing this concept to fruition as the project engineer. I built 2 tubes and the
coupling network before advances elsewhere made the project moot and the work was abandoned. I did, however, discover an interesting and unexpected phenomenon along the way. The grid wires were .0003 inches in diameter spaced 0.001 inches above the
cathode. They were about 1/4 inch long wound between precision ground tungsten support rods. The grid structure was about 1/2 - 3/4 inch long. It turned out that the inductance of the individual grid wires was high enough to make the entire grid structure, seen as a transmission line, almost 1/4 wavelength
long at the top of the band. This was not a fatal flaw, once the far end was properly terminated, but it was not anticipated and came as a great surprise when the first tube was built and tested with the far end of the grid left open. I left Bell a year or so later and came to Varian Associates in Palo Alto to join Willis Yocom who was in charge of and R&D effort to put Varian into the TWT business.

I am writing these memoirs for my g'kids and have absolutely no
plans to publish beyond One Reader At A Time. There is a lot of
personal stuff in them that would be of no, none, nada, zip
interest to anyone except close friends and family. I would not
object to leaving them in an archive somewhere where the
occasional archaeologist might come across them. 

The following was prepared for Bob K. (a Varian alumnus) who used
it as a prop for a moderately tasteless joke that only a few
insiders might appreciate. His Thesis: Since the U.S.A. won
WWII, Russel Varian and his brother, Sig, are recognized as the
inventors of the klystron. If, however, Germany had won WWII,
Oscar Heil and his brother, Sig, would be recognized as the
inventors of the klystron.**** Sorry 'bout 'dat, but not very.

Bob: You are almost right. If Germany had won WWII, Oscar Heil
and his wife would have been the inventors of the klystron. They
published a theoretical description of velocity modulation on
electron beams and indicated how the principle could be used to
make a microwave amplifier, but Russell and Sigurd Varian built a
working model at Stanford at about the same time. This work was
based on Russ' intuition with little or no theory to go on. All
indications are that they were unaware of the Heil's work. I am
also under the impression that the facts in the laboratory set the
theoreticians at Stanford into a frenzy of activity trying to
explain what was being seen. 

Heil was captured by the US Army in Germany before the end of
the war and, since his name was on a list of important German
scientists, he was hustled off to the US. He was sent to Dayton,
Ohio where the Air Force had a vacuum tube research facility. 
After the war, this operation was transferred to Ohio State and Dr.
Heil came along with it. I think his wife was captured by the
Russians and sent back there somewhere.

After I graduated from the U of F in 1950, I sent off several
scholarship applications hoping to continue my studies toward a
PhD. Ohio State came through with an offer, which I eagerly
accepted, of an assistantship to work in the vacuum tube laboratory
there. I was given $100/month plus tuition and was assigned to
work as Dr. Heil's assistant. This was a notable and unique
experience. 

Prof. E. Milton Boone, a descendent of Daniel, was in charge
of the vacuum tube lab. Pete was his foreman, primarily a tool and
die maker familiar with all manner of machine shop equipment and an
expert tube builder. Tony was our glass blower from England. He
had played a role in the microwave vacuum tube R&D effort over
there since the early 1930's and had a wealth of war stories and
jokes to tell. Sue and Ellen were our assemblers. They had both
been trained extensively at Bell Telephone Laboratories. Dr. Heil
was a staff scientist reporting to Boone. There were also half a
dozen people, including Dr. Jim Ebers of Moll-Ebers fame, who
either had or were close to getting the PhD under Boone. They all
taught classes and/or had government contracts to work on various
vacuum tube projects. I was one of three new people working toward advanced degrees under Boone including John Sullivan and Jack Cook.

I learned early in the game, probably from others, about the
Heil's early work on the klystron and how he was a little bitter
about not getting his just share of the fame and fortune. I was
familiar with klystrons from being a radar technician during the
war and I was pretty interested in all of the associated lore. 
There were several R&D contracts between Ohio State and the Air
Force for the development of high power, meaning 1 watt more or
less, millimeter wave klystrons and Heil was involved in all of
them. I think he was a consultant as opposed to a principal
investigator, but all of these projects depended on the Heil gun
which was still in the experimental stages. In any case there was
ample opportunity for friction between Heil and the PhD candidates
who were responsible for the success of these projects. Each could
blame the other for any shortcomings, and they did. 

The Heil gun was a rival of the Pierce Gun which was based the
theory of electron flow between concentric spheres. In a
convergent Pierce Gun, the cathode was a segment of a sphere while the anode bore little resemblance to a sphere because it was
necessary to put a hole in it to allow the electron beam to pass
through into some sort of electron beam device. As the distance
between the anode and cathode was reduced in order to draw more and more current, the effect of the hole in the anode was to reduce the electric field at the center of the cathode. The resulting beam
tended to be somewhat hollow with higher current density toward the outer diameter and lower current density near the axis. Heil's
idea was to make the cathode a segment of an ellipsoid in order to
bring the center of the cathode closer to the anode and thus
improve the uniformity of the current density. The main criticism
of this innovation was that the resulting beam was not as laminar
as one could obtain with the Pierce gun. I was never able to sort
out the controversy, but I did notice in later years that everyone
used the Pierce gun and no one had ever heard of the Heil gun.

All of the klystron projects and one or more versions of the
Heil gun were assembled on brass plates which also served to close
off a demountable vacuum pump. A thick walled glass cylinder,
called a bell jar, roughly 6 inches in diameter by 12 inches long
served as a standard vacuum chamber in our labs. Heil explained
that this glass cylinder was called a 'bell jar' because
experiments in vacuum chambers were often assembled on a flat
mounting plate attached to the throat of the vacuum pump and
subsequently covered by a glass vessel which looked more or less
like a bell. The present arrangement was far more convenient. The
tube lab owned perhaps a dozen such vacuum stations, roughly half
of them in working condition in our facility and the rest in a
warehouse somewhere nearby. My first job was to learn how to
operate one, prepare experiments for others, and record data. 
Before long, though, I would be preparing experiments for myself. 
Every morning when I arrived at the lab I would go to each of
the active vacuum stations, record all of the meter readings, and
fill the cold traps with liquid air which I drew from a container
called a Dewar, after its inventor I was told. The Dewar was a
large glorified thermos bottle on wheels designed to hold liquid
air for several days. When it was empty I would take the Dewar to
the cryogenics lab in a nearby building on the campus where the
liquid air was made and have the Dewar refilled. The only problem
was that the many pot holes in between the two buildings were
comparable in size to the diameter of the castors on the dolly. 

Heil was a very good teacher and he carefully explained the
role of the cold trap in producing and maintaining a high vacuum in
the bell jar. He was, as it turned out, a person like myself
intimately concerned with minutia, the basic fundamental details
about the way things worked. Nothing was too trivial to examine in
detail. The first day he went over a cut-away drawing of a
mechanical vacuum pump and explained that the lower limit of the
vacuum it could produce was the vapor pressure of the oil used to
seal the mechanism. This led to a discussion of vapor pressure
and how every substance known to man evaporated to one degree or another and how working with high vacuum would involve a lifelong study of this process. In the case of mechanical pump oil, the best attainable vacuum was roughly six orders of magnitude below atmospheric pressure... not good enough for any but the crudest vacuum work. The mechanical pump, referred to as a roughing pump, provided a rough vacuum at the bottom of an oil diffusion pump, supposedly invented in antiquity by Charlie Litton, the founder of Litton Industries. This pump consisted of a water cooled metal jacket roughly 24 inches tall by 4 inches in diameter within which there was a concentric chimney roughly 2 inches in diameter. A pool of special silicon based oil covered the bottom of both cylinders to a depth of perhaps 2 inches and an electric heater at the bottom of the inner cylinder was used to evaporate the oil. 
The oil vapor streamed up the chimney and came out through several sets of holes covered by deflectors intended to give the emerging vapor a downward thrust. 

The politically correct theory of operation was that gas
molecules from the vacuum chamber above the pump diffused into the path of the downward moving oil molecules and were driven toward the bottom where the roughing pump would remove them. After a degree of trust had been established between Heil and myself, he confided that this was a ration of nonsense in his considered opinion. His view of it was that the hot oil vapor dissolved a wide range of gases like nitrogen, oxygen, carbon dioxide, etc. and carried them to the water cooled walls where the vapor condensed and flowed down to the pool of oil at the bottom, carrying the dissolved gas molecules along. The cool condensate quickly warmed up to the temperature of the pool and released the dissolved molecules which were then removed by the roughing pump. The pressure at the top of the diffusion pump was some 3 orders of magnitude below the pressure at the bottom. Most of the gas at the top of the diffusion pump was oil vapor which could be frozen out and removed by a plate located near the top of the diffusion pump and cooled to the temperature of liquid air. 

Heil was also contemptuous of the idea that there was anything
special about the silicon based pump oil. He told me that he used
a variety of oils in Germany and when one type became unavailable
due to the war, he quickly found a substitute. The diffusion pump,
he said, was an automatic refinery because all of the highly
volatile molecules got removed quickly leaving behind those
molecules with a strong tendency to stick to each other as well as
the gas molecules we wanted to remove. 

Controversies like these are very common in technology. Both
theories tend to explain the end results and it is devilish hard to
imagine a simple experiment to resolve the matter. Later in my
career I learned that it was not always a good idea to look for
and/or find a resolution to such problems. People sometimes become
committed to flimsy theories to an extraordinary degree and may
even get violent toward those who question them. This was
certainly the case in the vacuum tube lab at Ohio State with regard
to a number of unimportant issues, Heil being on one side and one
or more of the faculty or staff or students on the other. Almost
from the start, however, I enjoyed his confidence and learned a
great deal from him. A factor in this, perhaps, was the fact that
I was always eager to hear what he had to say while not everyone
was so attentive. 

Heil's early lectures on vacuum practice dealt with the
concept of the mean free path, the average distance a gas particle
could be expected to travel in space before it collided with
another gas particle. Everyone talked about vacuum in terms of
pressure, but he thought this was a misleading concept. What
really mattered was the freedom of electrons to move about without
colliding with gas molecules. When a free electron collides with a
gas molecule, an ion is usually formed creating at least two
charged particles... one or more free electrons each with a
negative charge and a free heavy ion with a positive charge. Both
caused unwanted noise in the device while the heavy ion had the
potential to do some damage to the surfaces it might strike,
particularly the cathode. The better the vacuum, the lower the
noise and the longer the life of the cathode. Almost all
electronics like radio and television and scientific instruments in
those days used small glass vacuum tubes, called receiving tubes,
and the inter-electrode spacing was rarely over 1 cm. The major
exception was the cathode ray tube used in oscilloscopes and as
television picture tubes. The roughing pump could produce a vacuum
in which the mean free path was roughly 10 cm. Heil told me that
receiving tubes as well as cathode ray tubes were mass produced in
Germany using only a roughing pump and a liquid air cold trap to
remove the oil. A getter was used to finish the job. This was a
thin film of an active metal like barium which was evaporated by
induction heating onto a portion of the inner surface of the tube
just before seal off.

Our vacuum stations were equipped with two vacuum gauges... a
thermocouple gauge to measure the vacuum produced by the forepump
and a triode ion gauge to measure the vacuum in the bell jar. The
thermocouple gauge consisted of a pair of dissimilar wires welded
to each other at the midpoint. A current was passed between one
pair of wires to provide heat while the temperature of the junction
was read by a sensitive micro-ammeter connected across the other
pair. When the pressure was high, the gas carried away the heat
and the junction ran cool. As the gas was removed, the loss of
heat by gas conduction was reduced, the junction ran hotter, and
the meter registered the effect. Heil called my attention to the
fact that the meter did not begin to register an increase in the
junction temperature until most of the air in the pump had been
removed. He explained this in terms of an early puzzle in the
history of science. Clock makers several hundred years ago
discovered that the damping of a pendulum in a bell jar was
unchanged as the gas was removed until the vacuum, as indicated by
a column of mercury, was near the limit available at the time. 
This was eventually explained by the Kinetic Theory of Gases, a
major milestone in scientific thinking. It turns out that the
viscosity and thermal conductivity of a gas is hampered by
collisions between the particles. At high pressure, there are many
particles to carry heat and transfer momentum, but they aren't very
efficient because of frequent collisions. At lower pressure, there
are fewer carriers, but each travels further and is thus more
effective. Only when the mean free path is comparable to the size
of the vessel does the reduction in the number of particles show up
as a reduction in thermal conductivity or viscosity. 

The triode gauge was also the essence of simplicity, as Heil
explained its operation. A tungsten wire filamentary cathode was
used to provide an electron source while an open wire grid
structure nearby was connected to a positive voltage supply to draw
off the current. The anode, instead of being operated at a
positive voltage in the usual way, was put at a negative potential. 
The electron trajectories carried them into the space between the
grid and the anode where they formed positive ions by collision
with whatever gas particles were present. The electrons returned
toward the grid while the positive ions went to the negative anode. 
The current to the anode was a measure of the number of ions, and
thus the pressure in the device. Heil emphasized that this
principle could be used to measure the pressure in almost any
vacuum tube. I connected up a few tubes lying about to test this
assertion and found that most of them had a better vacuum than I
would have guessed from Heil's description of production methods. 
Heil said this was because operating the tube tended to improve the
vacuum through the formation of ions which tended to get buried in
the walls of the device or to get soaked up by the getter.

The new electron beam devices like megawatt klystrons and
traveling wave tubes, which were just now being developed,
required a much better vacuum than receiving tubes did. These
tubes could be as long as several meters and might operate at
hundreds of kilovolts so any ions formed in the beam could do a lot
of damage when they fell into the cathode. Heil had a way of
taking me off to a corner somewhere or into his office with the
door closed when he had some sensitive information he did not want
spread around. He explained in this way that the problem was not
the production of a very high vacuum inside a tube, as my
professors and most of our colleagues thought, but keeping that
vacuum once the tube was sealed off the pump. Gas particles
evolved into the vacuum from deep within every surface and quickly
destroyed even the best vacuum unless the walls of the device were
thoroughly purged of those gases. Our misguided fellow researchers
had somehow come to believe that only those gas particles sticking
to the inner surfaces of the device were of interest and that these
could be purged by a modest bakeout, like overnight at 400 DgC. 
Utter garbage, in Heil's enlightened opinion, but he had long since
learned to keep his mouth shut lest he cast pearls before swine.

This temperature, 400 Dg C, was about the limit for glass
envelopes, so there was considerable interest in ways to make
metal-to-ceramic seals for microwave windows and electrical
feedthru connectors. This would allow a higher bakeout temperature
and a better ultimate vacuum. Heil had a number of ideas, all of
them too sensitive to discuss even with me, at least at this time. 
He did, however, share some of his ideas on this matter at a later
date.

Among the artifacts Ohio State had got from the Air Force was
a helium Leak Detector. This device was sensitive to the presence
of helium in extremely small amounts. It was built into a piece of
hardware that could be placed between the table top vacuum flange
on any of our vacuum stations and the bell jar. Leaks were to be
found by passing a hollow needle, from which there was a tiny flow
of helium, over the outside of a vacuum chamber. When a leak was
found, a little helium would get into the vacuum and set off an
alarm. The source of the helium was a high pressure metal tank. 
This leak detector did not work properly and Heil was delighted
when I expressed an interest in getting it working again. He
carefully explained the principles of operation along with his
diagnosis of the malfunction which was that the student, now
departed, who had worked on it previously had screwed it up. He
gave me the manual and encouraged me to get it working again.

I read the manual and compared the operating circuitry with
the actual control unit and decided that Heil was probably right. 
My first chore, therefore, was to restore the control unit to its
original form. This turned out to be easier said than done because
a number of the parts specified in the original drawings were no
longer available and I was beginning to see why the previous
student had done some of the things he did. Eventually I came to
understand how the thing was supposed to work and was able to get
the control unit to function as I thought it should. When I
connected it to the detector head, however, it was apparent that
something was wrong here as well. Heil came to the rescue pointing
out that the hot wire electron source in the detector head was
burned out and that replacing it was a trivial matter. When that
was fixed, by Heil and Tony, the leak detector worked like a
champion. 

Under ordinary circumstances there was little need for a leak
detector in a lab where all of the vacuum systems were demountable
because the pumps were so fast that a leak had to be pretty large
to be noticed at all and there were much simpler methods for large
leaks. The situation became critical only when the vacuum device
under consideration was to be sealed off the pump. This situation
arose at about the time the leak detector was put back in working
order. Another of the pieces of equipment inherited from the Air
Force was an induction heating unit, called an rf bomber. This
device used a large triode vacuum tube as a radio frequency (rf)
power source and one day this tube was discovered to have an open
heater. I was on hand when Tony took the unit apart to recover the
triode for a closer examination which confirmed his initial
diagnosis. The tube was definitely shot because the broken pieces
of the heater/cathode filament were rattling around inside and
could be seen through the glass base. Tony said there were
several replacement tubes in the warehouse and he invited me to
come along with him to retrieve one. He got the key from Boone's
secretary and we walked several blocks to the warehouse. The
collection of junk there was truly impressive... shelves, desks,
boxes, lathes, milling machines, vacuum stations, receivers,
transmitters, oscilloscopes, electronics equipment of all kinds,
much of it in some stage of cannibalization, but some of it was
perhaps in working order, maybe. It was an amazing array of stuff. 
Tony knew just where to look for the replacement tubes. There was
only one left and the glass base was broken. Tony brought it back
to the lab, though, in order to study the details of the inner
arrangement of the various electrodes.

Back at the lab, Tony studied the tube he had removed from the
bomber and said that we would have to repair it because there was
no money in the budget to buy a new one. It would never have
occurred to me up to this time that it was possible to repair such
a device, but the idea did not seem to bother Tony very much. I
was very excited at the prospect that I might be allowed to witness
the process. Pete and Heil were called in to offer their
suggestions and neither of them thought the job was impossible, but
all agreed that it wasn't trivial either. The first thing Heil had
me do was to take the tube to an X-ray lab in another facility on
the campus and he called ahead to make sure the technician there
took the kinds of exposures he wanted. I waited while the
exposures were made and the film developed. The idea of looking at
the inner workings of such a device using x-rays was an entirely
new concept to me.

This tube was a high power triode with a directly heated
tungsten filament cathode and a copper anode which also served as a
portion of the vacuum envelope. The anode was perhaps 10 inches
long by 1.5 inches in diameter closed off at the far end and flared
out into a glass seal near the base of the tube. In operation, the
anode was immersed in a water jacket to carry away the heat. The
control grid and the filamentary cathode were mounted on heavy lugs
which were sealed into the glass base. Pete, Tony, and Heil
studied the tube and the x-rays for several hours discussing
various strategies and the risks involved in each and a tentative
plan of attack was formulated. The next day, we all got together
again and re-hashed the matter, this time coming up with a modified
plan that everyone was more or less happy with. The final plan was
that Tony would first crack open the glass nipple at the center of
the base where the tube had been sealed off the pump when it was
made. He would then seal a new piece of glass tubulation at this
opening. This would be used to control the pressure inside the
tube while the base was being worked in the glass lathe and finally
to seal the tube to the vacuum pump after the repairs were made. 
The glass lathe, also due to Charlie Litton as I have been told, is
a tool designed to hold both ends of a piece of glass tubing
concentric while a portion of the glass in between is softened by a
torch. The work is slowly turned on its axis all the while to
offset the force of gravity which would otherwise make the soft
glass sag in one direction. 

Tony explained that the big danger in all this was failing to
heat the entire glass body uniformly to roughly 300 to 400 DgC at
which temperature it would be soft enough to relieve local stresses
but not soft enough to flow significantly. Any degree of non-
uniform heating was likely to produce excess local stress and
consequent rupture due to the fact that the glass expanded when
hot. The copper seal was particularly tricky because copper had a
much higher coefficient of thermal expansion than the glass. Such
a seal, due to a fellow named Housekeeper, worked because the
copper was very soft and very thin. The metal electrodes through
the base were made using a special alloy which wasn't a very good
heat conductor, but it expanded just like the glass and was not
likely to cause us trouble if the heating was slow and uniform. 

After everything had been made ready, Tony took the tube to a
polariscope next to the lathe and showed me the internal stress
patterns present in the glass. The polariscope is a pair of light
polarizing windows illuminated from behind by an incandescent bulb. 
The light transmitted through the first plate is polarized in one
direction while the second window can be adjusted to transmit or
attenuate the light from the first window. A piece of glass or
transparent plastic placed between these two windows will rotate
the polarization of the light transmitted through it depending on
the local stresses. If the piece of work is free of stress, there
will be little change in the light pattern coming through the
second window, but if there are stresses they will show up quite
clearly. No glass shop can afford to be without one. I didn't
know what I was supposed to see at this early stage of my
education, but Tony assured me that the tube was free of
extraordinary stresses.

The first operation, cracking off the nipple where the tube
had been sealed off the pump, was the most risky. With the long
copper anode held in one chuck on the glass lathe, Tony used a
diamond point to score the nipple and immediately thereafter he
pressed the soft hot end of a glass rod to the score. The nipple
popped off clean and let the tube down to air. Then he used a soft
flame from an annealing torch to heat the entire base to a
temperature of perhaps 300 DgC. He handed the torch to me and told
me to continue heating the region as he had been doing while he
sealed a new piece of glass tubing to the nipple. This involved
simultaneously heating the nipple on the tube and the end of the
new glass tubing to the softening point using a sharp flame
produced by a torch burning natural gas and oxygen. Once soft and
with a little color, the two surfaces were pressed together and a
seal made. Immediately the torch was removed and Tony applied a
little pressure inside using his mouth on the end of piece of
flexible tubing. He explained that thick regions cooled slowly and
remained soft while thin regions cooled quickly and became stiff. 
Repeated heating and gentle blowing while the work cooled was the
way to get the wall thickness to approach uniformity. 

When he was satisfied with the new tubulation, Tony connected
the end of the flexible tubing he had used for blowing to a blowing
nipple on the lathe and brought up the free chuck to grasp the
glass base. The chuck jaws were fitted with soft asbestos pads
which held the work firmly, but not too firmly. Now, with the
lathe turning the tube at about 10 rpm, he held the sharp torch to
the glass roughly midway between the bottom of the base and the
Housekeeper seal and softened a thin ring all around. When the
glass was soft, he pulled the chucks apart slightly while blowing
slightly. The glass became thinner where it was hot. Repeating
this operation a couple of times, the glass got very thin and
parted, whereon Tony softened each end in turn to form a bead in
preparation for re-sealing after the repairs had been made.

Sue and Ellen, in the meanwhile, had been preparing the
heater/cathode. They took raw tungsten wire, heated it with a
torch to bend it sharply into a long hairpin, and coated it by hand
using an artist's paintbrush to apply a proprietary activating
compound. This product, from RCA, was a carbonated thorium oxide
powder in an acetate binder. The coated wire was mounted in a
vacuum chamber and fired to incandescence by passing current
through it. The activating agents diffused into the tungsten wire
during this operation and made the wire somewhat brittle and
delicate, but not unduly so. Feeding the new cathode inside the
grid structure, which was still mounted to the tube base, was a bit
tricky, but the girls managed in style. There was a long
discussion between Heil, Tony, Pete, Sue, and Ellen as to the best
way to attach the new cathode to the lugs in the base. The
original junction had apparently been made by a brazing operation
which they were not prepared to replicate, so a degree of
innovation was required here. Everyone eventually agreed and the
girls made the connection using nickel ribbon and judicious spot
welding. 

It was now up to Tony to seal the tube back together on the
glass lathe. This was essentially the reverse of the operation
used to separate the glass. When the job was finished it was
difficult to see where the glass had been separated. I was totally
determined at this point to learn glass blowing. I had been
wondering all along how Tony planned to evacuate the tube once the
repairs were made, but now he produced from a cabinet in the lab a
special section for attachment between the tabletop flange on a
vacuum station and the bell jar. This insert had a variety of
ports and fittings including one with a piece of glass tubing just
right for sealing the new tubulation on the base of the tube. The
recently repaired helium leak detector was fitted to another port
and we were soon ready to evacuate the tube. After the system had
been roughed out and the oil diffusion pump was hot and the liquid
air trap filled, the vacuum was quite good and it was apparent that
there were no large leaks. Then Heil fired up the leak detector
and went over the outside of the system with a jet of helium to
verify the vacuum integrity of everything. 

Heil explained that it would not be possible to bake the tube
properly under the circumstances, but since we had done nothing to
contaminate any of the vacuum walls, the only gas we had to worry
about was the oxygen and nitrogen from the air, a little oil from
the pumps, and the combustion products of the flame used to work
the glass. These would be confined to the inner surfaces and would
be relatively easy to purge. A makeshift oven was constructed
around the entire tube using several layers of aluminum foil and a
pair of heat lamps were used to heat the interior. The vacuum
gauge initially indicated that the pressure was near the lower
limit available with the system, but as the tube warmed up the
pressure increased as expected indicating that the gas trapped on
the inner surfaces was being liberated and pumped away. Tony and
Heil and I sat up watching the pressure go higher and higher and
finally lower and lower until nearly midnight. We filled the
liquid air trap and went home to bed. Early the next morning I
arrived to find Heil and Tony on the scene. The tube was hot,
roughly 250 DgC by their estimates, and the pressure was near the
lower limit available. 

The heater/cathode was connected to a power supply and a small
current was started to heat and outgas the cathode. Quite a bit of
gas was liberated at first, but as the filament began to show
color, the gas evolution subsided. Heil connected a current meter
between the grid and cathode and found a substantial amount of
thermal emission. Heil explained how this all worked and how
electrons fairly boiled off an active cathode. They came off with
such great velocity, moreover, that a retarding potential of
several volts might be required to turn all of them around and stop
the flow. This was a good means of measuring the activity of a
cathode, he said, and ours was definitely active. After a
conference with all concerned the decision was made to seal the
tube from the pump after a cooling off period, say, around lunch
time. This was done by Tony using a sharp flame to heat the glass
tubulation to the precise temperature to allow it to collapse on
itself. Once the opening was closed, the heat could be turned up
and the glass softened and pulled apart. I had not noticed
earlier, but Tony had made the tubulation a bit thicker at the
place he intended to heat for the seal off operation. Before we
went home for the evening, the rf bomber was up and running like a
new one. This was probably the highlight of my career as an
electrical engineer up to this time and the fact that I would have
been denied this powerful experience if we'd had the money to go
out and buy a new tube was not lost on me. 

There were, of course, detractors who pointed out that the
wages paid to everyone for the time spent was far in excess of the
cost of a new tube, but I reckoned that these expenses were ongoing
whether the tube was rebuilt or not. I was never able to resolve
the economic dilemmas implicit here, not was I ever able to
calculate the return on my investment in a new pair of shoes for
one of my kids.

Over the next several days, Heil gave me a number of lectures
regarding the experience we had just been through repairing the
triode. He had a new version of the Heil gun recently completed
and the two of us spent several days testing it and taking data. 
There were frequent arcs and gas bursts at first which poisoned the
cathode somehow and made continued testing impossible until the
pump had restored the vacuum and the cathode had recovered. Heil
gave me an extended account of the theory of oxide cathodes, but he
confessed without shame that there was a lot going on with them
that he did not understand. He believed that extremely small
traces of certain elements like chlorine or sulphur would poison
the emitting surface beyond repair, while a host of elements, like
gold for example, would poison the surface temporarily. In our
case, the poisoning was never total and electron emission would
slowly recover until full activity was obtained, usually within an
hour or less, although there were times when several hours was
required. During these waiting intervals, Heil had plenty of time
to go into all manner of interesting topics with me.

A prime concern was that he didn't want me to think that a
tube manufacturer could get away with the shabby bakeout schedule
we had used to get a vacuum in our tube. In the normal course of
events the copper anode would come from the mill full of oxygen,
nitrogen, zinc, arsenic, cadmium, antimony, etc., ad nauseam... all
deadly to vacuum integrity and/or cathode activity. Most of these
could be eliminated either by refining at or near the melting point
of copper in vacuum or hydrogen. Vacuum melting was best and
resulted in Oxygen Free High Conductivity, OFHC, electronic grade
copper. During fabrication into tube parts, however, the copper
would be contaminated on the surface with lubricating oils, at
least. These could be removed by de-greasing in Trichlorethylene,
TCE, or some equally obnoxious substitute and, finally, rinsed in
water. Pure water was best, but people often got careless and used
tap water, which contained chlorine. Chlorine was possibly the
worst of a whole host of cathode killers. The only known way to
remove it was heating to red heat or higher in either hydrogen or
vacuum. Vacuum was best, but hydrogen was far less expensive and
almost adequate. Unfortunately hydrogen firing left copper, and
most other metals for that matter, saturated with hydrogen atoms. 
The only way to remove dissolved hydrogen from metals was a long
term bakeout at high temperature, say 36 hours or longer at 400
DgC. We were going to get away with a puny bakeout schedule only
because the copper anode had previously been purged of hydrogen.

Even so, our tube would go soft in a big hurry, he said, unless
certain precautions were taken. The main worry was hydrogen
permeation through the copper anode. The water in the cooling
jacket was, without doubt, rife with free nascent hydrogen ions, H1
atoms as opposed to H2 molecules. These would zip through copper,
or most other metals, almost without impediment. Of course, this
would happen with new tubes as well, so why did these things work
at all? In normal operation the flow of electrons from the cathode
to the anode created an abundance of positive ions which would fall
into the cathode or the grid. Those that fell into the hot cathode
were quickly returned to the vacuum while those that fell into the
cold grid would eventually return to the vacuum and the gas
pressure would steadily rise. Our bomber was so designed, however,
that the anode voltage would reverse polarity for a fraction of
each cycle of the rf signal and during this interval most of the
ions would fall into the anode. A fraction of these would be
buried deep in the copper for long periods of time. Some would
migrate back to the inner surface and re-enter the vacuum, but some
would migrate to the outer surface and be gone forever. Thus the
triode would pump itself. This made a lot of sense to me, but Heil
suggested that I not share these ideas with others because they
were not likely to be well received. He also suggested that I not
take his word for it, but to keep my eyes and ears open and
discover the truth for myself.

The Heil gun used an oxide cathode and we had a lot of trouble
activating it and keeping it active. This led to a set of lectures
on the various types of cathodes I would run across in my travels. 
The most reliable cathode, Heil said, was a pure tungsten wire such
as we had recently replaced in the helium leak detector. It was
easily heated to such a high temperature that no poisoning agent
could reside on it and copious electron emission was always
available before the wire melted. The main problem was the excess
heater power required to keep it hot enough, but the wire was also
brittle and tended to break if the device received a strong
mechanical shock. If we diffused some thorium and a little carbon
into the tungsten wire, we could get plenty of electron emission at
a much lower temperature and such a cathode was not easy to poison. 
The main drawback here was that the material was even more brittle. 
The tube we repaired had probably died from a mild mechanical
impact. The heater power was also too high, if you could do
better. The Barium Oxide cathode was most widely used in receiving
tubes and elsewhere because it worked at a low temperature in the
visible red and could be formed into almost any shape. The main
problem with the oxide cathode was that it was easily poisoned. 
The dispenser cathodes, which operate at slightly higher
temperatures but are relatively immune from poisoning, were not
widely available for another 10 years. 

Heil's gun would work properly for a while and then we would
have an arc which shut down the power supply and generated a lot of
gas. He wasn't sure just what caused the poisoning, but he
suspected contamination from the pump oil somehow. It was simply
not possible to trap all of the oil from either the roughing pump
or the diffusion pump. These vapors inevitable coated all inner
surfaces of the vacuum envelope before the cold trap could take
effect. Furthermore, a high temperature bakeout which should break
down and eliminate oil films was not possible with a demountable
system, so we had to make the best of a less than perfect
situation. It would be a blessing, indeed, if someone could come
up with a completely oil free way to obtain vacuum, but Heil was
unaware of any such scheme.

Not all oil traps required liquid air. There was one vacuum
station in house, rarely used, that had a charcoal trap. Heil told
me how activated charcoal would absorb prodigious quantities of
almost all organic molecules. This was the principle behind the
gas masks used during WWI. The best charcoal was made by heating
black walnut shells to 1000 DgC while removing the evolved gases
using a vacuum pump. The charcoal thus produced was extremely
porous and had an enormous surface area capable of trapping a great
deal of pump oil. The advantage was that a supply of liquid air
was not required, while the disadvantage was that it took a long
time to re-activate it once it had become saturated. The charcoal
trap in the system we had could be re-activated using a built-in
electrical heater, but a long weekend at 400 DgC was barely
adequate.

After thinking about this for several days I approached Heil
with an idea. Why not, I asked, use one or more charcoal traps to
obtain a high vacuum without the use of any oil at all? From my
understanding of his lessons it would seem that a pot of activated
charcoal chilled to the temperature of liquid air should absorb
just about everything. Any remaining gas could be removed by
operating a triode as a pump. Heil replied without hesitation that
what I proposed was an excellent pump and an even better bomb. 
What would happen, he said, is that the charcoal would tend to
concentrate oxygen and become unstable. After WWI, he went on,
Germany was denied a whole range of explosives, like nitrates, that
could be used to make war. This put a crimp on industrial blasting
as in road construction and mining. Before the war, heavy earth
moving was often done by boring a hole in the ground, filling it
with ammonium nitrate fertilizer, adding kerosene or used motor oil
or the like, and setting it all off with a stick of dynamite. 
After the restrictions, this job was done by filling the hole with
charcoal and pouring in liquid oxygen by remote control. Premature
ignition before one could set off the dynamite was always a danger. 


Heil then related the story how, in the early days of
cryogenics in Germany, a linoleum floor exploded and killed a
worker in the lab. Oxygen liquifies at a higher temperature than
nitrogen, so oxygen from the air was condensing on the outside of a
pipe carrying liquid nitrogen and dripping off onto the linoleum
floor. When the linoleum was saturated with oxygen, a technician
entered the room and static electricity, most likely, ignited it. 
Heil didn't like using liquid air, preferring liquid nitrogen
instead, but liquid air was easier to make, so that is what we
used. 

Years later the need to provide an oil-less forepressure for
the newly discovered ion pump arose and I told everyone who would
listen about this line of reasoning and how it should be possible
to find a way to avoid the danger of an explosion. Perhaps we
could purge the oxygen out of the carbot trap before there was any
danger or perhaps we could find a non-combustible substitute. The
company chemists came up with a proprietary product on the market
called a molecular sieve which, when cooled to liquid nitrogen
temperatures, worked like a champion. We were, for the first time,
able to produce vacuum tubes in a totally oil-free environment. It
is not clear to me that better vacuums or more active cathodes were
produced as a result, however.

A time came when Prof. Boone suggested that I should be
looking for a thesis project. Before I could get into the PhD
program, I would first need a master's degree and writing a thesis
based on experimental work was one of the requirements. Heil had
several projects in limbo waiting until he had the time, or an
assistant, to work on them. One of these was an ion oscillator. 
He had conceived the idea in Germany during the war, but never had
the opportunity to do any work on it until perhaps a year ago when
he built a prototype and got a bit of preliminary data from it
before putting it aside in favor of more urgent work. The ion
oscillator was intended to be a residual gas analyzer as an
alternative to the helium leak detector. This latter device worked
by detecting the ion current carried by a single pass of helium
ions through a magnetic focusing system. The Heil Ion Oscillator
was supposed to use a specially shaped electric field configuration
to force the ions trapped therein to oscillate for many cycles,
thus producing an enhanced signal per ion. 

The configuration was quite simple. Two pure tungsten hot
wire electron emitters, cathodes roughly 1 inch long, parallel to
each other, were separated by roughly 1 inch. Three pairs of
parallel electrodes the same length as the cathodes were placed
between the two cathodes. The central pair was operated at cathode
potential while the other 2 pairs, located equidistant from the
cathodes and adjacent thereto were operated at a positive
potential. A strong magnetic field directed from one cathode
toward the other confined the electron flow to a thin strip between
the three pairs of electrodes. Gas particles wandering into the
vicinity of the electron beam would become ionized and a large
fraction of the heavy positive ions would be attracted to the
negative charge of the confined electron beam and forced to
oscillate back and forth in the direction of the magnetic field. 
If the electrostatic potential formed by the three pairs of
electrodes and the confined electron beam was purely parabolic,
then all ions would oscillate at the same frequency regardless of
the amplitude of those oscillations. They would, sooner or later,
tend to bunch up and produce a coherent signal between the two
outer pairs of electrodes at anode potential. It was a very clever
scheme.

This signal could be detected using an ordinary short wave
radio receiver. Another trip to the warehouse produced one of
several such in storage there. Heil had made a few runs with the
experimental setup he had, so as soon as we could seal up the
vacuum system and establish a vacuum, he was able to show me the
signal made by the oscillating ions. He wasn't able to tell me
what species of ion was responsible for the signal, but he was able
to show me how to find out. This involved setting up the electrode
configuration in an electrolytic tank and making an experimental
plot of the electric fields in the tank. From this plot and the
fact that there was an exact analogy between the fields in the tank
and the fields in the vacuum experiment, we could make a rough
estimate of the oscillation frequencies in terms of the mass of the
ions and the applied voltages. 

After perhaps a week of work getting to understand the
electrolytic tank and taking data, I came to believe that the
potential along the ion path was not even approximately parabolic
except near the center pair of electrodes. If this was true, the
frequency of oscillation would be markedly dependent on the
amplitude of the oscillation and I couldn't see why one should
expect a coherent signal at all. Heil had some hand-waving theory
based on his intuition to the effect that the bunching of the heavy
ions would somehow lead to a coherent oscillation, such as the
signal we were, in fact, seeing. How to calculate the frequency of
that oscillation in terms of the ion mass was not quite clear to
him at this point, but he expected that he would come to understand
how it all worked in time. In the meanwhile, my job was to take
data and make calculations and see if I couldn't come up with an
explanation or, failing that, a restatement of the question. 

After another week or so of intensive work on the electrolytic
tank, finding the fields in the ion oscillator, and calculating
transit times for a variety of ions, I approached Heil with my
results. They did not meet his expectations. The device was
putting out a clear signal whose frequency varied with the applied
voltage in just the right way, but a calculation of the mass of the
ions involved did not make any sense. Heil watched over my
shoulder as I took him through the entire line of reasoning. He
could find no fault with any of it, but there was the physical
evidence which made no sense. Either my reasoning or his model was
wrong. His intuition and experience told him my reasoning was
wrong. I was more than willing to accept this conclusion, but I
would have to be shown where I had taken a wrong turn.

Heil approached Tony to make an Argon source for him. This
was a standard type device common in the neon sign trade. After
the sign maker bends tubing into the desired shape, he pulls a
rough vacuum and seals the work from the pump. He then uses an
external magnet to control an iron slug inside the tubing to break
a septum on an appendage. This releases a small amount of gas,
like argon, helium, or neon into the work. The spent appendage is
then sealed off and removed. When a high voltage is applied to
electrodes at the ends of the tubing, the gas is ionized and gives
off light in a color characteristic of the gas introduced. Tony
created the item from raw stock in a very short time. I had never
seen anything like this before. I was completely amazement and
more determined than ever to learn glass blowing. Heil attached it
to the ion oscillator vacuum system and, after the device was
delivering a healthy signal to our receiver, he broke the septum to
allowed some argon into the vacuum. The signal did not change,
much to his surprise, and he said he would have to think about it
for a while. 

I discussed these latest developments with Prof. Boone and he
said that the ion oscillator looked like an ideal thesis projects
to him. There were a lot of unanswered questions which should
yield to mature investigation and I would either make it work as
Heil had intended or find out why it didn't work. Either case
would make a fine thesis. This suited me just fine and I agreed to
go with it and give it everything I could come up with.

I was taking Prof. Boone's class in electromagnetic theory at
the time and he had recently proved a theorem which said roughly
that if we knew the electrostatic potential along any closed
contour, then we could find the potential everywhere. I approached
Boone and Heil with the proposition that since we knew that we
wanted a parabolic potential distribution along the axis of the ion
oscillator, we could in principle at least, find the fields
everywhere and come up with electrode shapes which would give us
what we wanted better than the array of parallel rods Heil had
started out with. Both of them wished me luck, but Heil thought
the problem was too tough and totally unnecessary. He had been
solving vacuum tube problems with intuitive approximate solutions
too long to believe that his electrode configurations were
inadequate to the task.

Over the next couple of months I worked on a number of
theoretical approaches and Heil and I conducted a number of
experiments with the apparatus we had, but neither of us could come
up with an explanation for what we were seeing. All this time,
Heil spent more and more time working with his gun and the general
problem of metalizing ceramics. He showed me an artifact he had
brought over from Germany which was a small Alumina (Al2O3)
crucible in which he had melted some Titanium in vacuum. The metal
had wet the ceramic and migrated up the sides of the crucible. 
There had to be some simple and convenient way to metalize
ceramics, he reasoned, to enable vacuum tube engineers to get away
from glass and the 400 DgC limit on bakeout temperature. It was
pretty clear to him that we could make the required brazes in a
vacuum furnace, but that was expensive and inconvenient. Hydrogen
atmosphere furnaces were much more practical, but active metals
like Titanium formed hydrides which inhibited wetting.

One idea he had was to evaporate a film of some refractory
metal like Tungsten or Molybdenum onto the surface of the ceramic
and then use the Heil gun to bombard the surface with the electron
beam using pulses to heat the ceramic to its softening point. The
metal and the ceramic should form a bond under such conditions. Of
course, this was a vacuum operation, but Heil thought that if we
could metalize the ceramics in a vacuum and braze them in a
hydrogen furnace, that might be a step ahead. I helped him set up
an experiment to test this theory. An Alumina cylinder roughly 1/2
inch in diameter was mounted on a screw which was driven through
the vacuum wall using a lubricated rubber gland. The surface of
the ceramic was set at the focal spot of the Heil gun. It was a
bit of trouble getting the pulsed power supply working and the
gland to hold a vacuum, but we finally got it all working long
enough to convince Heil that the principle was sound even though
the setup left a lot to be desired.

These results inspired him to construct a high temperature
metalizing furnace. This project was not greeted with enthusiasm
from any quarter for reasons I was unable to learn. I thought it
was a great idea, but I could tell that Heil was running up against
opposition almost everywhere. No one confided in me, probably
because I was seen as his partner in crime by this time. Perhaps
his detractors were of the opinion that he could go to the
marketplace and buy a furnace cheaper than building one himself. 
Perhaps they didn't think the tube lab should be in the business of
metalizing ceramics at all. In any case, Heil's charter from the
Air Force allowed him great latitude and he went ahead with his
design and ordered parts to be made by Pete in the vacuum tube lab
as well as the main machine shop in the EE department. The final
configuration was a spherical copper shell about 12 inches in
diameter, silver plated and polished on the inside, and wrapped
with copper tubing for water cooling. A thin walled Molybdenum
cylinder roughly 1.5 inches in diameter along the axis of the
sphere was the heating element. It was electrically insulated so
that it could be directly heated by passing a large current through
it. 

The Saturday Evening Post, a weekly magazine popular in those
days, carried a regular feature entitled Tales Mein Grossfader
Told, by Dave Morah, in which corrupted versions of familiar fairy
tales were re-told in pidgin German. I found them to be terribly
funny, as did several of the other graduate students, one of whom,
Gary, made it a point to read the latest offering to Heil. He
either did not get the point of the stories, or pretended not to
get it, but he always smiled politely. Sometimes he might comment
or ask a question in a way to suggest that he took the stories
seriously and didn't realize that he was being made fun of. When
we were working on the leak detector, Heil explained to all present
how the device worked and Gary made a crack about the Sugar and
Ants method for finding leaks. Heil replied in all seriousness
that he had never heard of this method and asked how it worked. 
Gary explained that you put some sugar inside the vacuum and
watched to see where the ants went in to get it. Heil still didn't
seem to realize the joke and asked if people really used the
method. Of course, Gary answered, and for even bigger leaks there
was the Cat and Mouse method. Heil still didn't seem to get it,
but I digress. One time the subject of Tales Mein Grossfader Told
was Cinderella Hasenpfeffer including the line, "Achtung. 
Suddender das clocker bin up-sneaken mit ge-striken der midden-
nighten". 

Heil's association with the Ohio State tube laboratory came to
an abrupt end over an incident involving his metalizing furnace. I
came to work early one morning to find an inch of water covering
the floor in my lab, a couple of doors from Heil's lab. The campus
radio station studio was in an adjacent room and the people there
had recently arrived to find their quarters flooded as well. Pete
and Tony had arrived a short while earlier to find that a flexible
hose formerly connected to the cooling coils on Heil's furnace had
come off and was whipping about spraying water all over the place. 
This had apparently been going on for some time, perhaps a few
hours. They shut off the water and began cleaning up. By the time
Heil arrived on the scene, most of the water had been swept out,
but the floor was still wet and there was havoc everywhere. Heil
was stunned and asked what had happened. Gary was on and hand and
explained, "Achtung. Suddender das pressure bin up-sneaken mit ge-
poppen der hosen". Even then I don't think Heil quite got it.

So Heil moved back to Wright Field at Dayton after this
fiasco. The Air Force still had some arrangement with him and
provided a large truck to help him move. I offered to help and was
put in charge of driving the truck. An enlisted man checked me out
on the many forward and reverse gears and other features of the
truck. I had met Heil's wife earlier when Sue and Ellen put on a
going-away party for Dick Walter who had just got his Master's
degree and was headed west to Varian Associates in California. 
Heil's first wife had somehow been left behind in Germany and I
gathered that this was a fringe benefit so far as he was concerned. 
In any case, his new wife was quite young and charming. I ran into
him years later at a conference somewhere, and we visited for a
little while going over old times. I am under the impression that
he was working for Litton at the time, but I couldn't be sure. In
any case, I lost touch with him after that.

Toward the end of the Spring quarter I had made numerous
calculations and measurements with the ion oscillator and had come
to suspect that the signal we were seeing was due to ions
originating on the positive electrodes themselves. I knew the
shape of the electric fields between these electrodes reasonably
well and was able to calculate the transit times for ions
originating anywhere in the region. The consequent frequencies
were far spread and I saw no reason to suspect that they would
somehow all get in step and produce a coherent signal. On the
other hand, if the positive electrodes were the source of the ions,
I could imagine that they might oscillate in step for many cycles
before drifting out of step. I explained my hypothesis to Prof.
Boone just before he took off for a summer of climbing about in the
Grand Tetons. He thought there might be some merit to it and asked
Jim Ebers to supervise my work until he got back for the fall
quarter.

Jim was working on a millimeter wave klystron project and
taught a course on klystron theory which I had taken from him, but
he had little knowledge of what I was doing up to this time. When
I explained the project and what I had done so far, he caught on to
the subtleties right away. At first he suggested that I calculate
the transit time through the existing structure, assuming that the
positive electrodes were the source, in terms of the ion mass and
try to find out what ions we were looking at. After a week at the
slide rule doing very tedious calculations, I told him that if I
had to guess, I would most likely pick chlorine. He suggested that
we take the bull by the horns and pour a lot of power into the
anodes, getting them red hot to drive off any chlorine, and then
looking to see if we had got rid of the signal. This was done and
the signal was markedly weaker, though not gone entirely. 
Repeating this process several times made the signal go away or at
least below the noise floor of the receiver. I was about ready to
write my theses, but Jim suggested that we construct an experiment
to highlight the principle. He suggested an ion oscillator in
which the ion sources were parallel plates which could be heated
individually using Heil gun heaters, which we had in surplus. A
cathode consisting of several tungsten wires in parallel midway
between the two plates should be used as the electron source for
making ions. This experiment was quickly set up and we immersed
the plates in salt water for starters. The transit time was easy
to calculate almost exactly and it was apparent from the start that
we had large signals from both sodium ions and chlorine ions. We
were never able to eliminate either element from the plates below
clearly detectable levels although we were able to heat them to a
dull red color. My thesis was ready for Prof. Boone when he
returned from mountain climbing. I was granted the MSc in EE and
admitted into the PhD program.

After Heil left, I started spending more and more time with
Tony and Pete learning to blow glass and operate lathes and milling
machines. Ebers and Boone tried hard to discourage my investment
in these skills on the theory that industrial employers were not
going to be happy paying me research engineer's wages to do menial
labor work. They were, however, unable to dampen my enthusiasm for
these basic technologies. I left the PhD program to get married
after several quarters and took an offer to work at Bell Telephone
Laboratories. I was told that the academic freedom there was
absolute, even moreso than at most universities, and that I could
work on anything that interested me. This was pretty much the
case, but for most of the four years I spent at Bell, I felt like a
pair of brown shoes at the Tuxedo Ball.

Before I left Ohio State, Tony quit the labs to devote full
time to an enterprise he had been working on in his off hours. He
had been buying electron guns and similar assemblies direct from
RCA and using them to rebuild television picture tubes. He was
also able to provide some moonlight employment to Sue and Ellen. I
heard later that he had rented industrial space and bought new
equipment and was on his way to making a superior living, if not
yet a fortune.- RMR