don’t hear a lot about homebrewing radio gear these days; it’s
just too easy to call a salesperson or go online and find what you
James Rockwell, right, and Clyde Haehnle. The occasion was Rockwell’s
1965 retirement party.
doubt anyone would consider homebrewing a broadcast transmitter, not
even a pre-sunrise AM rig that could output only a few watts. If
someone told you that he’d decided to roll his own 50 kW rig, you’d
probably laugh him all the way to the door.
there was someone who once did just that. Not only did it work, it
was touted as delivering the “highest fidelity radio transmission
in the nation.”
someone was Ronald James Rockwell — better known as R.J. Rockwell —
who reigned for many years as the vice president of engineering at
why did Rockwell decide to take it upon himself to create a 50 kW
one knows for certain; however, it would not have been due to lack of
funding. The Crosley Company (Avco by then) certainly had plenty when
it was time to think about replacing the aging Western Electric 50 kW
transmitter that had been in operation since the late 1920s at WLW.
(Even in the station’s “superpower” phase, 1934–1939, when
the WE rig wasn’t connected to an antenna, it served as the driver
for the big RCA 500 kW “amp.”)
to Clyde Haehnle, who had been with the station since the early
1940s, and was a senior engineer at WLW then, it was believed that
Rockwell wasn’t happy with what he found in the way of commercial
wanted a transmitter with very low distortion,” said Haehnle. “None
of the available models came close to meeting his requirements, so he
set about to design and construct something that could. One of the
things he wanted to accomplish was to eliminate the big plate
modulation transformer, as he saw this as a source of distortion.”
one-of-a-kind WLW Cathanode transmitter.
(Click to Enlarge)
remembers that Rockwell carried such a high profile within the WLW
organization, there were no issues in convincing upper management to
let him take on the job of designing and constructing a radically new
type of 50 kW transmitter.
had an absolute carte blanche to build the transmitter. Rockwell
never had any trouble on selling projects to management.”
DID IT START?
more than half a century, it’s not certain exactly how long
Rockwell took to arrive at a totally new approach to amplitude
modulation. The idea may have its roots in the ’30s when WLW was
still running half a megawatt.
of a Rockwell patent application from mid-July of 1939 — little
more than four months after a court decision throttled the station
back to a mere 50 kW — provides evidence that he was thinking about
new modulation methods.
patent (granted in 1941) described a method for simplifying a
transmitter’s modulation system, and began with description of the
Class B (Barton) modulator hardware needed for the 500 kW transmitter
— eight 100 kW triodes.1
devised a way to reduce the tube count to four by shifting the
alternate (“bottom half”) cycles, which would normally be handled
by the lower half of a Class B circuit, back to the same tube that
amplified the “top half.” While this sounds a bit like
single-ended Class A operation (and Rockwell’s circuit used only
one high-power triode), it wasn’t. He employed a couple of small
triodes and two rectifiers as a means of “switching” alternate
half cycles to the big tube (which was biased for Class B). It’s
not known if the circuit was ever tried on the air, but it’s
doubtful, as by that time, WLW was not making much use of the 500 kW
machine due to politics and the FCC.
cover sheet of the 1956 application to the FCC for type acceptance of
the Cathanode rig.
1946 Rockwell patent application describes a somewhat more advanced
scheme for attaining Class B modulation, with elimination of the
conventional modulation transformer as its goal. Rockwell claimed his
new circuit provided 20 Hz to 50 kHz bandpass with less than 1
percent distortion, and very low noise figure. The circuitry in this
patent begins to bear some resemblance to the finalized Cathanode
circuit disclosed some 10 years later. Curiously, while many patents
are written in such a way as to conceal critical details of an
invention, Rockwell provided a “parts list” with tube types and
resistor and capacitor values, almost encouraging others to build it.
1949, Rockwell applied for another patent that has additional bearing
on the Cathanode circuit. Its title was “Push-Pull Wide Band
Amplifier,” and Rockwell cited his earlier “Modulation System”
patent as the basis for the developmental work leading up to this
invention. Again, a complete parts list was included.
followed this with another patent application in July 1953 for a
“High Fidelity Amplifier.” In its preamble, he cites
developmental features set forth in earlier patents leading to this
application, which describes a “double-ended high fidelity
transformerless amplifier now
known as the Cathanode amplifier” [Rockwell’s emphasis
here and below].
circuit in this patent application (it was granted Sept. 18, 1956) is
virtually the same as the Cathanode circuit used in his broadcast
transmitter. Rockwell, in the application, noted: “It is a still
further object of this invention to provide a very high fidelity
amplifier in a relatively small number of stages, which
is especially suitable for use as a Class B modulator.”
THEORY INTO PRACTICE
the 1953 patent application indicates Rockwell was on the road to a
new form of modulation, there are no surviving records indicating
when he presented the idea to WLW’s corporate chieftains, or when
he ordered parts for the new rig.
the documents that do survive is an engineering report prepared for
FCC type acceptance of the transmitter, and which completely
describes the new machine. The report logically would have been
prepared after the project was successfully completed — it’s
dated Sept. 6, 1956. Curiously, Rockwell didn’t get around to
disclosing to the broadcast engineering world the workings of his
radically different transmitter for more than a decade, publishing a
paper in the IEEE’s January 1967 edition of “Transactions on
the IEEE document, Rockwell describes a “scaling up” exercise as
part of the path to the final transmitter design. This was obviously
done sometime prior to construction of the rig.
order to facilitate investigation of the Cathanode circuit prior to
construction [of the 50 kW transmitter],
a low-power modulator stage was set up using Western Electric 300B
tubes since they were close in characteristics to the Federal Type
6921 tubes to be used in the transmitter.”2
recalled the actual design and testing of the new rig.
guess that it took about a year from the time the project started
until the transmitter was on the air,” he said. “Once it was
ready, we did a hell of a lot of testing before putting it on the
air. We tested it for months and were very confident when we did the
circuit drawing from one of Rockwell’s many patents. The striking
symmetry portrayed in this diagram is evident in most of Rockwell
schematics. It typifies the inventor’s love for symmetry in
everything he constructed.
(Click to Enlarge)
of the surviving station employees remembers the exact date when the
Cathanode rig was placed in service, but Dick Perry, who captured
many of the WLW goings-on in his 1971 book, “Not Just a Sound: the
Story of WLW,” pegged it sometime in January of 1959. Perry noted
that the new transmitter was part of a $300,000 program to rebuild
the station’s transmitting facility, stating:
these engineering changes the station management began calling WLW
‘The Nation’s Highest Fidelity Station.’ The WLW transmission
was tested by the McIntosh Laboratory, Binghamton, New York.
According to the report by Frank McIntosh, president, the WLW signal
ranged from seventeen to 21,500 cycles per second — more than ten
octaves — with a distortion of 0.3 percent. Mr. McIntosh stated
that: ‘it should be recognized that while FM is capable of this
same order of fidelity, many stations have not achieved it because of
limitations in microphones, preamplifiers, circuits, and program
1959 startup date is confirmed in a Billboard magazine article that
pegs it precisely as Jan. 19.4
NEW FORM OF MODULATION
concept of a “plate-modulated” transmitter constructed without a
modulation transformer takes a little getting used to, as what’s
going on in the circuit is not really that obvious on first reading.
However, Rockwell’s development of unconventional circuitry was
had a brain that wouldn’t stop,” Haehnle recalled. “He could
out-think what a vacuum tube could do. He was brilliant. He could do
things with circuits that you wouldn’t believe. He was a real
bulldog — he just wouldn’t give up.”
summed up the transmitter’s modulation system:
modulation system … involves a process in which audio voltage is
obtained at any instant from both the anode and the cathode of a
modulator tube. This voltage is either added to, or subtracted from,
the split high voltage rectifier, and the resultant is applied to the
anode and cathode of the final RF Class C amplifier as modulated DC —
with no modulation or coupling transformer being involved.” 5
only did Rockwell successfully eliminate the modulation transformer,
he made sure that the audio had a very clean path.
coupling has been utilized between the [input audio
amplifier tube] and the [second
audio amplifier], and between the
[third and fourth audio amplifiers],
thus leaving only two coupling condensers in the entire five stage
audio system, thereby improving low frequency response and greatly
simplifying low frequency feed-back circuitry.”6
actual modulator stage consisted of two previously mentioned
water-cooled Federal F-9028 triodes with 13 kV on their plates. These
were biased “practically to plate current cut-off,” allowing the
stage to operate in Class B mode. In a general description of the
modulation scheme, Rockwell stated that each tube functioned
independently in developing a 5 kV peak potential at its anode and at
its cathode, with the two excursions amounting to 10 kV. This
potential was applied to the plate and cathode of the Class C RF
finals via two 5 kV rectifiers. These rectifiers were in a
series-adding configuration and thus supplied 10 kV to the RF
provided a precise description of the modulation process:
modulator tube adds its 10,000 volt peak output to the 10,000 volt
Class C rectifier, thus providing 20,000 volt positive peak modulated
DC as cathanode modulation … to the 50 kW RF amplifier; the other
modulator tube subtracts its 10,000 volt peak output from the 10,000
volt Class C rectifier, thereby providing negative peak modulation
voltage to the final amplifier.”7
added that his circuit also reduced the possibility of “flashover”
to ground in the RF output stage, as voltages were only half of what
would be needed for modulation in a conventional circuit with the RF
amplifier’s cathode at ground potential.
Cathanode terminology for describing his new type of modulation is
indeed descriptive, but the term did not originate with him.
“Cathanode” appears in at least one Bell Labs patent from the
1940s, and seems to have its roots in a 1932 patent application made
by Raytheon’s James D. LeVan. However, in both patents, “cathanode”
is used to describe an element in a cold-cathode “discharge”
to Haehnle, and another former WLW engineer, Charlie Stinger, the new
transmitter was almost completely fabricated in house. Even the sheet
metal for the transmitter cubicles was formed and painted by WLW
personnel, along with the winding of power supply transformers.
Rockwell obtained some power line transformers from Cincinnati Gas
and Electric,” Stinger recalled. “We took these apart and
completely rewound them to his specifications. One of the things that
he was particular about was the wire — it had to be square, not
round. I believe that it took a while to find that, but he did, and
we rebuilt the transformers to his specifications.” Stinger
suspects Rockwell specified square wire so that it could be wound
tighter and had less chance of moving in the magnetic fields
generated in the transformers. Haehnle suggested it was used to allow
more copper conductor to be placed in a given space.
the rig’s power supply was unconventional. Mercury vapor rectifiers
were the norm then, but Rockwell insisted on using Cooper Hewett
(mercury arc) rectifiers, a technology that hadn’t been used in
transmitters since the 1920s. It’s not sure why he took this
approach, but it may have been an attempt to save money on the
heavily insulated filament transformers that would be needed for a
conventional three-phase mercury vapor power supply. Alternatively,
he might have wanted to use tubes without filaments to improve
recalled the English-made mercury arc rectifiers were very reliable.
(They carried a maximum rating of 270 kW at 15 kV.)
don’t remember that we ever had a problem with them,” said
Haehnle. “In fact, we had less trouble with arc-backs with them
than we did with the conventional mercury vapor tubes used in the big
500 kW transmitter.”
Rockwell’s painstaking efforts to develop a “perfect”
modulation scheme, some problems did arise. However, these were not
directly related to the audio performance.
recalled by Haehnle, one issue involved some “interferences” with
patents held by audio legend Frank McIntosh. Haehnle stated that an
amicable settlement between McIntosh and Rockwell was reached fairly
quickly, with McIntosh journeying to Ohio to inspect the new
transmitter and measure its performance.
front panel of R.J. Rockwell’s “home-brew” ham radio
transmitter. Notice the symmetrical placement of its meters, dials
and knobs. The unit is on display at the Gray History of Wireless
Museum in Bethany, Ohio.
(Click to Enlarge)
Cathanode transmitter also incurred an FCC citation during an
inspection by John Reiser, who worked out of the commission’s
Detroit office then.
issued a notice of violation because the transmitter’s antenna
current meter didn’t conform to current FCC standards,” Reiser
said. “It didn’t have the required number of scale divisions.”
recalled some other problems that could have brought an additional
were parts on the inside supported by string or cords. They weren’t
really permanently mounted, just suspended in free space. This sort
of surprised me. I could have cited them for violation of good
the biggest problem stemmed from Rockwell’s 1965 retirement. The
transmitter’s design was so unusual, it wasn’t that well
understood by those left to maintain it. Stinger remembers that the
rig required quite a bit of attention to keep it in proper operation,
and Haehnle observed that it was designed for fidelity, rather than
efficiency, with a resulting large power bill.
a few years of Rockwell’s departure, the Cathanode rig was gutted
and rebuilt with conventional high-level Class B (Barton) plate
modulation. It backed up the venerable 1920s Western Electric rig
until 1975, when a Continental 317C1 became the station’s main
transmitter. The “converted” Cathanode transmitter is still a
part of the WLW transmitter facility.
QUESTION OF BALANCE
that knew him say that Rockwell loved symmetry — maybe was even
slightly obsessed by it. His circuit drawings reflect this, as
components appear to be in perfect balance. Rockwell’s homebrew ham
radio transmitter is also a good indicator of his quest for a
balanced and symmetrical world. The left-side front panel meters,
knobs and switches perfectly complement those on the right side. This
quest for symmetry may explain Rockwell’s development of the
Cathanode modulation scheme, which exhibits a high degree of balance
author expresses appreciation to the individuals who assisted in
preparation of this story including Clyde Haehnle and Charlie
Stinger, former WLW employees who worked under the direction of R.J.
Rockwell; John Reiser; and Jay Adrick of Harris Corp.
O’Neal is technology editor of RW sister publication TV Technology.
U.S. Patent 2,233,961 — “Amplifying System and Process” granted
to James Rockwell on March 4, 1941 and assigned to The Crosley Corp.
Rockwell, R.J., “Cathanode Modulation System.” IEEE Transactions
on Broadcasting, Jan. 1967, p. 22.
Perry, Dick; “Not Just a Sound: the Story of WLW,” Prentice-Hall,
Inc., 1971 p. 79, 80.
“Crosley Debs Hi-Fi System,” The Billboard, Jan. 24, 1959, p. 4.
Rockwell, R.J.; Alberts, William S.; Haehnle, Clyde G., “Engineering
Report: Application for Type Acceptance Rockwell Cathanode
Transmitter Type EDS-530,” Sept. 6, 1956, p. 3.
Ibid; p. 6.
Ibid; p. 7.