ORIGINAL B-29 RADIO OPERATOR'S POSITION-- photo notation courtesy Mike Hanz.
THE AMAZING ART-13
In 1940 a fledgling radio company in the Midwest, with fewer than 200 employees, participated in a US Navy design competition for a new transmitter, going head-to-head with the biggest radio manufacturers of the time. The story goes that in spite of winning the competition, the Navy, concerned with the small company's ability to execute on a major contract, was slow to sign a contract.
This quickly changed when a British procurement officer told Art Collins that the UK would buy as many of the new radios as his Collins Radio Company could produce. The transmitter, Navy designation "ATC", was used in every WWII Navy airplane with a crew of two or more. In 1942 the little company that at first had trouble selling its innovative radio, did fifty million dollars worth of business with the US Navy. The radio was later picked up by the Army Air Corps as the "ART-13", but too late to make it into most of the Army Air Corps planes. It was known to be installed in some B-24 bombers and is believed to have been installed in every B-29 bomber deployed to the Pacific Theater of War. The revolutionary new radio was the first transmitter ever designed to operate across most of the HF spectrum that did not require a dedicated radio operator.
Prior to US entering the WWII, the War Department had been injecting funds into strategic suppliers. A major addition to the new Collins “Main Plant” on 35th street in Cedar Rapids was built with such funds, and the company was still in the process of occupying the new building during the weekend of December 7, 1941. Ben Stearns has documented that Art Collins and the team were busy the weekend of December 6th and 7th preparing the radio for a meeting with the Navy in Washington that had been scheduled for December 8th. ART-13s were soon rolling off those Main Plant production lines. Eventually, more than 90,000 ART-13s were built by Collins and several other licensees and many of these radios saw military service into the 1960s.
The technical advances of the ART-13 were due to the an elegant marriage of superior electronic design coupled with precision advances in mechanical and electromechanical innovations.
An essential part was the Collins Autotune -- an electromechanical shaft positioning control device capable of precisely locating ten preset-able angular positions with the flick of a rotary switch.
Also key was the Collins Permeability Tuned Oscillator (PTO), a radio tuning device that brought stability, precision and linearity to a new standard.
The radio was nothing short of amazing considering the state-of-the-art at that time. Through the magic of technology called “Autotune” patented at Collins, ten frequencies could be easily pre-programmed and then selected in-flight without the aid of a radio operator. This gave naval pilots, including those flying torpedo bombers, a considerable advantage over previous radios utilizing a single hard-wired frequency that were well known to the enemy.
Eventually the ART-13 was picked up by the Army Air Corps, and found its way into virtually every B-29 Bomber. Matched with the venerable BC-348 receiver, the pair was used as the "liaison" (or long distance) radio set. The HF transmitter used in earlier bombers was the BC-375, a much taller unit requiring a cumbersome manual operation for frequency changes.
The five autotunes of the ART-13 shown from below. The single transversely mounted motor in the middle drives all five autotune mechanisms.
The PTO (Permeability Tuned Oscillator) is on far-right autotune, and makes 20 full revolutions from stop to stop.
The ART-13 was the first Collins radio to use the PTO, but not the first to use the auto tuning method.
upper shelf left to right: COMMAND SET BC-458/459 BC-454/453/455
lower shelf left to right: LIASON SET COLLINS ART-13 TRANSMITTER, B-348 RECEIVER
ART COLLINS TALKS AUTOTUNE
First Person Radio History
This speech was written and delivered by Art Collins at a meeting of the Collins Radio Technical Association (CRTA) meeting on May 2, 1944.
Since Ted asked me to tell this group something about the early history of the Autotune I have
been having quite a bit of fun making up some notes covering my recollections of the past six or
seven years dealing with this subject. I will not attempt to give you dates and I may have my
facts out of chronological order.
Our first multi-frequency transmitters made use of separate r-f sections complete with tubes and
circuits for each channel and employed a common power supply and common audio systems.
We made up several of these sets providing operation on three or four frequencies. Each
channel had a power of about 100 Watts and the construction was of the relay rack type. We
sold these sets to Braniff and American Airways, to the Standard Oil Company of Venezuela
and to a few other customers. They were very practical and satisfactory sets and a number of
them are still in use today after eight or nine years service. At that time Western Electric and
RCA placed on the market some multi-frequency transmitters in which they used common r-f
tubes but switched different sets of tuned circuits in order to change frequencies. RCA built
some four channel sets and Western Electric built a ten channel set. These transmitters were
only moderately successful. RCA and Western Electric at that time did not have very much
competition in the high frequency transmitter field and were inclined to be a little bit careless as
a result in executing their designs. As soon as the airlines began to extend their operations in
this country they found that they had a real need for some really reliable multi-frequency
transmitters with as many as ten channels. One day after I had been visiting Frank Dyer in
Dallas and while I was flying back home, I decided that a new approach to this problem was
needed and that if somebody could build an extremely reliable set there would be a large
market for it. The idea of building a continuously tunable transmitter and shifting frequency by
actually mechanically tuning the circuit appealed to me and on my way back I made some
mental notes as to how the job could be done.
I have asked several of my friends who are responsible for various inventions and new ideas as
to what particular environment they have found to be most conducive to original thought. I have
noticed that in thinking about problems and attempting to find a new solution for them that the
answer usually comes to me while I am either flying or riding along somewhere. Other people
tell me that they get ideas while shaving and in various other particularly congenial
surroundings. So far I have never had anyone tell me that they were able to get an intelligent
idea on a golf course. I personally am still experimenting with this procedure, however, results
have been entirely negligible but I haven’t given up hope.
When the decision was made to go ahead with the mechanical tuning idea it seemed apparent
that it would be necessary to really do a good job of building a precise and reliable mechanism
and that the amount of work involved would be such that the design should be more or less
universally applicable to any kind of a transmitter. At that time the Collins Radio Company was
very poorly equipped in its facilities and experience to undertake a program of this kind. The
only machine tool which we possessed was a broken-down 16” South Bend Lathe. This lateh is
still in our shop having been rebuilt three or four times and I believe it is referred to
disparagingly as the “old gun lathe”. It never did qualify as an accurate tool. It was obvious,
therefore, that we needed some new equipment and probably some new talent before we could
get very far with this new project. I went over to Ohio State and hired Pete Morison, a little
Scotchman, then an assistant professor in the Mechanical Engineering Department. Pete
looked to me like the sort of fellow who could do a nice job of building gadgets. He talked one
of his students, another Scotchman, by the name of Bob Cochran, into coming over to Cedar
Rapids with him, and then I sent Pete to Chicago to really splurge himself buying some machine
tools. Pete spent one thousand dollars for a secondhand Brown & Sharpe Vertical Milling
Machine. This mill was in reasonably good condition after Pete got through lining it up. He
spent the other $500 we possessed at that time for tools, cutters, collets, etc., and for a South
Bend work shop lathe. This equipment was set up in the basement of the building at 621 First
Avenue which is not the Troy Laundry and we were all ready to go into the Autotune business.
Poor Bob Cochran went back to Ohio State for an operation and just as he was ready to leave
the hospital he died from an embolism, so we had to carry on without his help.
Although I had some reasonably concise ideas as to how an Autotune unit should be built, at no
time did I have very much to do with the actual engineering of any of the models of the
Autotune. I found that there have always been several of you people who have been in a
position to do a very good job of engineering this type of equipment. Bob Samuelson, who is
now Chief Engineer for Hallicrafters, did quite a but of the detail layout work on the first models
and he was assisted by Don Holmes in laying out the relay control circuits. Bob Davis, John
Giacoletto, Roy Olson and several others of you have had a great deal to do with engineering of
The first model of the Autotune looked quite a bit different from the present models which you
are familiar with. It used stop rings clamped in somewhat the same way as the present model
but instead of using pawls it employed a lead screw which moved a dog up and down parallel to
the main shaft into position to engage one of the ten stop rings. The lead screws were driven
separately by one motor and the tuning shafts were driven by another motor though a disc
clutch. This version of the Autotune had several things wrong with it which we were not able to
overcome in that form. First of all the motor driving the lead screw had to stop very accurately
in order that the dog would engage a single stop ring properly. We had a lot of difficulty getting
a motor and control system which would position itself without over-shooting or hunting.
Numerous schemes were tried to reduce this trouble and we finally got fairly good results. The
other trouble with the first model of the Autotune was the disc clutch which was subject to
considerable variation in torque depending upon condition of the clutch surfaces. We never did
get those disc clutches working very satisfactorily, but we soon got around that problem by
substituting overrunning coil springs for clutches. Quite a few transmitters were built using the
lead screw arrangement and coil springs in place of clutches. The most successful models
were the 17D’s which were 100 Watt aircraft transmitters built for Braniff Airways. The 17D’s
proved after going through the headache stage to be quite reliable sets. They have been used
successfully by Braniff up until the last year or so. The 17D’s were the first high powered set
used in an airplane having a power something over 100 Watts and they were the first tenchannel
aircraft transmitters. I believe they were also one of the first transmitters to receive
CAA type approval. The 17D was not particularly dainty, it weighed 60 to 70 pounds and was
built in a rather crude but substantial fashion. One of the 17D’s went through Braniff’s only bad
crack-up and although the airplane was completely washed out and the 17D was thrown several
hundred feet clear of the wreck, it was found to be in operating condition after a broken tube
In thinking back over the history of the Autotune, I have been touched by a feeling of
appreciation of the part played by our customers in fostering this development. Some of our
customers played the part of sponsor consciously, others couldn’t help themselves. Old Bill
Beakes of Tropical Radio Telegraph bought the first Autotune set along with a group of other
transmitters they were buying. I am sure that Bill realized that the damn thing wouldn’t work and
he more or less wrote off its price, some $1500 or $2000, as a contribution to the advance of
science. Frank Dyer of Braniff recently made the remark to me that he felt sure that he had
bought the first model of every piece of equipment we had ever built. I think that statement is
almost literally true. Frank is an incurable optimist and he also has a knack of making things
work. I would say that Braniff had made a very real contribution to the development of Autotune
equipment. They were placing some orders at a time when orders were urgently needed in
order to keep going. Jim Flynn of American Airlines was another hardy pioneer. After we
began to get the bugs out of the Autotune he came along with an order for $100,000 worth of
aircraft transmitters which were to be of the new design which was still just a glint in our eye.
$100,000 in those days was a lot of money. Roy Olson built this set for Jim and it became our
17F which was really a good set in that it embodied a type of Autotune which is substantially the
same as our present model.
I am getting ahead of my story a little bit. The turning point in the Autotune history came with
the set we sold to TWA. Some years ago TWA seemed to be long on boards of directors and
creditors and short on imagination. I don’t know how they managed to screw up their nerve to
buy anything so questionable as an Autotune set, but Howard Morgan and Jack Franklin finally
took the leap and ordered a ten-channel ground station equipment to be installed in Kansas City
on an experimental basis. This set was to be complete with all sorts of bells and whistles and
automatic gadgets which it was decided were essential to airline ground station equipment
which would be standard throughout their system. At this stage of the game we were still
building the two motor lead screw type of Autotune. We had, however, graduated from the
friction clutches to the over-running springs. I think everything that possibly could go wrong with
a piece of radio equipment went wrong with that set in Kansas City. I have always had a
suspicion that one of the troubles was that the TWA transmitter received the helpful attention of
two many engineers and not enough Ford mechanics. In any event, after the TWA set had
blown up for about the fourth time, Myron Smith, Frank Davis, and myself were driving down to
Kansas City in the middle of the night and I decided while we were sailing along through the
Missouri brush that there ought to be some easier way of doing the Autotune job than using two
motors, which had to stay out of each other’s way. Myron and Frank and I talked the problem
over and after we got back to Cedar Rapids we started the gang out to build the type of
Autotune which has proved to be most successful. A single motor was used to rotate the tuning
shaft and also to rotate a cam actuating pawls to engage the stop rings. Another feature which
was introduced at this time was the type of clutch we are now employing which has a torque
which is substantially independent of changes in the condition of the clutch surfaces.
The Autotune has proved to be a very important factor in the development of our Company. At
various times I have tried to analyze as accurately as possible why this is so and I have come to
this conclusion: As an invention the Autotune is not an outstandingly brilliant idea; in fact it is a
very garden variety sort of idea. It makes use of more or less obvious means of performing its
function. I don’t think anyone could truthfully describe it as being the result of any “flash of
genius”. There have been too many headaches connected with it to fall in that category if there
is such a thing. The important lesson to be learned from the Autotune, as I see it, is that either
through luck or good judgment we accurately defined the problem and we also set a sufficiently
high goal to shoot for in respect to standard of performance and reliability. There have been
other tuning devices, several of which are really more ingenious than the Autotune. Some of
the push button tuning mechanisms on car radios for instance are really extremely cleaver.
However, they have all been developed for a specific purpose and in no case has the designer
undertaken to find a full answer to the whole problem of precise tuning. What we set out to do
when we undertook the Autotune project was to develop an automatic tuning system which
would be to radio communications what automatic dial switching is to telephone
communications. It was a type of mechanism which is ultimately capable of extreme precision
and we insisted on building the units to the best possible standard of design and workmanship.
I think if we had been satisfied with designing and building a simpler and perhaps smarter
device which would have been suitable for a particular application but which would not have
been capable of use on a transmitter requiring a large number of tuning controls, even more
than was needed on some of our earlier sets, and also which was capable of even greater
accuracy than we had need for at the time, we would not have established the reputation we
have in this particular field. Therefore, I believe that if we are to take a lesson from our
experience with the Autotune in setting future Company policies, and also possibly in serving as
a guide to individuals who are working on their particular projects, I think we can find support for
the principle of employing methods which are capable of giving performance greater than is
necessary for the needs of the moment. To illustrate this point in relation to the Autotune, I
might refer to Ted Hunter’s new oscillators. The idea of tuning high stability oscillators was only
a vague possibility at the time the Autotune work was started because we didn’t have any very
stable oscillators at that time. However, the Autotune arrangement chosen was capable of very
close accuracy and now that Ted has come along with some really good oscillators we have a
tuning mechanism to serve as a complementary and necessary device capable of giving an
overall system with substantially crystal standards of frequency resetability.
Some of the recent developments in respect to our Autotune patent situation may be of interest
to you. We have, of course, applied for and received patents on all the various versions of the
Autotune and now have a considerable number of applications pending on some of the recent
improvements. We also have made applications on numerous alternative schemes which we
did not intend to build but which we wished to have covered in order to afford more complete
patent protection for our work. One of the early Autotune sets we sold went to the French
affiliate of I. T. & T. Somehow this set found its way into Holland, and apparently started the
Philips Company of Eindhoven working along similar lines. Before the war some of the Philips
publications showed illustrations of experimental transmitters with quick shift devices which
appeared to be copied from this set. Philips filed patent applications on their work and some of
their claims interfered with claims in our patents. These applications were filed in this country
and RCA through their exchange of patent rights under their agreement with Philips had a
certain interest in these patent applications. Last year RCA brought an interference action
against our patents under the Philips applications. After several preliminary hearings and
extended discussions with RCA we were able to show RCA that our dates were substantially
prior to the dates which might be established due to Philips’ work, and RCA gave us a
concession of priority. This concession on the part of RCA and Philips substantially enhances
our patent position so that at the present time we have no serious challenge to our Autotune
patents. Of course, we may expect increased activity in this field and if we are to continue in the
lead, we will have to continue our present policy of intensive development work.
The most recent history of the Autotune as of the last three or four years is no doubt familiar to
most of you. Although the airlines were the first important customers, military applications were
kept in mind from the beginning. The Army was not very receptive to the idea at first, although
they later became quite interested. The Navy, however, took the lead in sponsoring the
application of the Autotune quick shift idea to military equipment. A couple of Navy pilots, Roy
Jackson and Ham Dow, who were radio amateurs on the side and had been doing a good deal
of flying on commercial airlines and had seen commercial airline equipment, landed in Cedar
Rapids on two or three occasions and looked over the equipment we had been building for the
airlines. Apparently they were in a position to make some recommendations to the Bureau of
Ships because when we approached the Navy Design Section with a proposal to build some
quick shift ground stations transmitters for the Navy bases we had little difficulty in enlisting the
Navy support and we were granted several substantial contracts, in spite of the fact that at the
time the Navy was severely limited as to the amount of money it could spend for
communications equipment. Later on we undertook a development program directed towards
producing an Autotune aircraft transmitter for scout and patrol planes. This development job
eventually grew to be the ATC. The Army has followed the lead of the Navy and has now
adopted the ATC to completely replace the existing liaison sets. The eventual combined
requirements of the Army and Navy for Autotune ATC’s are so large that we have granted to the
Navy Department a free license for the duration under our patents to permit the Army and Navy
to have additional quantities of equipment made by other manufacturers to be an exact copy of
Since Dick May has joined our organization he has been given general responsibility for
Autotune development and he has set himself down to make more or less of a career of this
work – at least until such time as he can devote his effort to some even more promising project.
Dick has combined a large amount of ingenuity with the careful and scientific approach and has
been getting results far beyond anything I imagined when I started out on this work years ago.
In fact, I find myself considerably at a loss every time I drop into Dick’s lab in trying to keep
abreast of the fine points of his work. Dick’s latest Autotune units are much smaller, more easily
adjusted, more accurate and more reliable than anything we have had before now.
At this point I would like to take a back seat and let Dick do something which would be very
difficult for me to do – explain how the Autotune unit works.
Arthur Collins, founder of Collins Radio
We thought warbird gear heads might be interested in the results of Dave Graham's design and prototype build of the AM Filter module for the BC-348 Receiver. Dave has put a lot of work into this design and the results show. You will notice a custom designed "octal" plug at the bottom of the module. This resulted from mechanical problems using the conventional octal plugs. Their mounting forced the lateral size too large to fit in the space in the receiver.
The design features include:
1 - Uses standard Collins 455 kHz AM mechanical filters to provide a nominal 6 kHz of IF selectivity (compared to 16 kHz for BC-348)
2 - A "No holes - No soldering" modification. Simply remove the 1st IF tube and plug in the new module
3 - All solid state design; converts the 915 kHz IF down to 455 kHz then filters and converts back to 915 kHz; utilizes the radio's AGC buss
4 - Will also work with the "non-Q" or "grid cap tube" versions of the
BC-348 with a minor wiring change to the module to bring in the grid signal
1 - I have the radio for some additional test and evaluation. I want to run it with the switching PS and the AM module at the same time.
2 - Dave is going to build up a second module that we will to deliver to Dallas for installation on FIFI's receiver.
3 - Working on a second "spin" to the design - tweaks are:
a) Using a dual frequency programmable synthesizer chip instead of the present crystal oscillator (see explanation below)
b) Shorten the module height slightly (we almost line-to-line with the case which sometimes drags on the module)
We ordered three LO Crystals from ICM in Oklahoma City a few months ago for about $35/ea. Dave did a spurious analysis and found that we could improve things by using the Collins AM filters from the 51S-1 receiver as they are on a 500 kHz IF and same package as the 455 kHz filters. The design would accommodate either filter by simply changing the LO crystal. When we went to order the new LO crystals, ICM informed us that they would be $70/ea as they made a mistake in not charging us correctly for the first three. So we went on the hunt for a more reasonable crystal supplier - but found none so far.
But the little synthesizer chip is only $3.50 and allows either of two
frequencies by adding/removing a jumper. We will need to re-spin the PCB through our vendor in China to use this chip.
Rod Blocksome, KØDAS