One of my favorite pieces of transistor memorabilia is this Lucite paperweight which encapsulates six of the most significant electronic components to be developed by the Bell Telephone Laboratories and later produced by the Western Electric Company. It appears to date from the mid-1950's. Of particular note is the point contact transistor which has holes on both sides for adjusting the actual contact "whiskers".
Above left is a close-up view of an original BTL "type A" transistor which clearly shows the point contacts touching the tiny piece of germanium inside. This cartridge style case was the very first design used to manufacture transistors. Above on the right are various examples of early 1950's Western Electric transistors. Note the wide variety of case styles and colors that were originally produced. As the transistor industry matured in the late 1950's and early 1960's most of this design diversity disappeared, making the resulting devices somewhat boring. Top row l to r: 2N21 point contact from 1955 (the lowest registered "2N" number), 2N110 point contact from 1956, 1698 "type A" developmental point contact unit (1952?), unknown paint dot coded junction developmental unit from 1953, 3N22 tetrode from 1957. Bottom row l to r: 2N23 point contact from 1954, 1729 "type A" point contact from 1953 (in a somewhat rare socket), unknown paint dot coded junction developmental unit. Click here to see my transistor manufacturers code page for more details on how to identify transistors. Below is an example of the attractive Western Electric packaging which contained a 2N110 point contact transistor manufactured in 1959. The transistor inside is identical in appearance to the one pictured in the group above (except that it has full length leads).
The device shown below is one of the few examples of a commercial application using Point Contact transistors. It is a type 149A amplifier module made by Western Electric for use in the Bell Telephone system. I am indebted to my good friend Terry Hosking for sharing one of these wonderful devices with me. Terry performed an "autopsy" on one of the modules and reports that whatever its purpose was, it wasn't to amplify voice signals. The manufacturers logo, WE, is visible in the left view. The transistor type (not visible) is 2A.
On the right is a view from the opposite side. The date of manufacture is 2-58 (February, 1958), and I believed this to be among the latest P-C transistors to be manufactured until I discovered the box of 2A's which is described later on this page. P-C's quickly became obsolete after the 1951 invention of the junction transistor and here we have a WECO "type A" cartridge style case (including the black plastic tubing) who ancestry dates all the way back to 1948. The Bell System obviously made their equipment to last and the fact that P-C equipment was still in service a decade after the invention of the transistor is very interesting to say the least. The entire device is mounted in an aluminum cylinder 3" tall and 1-1/4" in diameter. A aluminum cap slips over the transistor end which is shown in the photos. At the bottom of the housing (not shown) is a standard 8 pin octal tube socket.
It was the practice of Western Electric to donate obsolete transistors to various colleges so students could experiment with them. Shown above is a partial box of WECO 2A's which somehow survived over the years. The inner drawer had holes to accommodate 20 transistors. As can be seen from the hand writing, this box originally contained 20 devices. Apparently 10 were removed for use elsewhere. Of the remaining 10, only 6 were in the box when I acquired it. There are two date codes on the transistors, 12-61 (December, 1961) and 1-62 (January, 1962). These are the latest manufacturing date codes for cartridge style point contact transistors that I have seen so far. I assume the 6213 represents the date of the gift, the 13th week of 1962, or sometime in early April. As can be seen, the box was sealed with stickers on each end indicating that the transistors were a college gift from Western Electric.
Here is a closeup look at one of the 2A's. These were obviously intended
to be replacements for the devices which were used in the 149A amplifier
modules described previously.
Above is an early Bell Telephone Labs prototype of the historic M1752 junction transistor. This was the very first design of a junction transistor to ever be made. Research seems to indicate that the earliest M1752's had no external markings while later ones carried the typical BTL three paint dot (violet-green-red) identification scheme. All units had serial numbers in order to track individual performance characteristics as production techniques were "fine tuned". The earliest units had tags like the one shown above, while later units had the serial number hand painted on one side of the case. Below is the cover of the July, 1952 issue of Scientific American which contains an excellent article on junction transistors. Note the color coded M1752 among the pieces of germanium. The article predicts: "this tiny crystal is likely to have a profound effect on our technology and society, making possible developments in electronics which could never be achieved with vacuum tubes." This was a very bold statement at the time.
Below is a close up of the M1752 showing the beautiful hand-made texture of the case. The material was a gray plastic resin which was then painted black. The earliest junction transistors were made using molded plastics but it was soon discovered that this contributed to early device failure due to contamination by moisture. Subsequently, a change was made by most manufacturers to hermetically sealed metal cases.
One can only appreciate the true appearance of these prototype devices by viewing them under high magnification since they are very small.
Above left is a WECO 1740 germanium junction photocell. This device is described in the 1956 book "Transistors Handbook" by William Bevitt which includes the drawing on the right. The packaging is some of the earliest to be found for transistors and many interesting features are present. Apparently WECO was making a generic transistor box at this time which could be ink stamped (probably by hand) with the specific product number and date of manufacture. Other nice touches are the transistor schematic symbol and proud announcement "MADE IN U.S.A."
The opposite side of the box lists two patent numbers... 2569347 is William Shockley's historic junction transistor patent which was originally filed on June 26, 1948, just nine days after the original point-contact transistor patent by John Bardeen and Walter Brattain. I am at a loss as to the meaning of the number PS6858 on the end flap. Click here to see my transistor history page for more details on the original patents.
On the right is a closer look at the business end of the WECO 1740 photocell. Bevitt indicates that the tiny package is made of clear plastic with all faces but one painted black, leaving a transparent window on top revealing the single-crystal bar of germanium. The PN junction is in the middle and forms the photoelectric element. One side also has a swatch of white to make the identification paint dots more easily readable. The yellow dot on the end signifies the positive lead for the biasing circuit. The purple-yellow-black dots on the side use the same color code as resistors which translates to 7-4-0, with the leading "1" being implied.
I find it interesting that even though the name transistor is shown in very large letters, the manufacturer still felt obligated to clarify in parentheses that these devices were crystal amplifiers. Of course in reality the term transistor was still not well known at that time. I wish the price had been included in the ad, but it was probably omitted to avoid "sticker shock" among prospective purchasers.
Click here to see a larger version of the ad.
Above are two photos of one of these extremely rare early devices. Notice the very crude hand-made appearance. One of the aspects I find so fascinating about the unique physical design. At the dawn of the transistor era there was certainly no standard for common appearance. It is likely that only a few hundred RD-2517's were ever produced. The case is a plastic shell into which the germanium core assembly was placed and then the whole thing was sealed up with an epoxy-like substance. The collector lead was indicated by the red marking seen on the left side. The letters G P (Germanium Products) were molded into the bottom of the case as seen on the right. Electrically the RD-2517 was an NPN device.
In addition to the RD2517, GPC eventually offered three other grown junction NPN transistors in this initial series. They were the RD2520, RD2521, and RD2525. It was typical of early transistor manufacturing techniques to obtain devices with a wide range of operating characteristics. Basically the manufacturer would make a batch of transistors and then test the resulting units and categorize them by "type". The various resulting "types" of transistors were then offered for sale. The variations in characteristics are shown in the following GPC specification listing.
These early GPC transistors were not hermetically sealed and often suffered from failure due to the entrance of moisture over a period of time. I suspect that the apparent early demise of Germanium Products Corporation was at least partly due to their inability to solve manufacturing problems and build a substantial customer base. A few of these early GPC transistors did find their way into hearing aids, particularly the hybrid Sonotone model 1010 which is shown in more detail on another of my pages. In that application the device was known as the TN-10. GPC later built transistors with the more typical oval-shaped metal cases which were hermetically sealed.
Above is a an ad which appeared in the May, 1954 issue of Radio Electronics.
As noted in the ad text, when GPC shifted to a more modern hermetically
sealed type of transistor they were left with a quantity of the original
plastic encapsulated transistor inventory to dispose of. Their sales agent,
Federated, apparently came up with this approach targeted at hobbyists
and electronics professionals. Although somewhat high by 1954 standards,
the price was an incredible bargain for those who could afford it. It is
interesting to see that no type numbers or specifications were mentioned...
apparently when you ordered they simply shipped whatever was at the top
of the storage bin.
Germanium Products Corporation also did some early work in the field
of high frequency Tetrode transistors. Here is a nice looking RDX 301 prototype
(note the hand painted serial number on top). Another clue to its early
origins is the fact that the leads are not spaced to "key" the collector.
Rather the collector has been identified with a swatch of red paint. The
three photos on the right show the lead spacing difference between a production
RDX 300 and this particular RDX301. The collector is on the left in all
instances. This device was made around 1955 and is cross referenced in
early specification tables as also having the RETMA designation of 3N23A.
It had a cutoff frequency of from 20-35 mc. which was an order of magnitude
higher than most conventional transistors of that period.
Above is a TI model 800 semiconductor. During my research I found
it listed in the 1958 Newark Radio catalog where it is described as being
an NPN photo transistor. It sold for a hefty $8.15 and one of its uses
may have been in first generation computer punched card readers. According
to early text books such a two terminal device was still considered to
be a transistor because it had three layers of N, P, & N material.
The difference was that no base connection to the "P" layer (where the
incoming light was focused) was brought out of the case. This particular
800 dates from late 1957 and has a small clear plastic lens (hard to see
in photo) embedded in the top of the case.
National Union was a well known tube manufacturer that briefly ventured
into transistors. They did not become a major player in the semiconductor
business which makes surviving devices somewhat scarce. Pictured above
on the left is a 2N39 (PNP junction type) from late 1953. The beautiful
green package on the right is flanked by the 2N39 on its left and the 18-B
(point contact type) on its right which has a date code of 328 from mid
1953.
Below is a Transitron 2N85 which was a very early attempt to produce a "power" transistor by soldering a standard case to a small heat sink. The Transitron Corporation (note the stylized "T" logo above the type number) was one of the original Bell Labs licensees and was a pioneering first generation transistor manufacturer. Unfortunately they did not gain a solid market share and left the transistor business by the end of 1955. There is a hard to read date code on the heat sink of this device and it appears to be 515. In my collecting experience Transitron products are among the most difficult to find.
Transistor Products Incorporated (later Clevite) was another pioneering
company that soon faded from the manufacturing competition. In 1953 they
had a series of point contact types from the 2A through 2G. TPI was one
of the few companies to produce point contact transistors with exactly
the same style "type A" case that Bell Labs originally introduced. Surviving
examples are scarce.
Even CBS had a division which did developmental work with transistors.
This is a PT-2S point contact device which appeared in 1953 product literature.
The PT-2S was intended for Switching, while the PT-2A was for Amplifying.
The design used was somewhat unusual for a point contact transistor
of this vintage since it has no adjustment access. The top of the PT-2S
can has a solder seal but otherwise the case and its long triangularly
spaced leads give every impression of being a much later TO-5 type design.
I assume that the three dots on the right side indicate the collector terminal.
It appears that CBS design engineers were ahead of their time.
Above is a photo gallery of the Westinghouse 2N54 which was a germanium PNP audio transistor. It was announced in the new products column of Radio-Electronics magazine in January, 1955 but appears to be very crude in terms of construction. Most other manufacturers had adopted hermetically sealed metal cases by this time but the 2N54 appears to be an orange epoxy plastic case with a coat of gray paint and black identification stamps. On the reverse side 337 is the RETMA code for Westinghouse and 518 is the date code. Note the wonderfully pock marked and rough surface which probably indicates that these devices are engineering prototypes. They may not have evolved any further than this, but if anyone has a more refined and later example I would welcome any new information.
The 2N54 has a most unusual pin out arrangement as detailed in the
diagram above which appeared in the first edition (1956) of the General
Electric transistor manual. For some reason GE was very liberal in providing
detailed technical specifications for all of its competitors, at least
at the very beginning. I don't believe any other manufacturer used this
particular layout for a transistor product. Survivors of this series (2N54,
2N55, 2N56) are most likely extremely rare and may only exist today as
the fortunate result of "escaping" the factory in an engineer's lunch box
to become a souvenir.
The internal assembly as seen above is similar to the 2N32 commercial production units which were announced by RCA in September, 1953. A construction diagram for the RCA developmental point contact transistor appeared a year earlier in the September, 1952 issue of Radio and Television News.
I was fortunate to receive the historic devices below from a retired employee of the RCA Research Labs who worked there in the early 1950's. The two transistors on the left are among the earliest commercially available RCA transistors and date from 1953 and 1954. The green 2N35 is an NPN junction type and the white 2N32 is a point contact type. The other transistors date from around 1952 and were used by RCA engineers to develop the worlds first experimental all-transistor television (excluding the 5" black and white picture tube).
Red markings indicated PNP types and green markings or case color indicated NPN types. The RCA television contained about 22 experimental transistors, weighed 27 pounds and received one fixed channel at a distance of up to 15 miles. It was first demonstrated in early 1953. The successful construction of such a device was an incredible accomplishment and pushed the transistor technology of the time to its limits!
Above on the left is a close up look at a historic RCA TA153 PNP developmental junction transistor. This device was the predecessor of the 2N34. Notice the serial number on top and the extremely hand made look of the molded plastic case. This device obviously spent many hours in the research laboratory as scientists pursued their goal of learning how to make commercially viable junction transistors. The first commercial RCA transistors were packaged in nice looking 2-1/2" x 3-1/2" envelopes like the one shown above right.
Shown above are some more early envelopes and their celluloid "master
carton". This type of packaging was used at least through 1958.
Below is a "family portrait" showing the evolution of the 2N35 packaging from its release in 1955 through the late 1950's. The scale is the same in all images... the case size shrank considerably as production techniques were refined.
Sylvania distributed a hobbyist booklet entitled "28 uses for Junction Transistors" in early 1955. It was even sold through mail order distributors in 1956 as a "kit" containing some of the components needed to build the projects.
Sylvania was an early developer of power transistors and produced two "non-standard" case styles before the familiar TO-3 package was introduced. The first power job was a the 2N68, a PNP type which was rated at 2.5 watts. A similar looking NPN device, the 2N95 had the same rating. Shown below are a production device on the left (white lettering) and an early unit, possibly a prototype, on the right (green lettering). Note the much thicker leads on the production device.
The second style, designated the 2N101, omitted the cooling fins and was a PNP type rated at 1.0 watt. It also had an NPN complementary partner, the 2N102. Again, note the much thicker leads on the production device on the left.
The third style was produced with what became the industry standard TO-3 case. Shown below are a PNP 2N255 along with an unmarked Sylvania prototype TO-3 device. Note that this particular example has a later date code from 1960 and the box is labeled "Semiconductor" rather than "Transistor". It is interesting to see that the case was not totally painted.
Sylvania, while not as successful as the major players such as RCA
and GE, did play an important role in the formative years of transistor
development and had a significant impact in many early transistor based
commercial products.
GE also quickly developed junction transistors using distinctive case styles for their PNP and NPN designs. Shown below is an early prototype PNP grown junction germanium transistor. This laboratory device was hand lettered as a ZJ1 and has a date code of 4-16 or the 16th week of 1954. The meaning of the other characters is not known but they were probably serial numbers. This device was a precursor to the famous 2N43, 2N44, & 2N45 series of transistors. These were originally developed for military applications and eventually distributed in large quantities commercially and were designated by the addition of an "A" suffix on the RETMA type number (i.e. 2N43A).
Here is the new product announcement for the 2N43, 2N44, and 2N45 which appeared in the November, 1953 issue of Radio-Electronics.
Below are a pair of GE 2N167 NPN transistors being used in a very early computer logic card. GE used a flattened oval for NPN devices and a round case for PNP devices. All first generation GE transistors had a pinched off evacuation tube through which all air was removed from the devices at the time of manufacture. As a result the early GE products are very distinctive and unmistakable.
Below is a view of the typical GE packaging which was used in the mid 1950's. Orange and Blue colors were applied and nifty looking plastic tubes were used to protect the transistors.
Shown below is a pair of early GE PNP transistors. On the left is a 2N45 from 4-48. On the right is a 4JD1A17 from 5-52. Descriptions from the first GE transistor manual, dated December 1,1956, appear below the photo. The purpose of the 4JD series of devices is unclear to me and they are much less common than the conventional RETMA series. From the 1957 Lafayette Radio catalog the 2N45 was priced at $3.05 and the 4JD1A17 at $6.02. Clearly there was a premium associated with the 4JD devices.
GE naturally also introduced their own version of a low priced audio
transistor for hobbyists and experimenters, the germanium PNP type 2N107.
It was announced in a press release dated Monday August22, 1955 and had
the same physical design as other similar GE PNP transistors such as the
2N45 shown above. Here is an excerpt from the press statement: "According
to James H. Sweeney, G-E manager of marketing for semiconductor products,
the suggested distributer price of 'well below two dollars' for the new
2N107 transistor makes it the least expensive of any transistor currently
available."
It is doubtful that more than a handful of CK703's were ever sold to hobbyists due to the steep price of $18.00 (equivalent to over $110.00 today). In fact I only know of a couple in the hands of collectors today. A Raytheon engineering memo indicates that the CK703 was apparently not a good performer and it was replaced by 1952 with the somewhat improved but nearly physically identical CK-716. These were also priced at $18.00 and even appeared for sale in the 1953 and 1954 Allied Radio catalogs.
Most likely Raytheon was trying to move their unsold inventory since by that time the point contact design was already obsolete. I imagine that hobbyist sales for the CK-716 remained poor for the reasons previously cited. Based on my collecting experience, it seems safe to say that only a handful of CK-716's are preserved today. The piece of history shown below is undoubtedly the greatest rarity in my collection. After years of searching I was beginning to resign myself to never finding one. Happily I was wrong.
I was told that my CK-716 came from the famous Sandia National Laboratories which were established in 1949 to develop science-based technologies in support of national security. The paper label has partially collapsed over the "access hole" in the side of the case. This design aspect is similar to the Bell Labs "type A" units and was used to adjust the point contact whiskers at the time of manufacture. Also notice the socket which was made exclusively for the Raytheon CK703/716 design. These too must be extremely rare since so few were ever needed. On the right is a view of the unique "double stubby" pin arrangement. The brass case itself is the base terminal, the emitter is the pin on the left and the collector is the pin on the right.
Below is the famous Raytheon CK722, which was the first mass-produced PNP junction type germanium transistor. It was intended for general purpose audio frequency applications.
This is what you got in February, 1953 if you were one of the first
electronics experimenters to visit your friendly neighborhood radio parts
store and plunk down a breath-taking $7.60 for a brand new Raytheon CK722
transistor. 
Click here to see the very first ad for the CK722 in the March, 1953 issue
of QST. The earliest commercial units were shipped in full-size
tube boxes which were stamped "CK722" on the end flap. Inside, almost lost
in the generously supplied packing material, was the transistor itself.
What a thrill it must have been to marvel at its incredibly small size
and anticipate the wonderful projects it would surely allow you to build.
A close-up view reveals the details of the stamped logo on the front and the date of manufacture on the back. This is an extremely early unit which was made during the 7th week of 1953 (February). The first non-trade related magazines to announce the CK722 also had cover dates of February, 1953.
The CK722's distinctive rectangular case was produced first in black plastic (1952 - 55), then blue painted metal (1955 - 57), and finally unpainted metal (1958 - 63). The vacuum tube style shipping cartons soon gave way to a relatively compact envelope style of packaging. The CK722 was a very popular device with the first generation of transistor experimenters in the 1950's due to its affordability. As production ramped up, the price of individual units soon dropped from the initial $7.60 of February, 1953, to a mere $0.99 by December, 1955. The abundance of CK722 based construction articles from this period stand as proof that it was arguably the most popular transistor type in semi-conductor history. The red labeled CK722 on the right is somewhat of a mystery. It was purchased long after Raytheon stopped making this type of case. One theory is that many unmarked surplus CK718 (an identical looking hearing aid type) transistors were bought, relabeled and imported from China. The unit came in a small vacuum tube style box with the name "National Electronics" on it.
As transistor manufacturing technology quickly advanced, Raytheon was able to dramatically reduce the size of their products. Interestingly, in order to retain the original external dimensions of the CK722 they simply embedded a small transistor inside a largely empty case.
The ad above appeared in the September, 1957 issue of Popular Electronics. It clearly illustrates the first generation black plastic case style of Raytheon transistor. This style of transistor was last made in 1955 and as succeeding generations replaced them, the surplus CK718, CK721, and CK722 transistors found their way to middlemen who offered them for sale to experimenters at bargain prices. The three transistors shown above the ad are from this surplus and are completely devoid of any identifying markings other than a red dot to indicate the collector lead. Although quality was somewhat variable (of the units shown, one is barely good, one has average gain, and one has high gain), the low price made it easier to justify taking a chance.
Amazingly, I even found an ad with the venerable black cased CK722 apparently being offered for sale in 1970 at about 33 cents a piece by the well known Kansas City supply house of Burstein Applebee. Of course there is no way to tell now if they were actually shipping the real thing or simply a functional equivalent. Regardless, the catalog picture is certainly correct.
One of the most unusual variations of the black plastic first generation Raytheon transistors is shown above. As noted elsewhere in these pages, early transistor production had a significant percentage of defective units. Rather than simply discard all the non-functional devices, Raytheon had the novel idea of using some of them as promotional advertising give aways. Because the transistor was still a brand new and uncommon device in 1955, most of the general public had only seen them in pictures, if at all. These non-functional CK7XX units were stamped "MECH SAMPL, ELEC REJCT" on one side and "RAYTHEON" on the other. Some of the look alikes were stapled to 3" x 6" cards which were given away at radio shops to promote the introduction of the Raytheon model 8TP transistor radio. Click here to see both sides of such a card. The model 8TP radio is described in detail in my transistor radio web pages.
Here are a few other examples of different manufacturers case styles.
From l to r: a rare Hydro-Aire type CQ-1 from the mid 1950's, a Philco
2N62 from 1955, Texas Instruments unpainted and painted units, a Sylvania
2N35 from 1954 and a General Electric 2N167 from 1959.
\
If you check a modern transistor cross reference list you won't see the 2N626 listed (nor the other three RETMA numbers that ARA reserved... 2N676, 2N1019, and 2N1020). The fact is that nothing comparable to these device was ever made! The four module types were apparently intended to be marketed by ARA but based on a search of the period literature I can find no mention of them ever being used or sold commercially. They were most likely a developmental concept which never successfully took off.
The $150.00 cost for the 2N626 was astronomical in 1958 and the 2N676 PNP version was priced even higher at $175.00 for 1-10 units. This would be in the range of $955.00 to $1100.00 today after taking inflation into account! I can only assume that ARA was hoping to win a lucrative military contract for these products, since the U.S. government was among the few potential customers with pockets that deep. Apparently it didn't happen as far as I can tell from my research.
One of the primary features of the 2N626 composite transistor was
a huge current gain which was listed as ranging from 18,000 to 30,000.
The device would also dissipate a whopping 10 watts of power. Click
here to see the 2N626 data sheet. The four ARA devices shown here may
be among only a handful that remain in existence today. They came from
the estate of pioneering transistor experimenter and author, Lou Garner
and were probably supplied to him as a professional courtesy for review
purposes. A special thanks goes to my friend Terry Hosking for generously
providing access to a copy of the original ARA product literature.
I am also greatly indebted to my good friend and fellow transistor collector, Jack Ward, for helping to make this display possible.
Jack has created what is undoubtedly THE definitive Internet site for the Raytheon CK722 transistor. To learn more, please visit his web pages. Jack has also authored an excellent reference work on the CK722 which he is selling as a 2 CD set.
