RCA's 1600 series tubes were usually developed for special applications. Some are relatively well known (like the 1625, which is a 12V version of the 807). Others, however, are so obscure as to be mentioned nowhere in any literature that I have ever seen. The following is an article that I wrote that was published in the February 1998 issue of the Antique Wireless Associations "Old Timer's Bulletin" (OTB) on 5 'new' 1600 series tubes.
Since the publication Lud Sibley's article on Sylvania's 1200 series tubes and RCA's 1600 series tubes (OTB May 1994), I have been trying to gather a collection of these unique and interesting tubes. In doing so, I have come across five types not mentioned: types 1669, 1672, 1680, 1682, and 1684. The 1680, 1682, and 1684 appear to be some of RCA's initial efforts to design tubes especially for computer use, and the 1669 and 1672 are phototubes.
About a year ago I acquired types 1682 and 1684 from my good friend and fellow collector John Walker, and subsequently was able to return the favor, when I found a pair of 1680s and sent one to him. These tubes are all seven pin miniatures. Although John and I both have large collections of tube data, neither of us had any references for these tubes, and I set about the task of figuring out what they were.
The first clue was that the type 1684 also had an IBM part number: 304994. The 1682 had a partial IBM number, but it was unreadable. So, these tubes were probably used for computers. Tubes for computers were specially designed to prevent "sleeping sickness", which is a cathode interface condition caused by prolonged periods of operation with no plate current. A phone call to IBM's headquarters pointed me to their parts line. This is a 900 number, where for a fee, they will help identify parts. IBM's parts line equated part number 304944 with a type 5844 tube. The 5844 is a computer-rated variant of the 6J6. So, the 1684 had been superseded by the 5844, and was likely very similar in characteristics. A comparison of the 1684 with a 5844 from my collection showed them to be of identical construction, and the 1684 tests good on my Hickok tester when set for a 5844.
None of the other tubes had clear IBM part numbers, but, given that the 1684 had a known industrial equivalent, I decided to draw basing diagrams for the 1680 and 1682 and search for likely candidates. The 1682 has the same basing as the 1684, which surprised me, since they are of very different construction. Through a search for tubes with like basing, a visual comparison, and testing, I quickly identified the 1682 as a probable developmental equivalent to type 5964, another computer-rated 6J6 variant. The same process took a little longer for the 1680, as the basing was a bit more difficult to determine visually. But, the 1680 was finally identified as being a probable developmental equivalent to the type 5915, a computer-rated 6BE6.
While the lack of RCA factory data prevents exact dating, types 5844, 5915, and 5964 were all registered in 1950 or 1951. Also, I found the 1680, 1682, and 1684 for sale in surplus catalogs of the late 1950s/early 1960s. All of this leads to their probable development in the late 1940s.
Two other heretofore undescribed 1600 series tubes have come
to light. Both are phototubes. The first is the type 1669. This
is a cartridge type photocell, and physically resembles a RCA
type 921 with the same T-7 tubular bulb and end connectors, except
that the tube and the photocathode are about three times as long
as the 921. The other phototube is the type 1672. The example
that has come to light is a GE tube dual numbered "1672/441".
So, it's characteristics assumedly closely resemble the GE type
441, which is a four pin vacuum phototube with a T-8 tubular bulb
and an S-4 (blue) spectral response, and, as such, is essentially
the 4 pin equivalent of the type 929.