Vintage CV series semiconductors

Many UK transistors and diodes, including early types whose manufacturer is identifiable by their unique shape, can be found, usually unbranded, bearing a part number CVnnnnn. These are UK inter-service military-specification devices, the prefix CV being an abbreviation of 'common valve' ('thermionic valve' being British for what the Americans call a 'vacuum tube'). This designation was introduced in 1941 before the transistor had been invented and the series contains both vacuum and solid-state devices. It runs from CV1 to about CV12000 (!).

Each particular CV number is stated as having one or more 'prototype or possible substitutes', which are usually commercial device part numbers. Thus one CV type may be made by several manufacturers and I have found that sometimes a type may also be made by a manufacturer not listed in the Register entry for that CV device. I suspect that most CV types were normal commercial devices, possibly with tighter parameter selection. If anyone knows more details of this, please

There is good information on the Web about the CV series in general, because of the large number of valve/tube enthusiasts who share information about it. The Virtual Valve Museum describes the markings on the CV series types which convey information about the date and location of manufacture. A CV semiconductor device usually (but not always) bears the markings:

• The part number CVn, where n is a three, four, or five digit number.
• A 'K' that means that it has been manufactured to specification,
• a 'B' that indicates that approval was given by UK authorities,
• a one-, two- or three-letter factory code that also identifies the manufacturer, and
• a two-letter date code.

The Virtual Valve Museum allows download of scanned images of the CV registers in 1944 and 1946 (which contain only a few early semiconductor diodes) and the definitive source of data for CV semiconductors: the CV register of 1963.

Because of the unusual nature of the series, there is little structure to its numbering, although there are some small groups of semiconductor devices that correspond to parts of a series from the same manufacturer. Other oddities include:

• I suspect the series is largely but not wholly chronological. It seems likely that gaps were left, some of which were filled in later.
• It is common for a CV type to specify one or more other CV types amongst its 'prototypes and possible substitutes'. In some cases, only CV types are given. This may result in circular CV references with no apparent commercial prototype.
• Each CV type has a short description. For semiconductor diodes this is not very consistent. Many low-numbered diodes are described as a 'crystal', higher-numbered ones are mainly a 'semiconductor diode' but sometimes a 'semiconductor rectifier'. Other values are used, albeit rarely.
• For reasons unknown to me, a large number of American diodes and transistors are included.
• Some types are 'Cancelled'. I don't know what that signifies. Others are 'Superseded' by another CV number.
• For some types the 'prototypes and possible substitutes' field just contains a Ferranti drawing number.
• There are quite a few types described as some kind of semiconductor for which the prototype has a part number unknown to me.
• There are quite a few types for which the prototype is a VX series device. I can find no information on this series, but I suspect that it was originally a valve experimental series. In fact, VX semiconductors are hard to find.
• Manufacturer codes can be found that are not listed in the Joint Service Specifiation K.1001. These are not obscure manufacturers: a common one is NTN which appears to be Newmarket Transistors (listed in K.1001 as CO) and another is FRD which is Ferranti (listed in K.1001 as four sites: R, RA, RB, RC). I have others that I cannot identify. I'm unaware of a later manufacturer list that may explain these.

Contemporary semiconductor manufacturers' literature often contains lists of their CV types, and this often specifies equivalences that are different from, or additional to, those in the CV Register.

Obviously I do not possess all the thousands of semiconductor devices in the CV series, in fact I am seeking many of them. It is noted in the text where I am seeking examples of any particular type: if you have some for sale or exchange, please

It's hard to know how to show this enormous series. The lowest-numbered semiconductor is CV101, and the lowest-numbered transistor is CV2389.

The oldest semiconductors in the series are microwave diodes, mostly silicon, some of which were used in radar equipment during World War 2. These were made by either BTH or GEC.

The 1944 and 1946 CV Registers are in two parts, A and B. Part A covers CV1 to CV290 (with some gaps) and lists each type with a single prototype and a short description. Part B covers higher numbered types and is less informative.

The 1944 Part A contains the following semiconductors:

• CV101 to CV103 are described as crystal valves with no prototypes.
• CV111 to CV113 are described as crystal valves with no prototypes.
• CV241 is described as a crystal valve 'For wavemeters'.
• CV246 is described as a crystal valve 'Detector version of CV111'.
• CV247 is described as a crystal valve 'Detector version of CV112'.
• CV253 is described as a crystal valve 'CV111 selected'.

The silicon diode CV253 shown has the part number engraved on the metal.

The 1963 Register is more informative and allows multiple prototypes 'or possible substitutes'. It exists with various 'amendments' which do change some of the details of semiconductors. It has a larger number of 'crystals', the following are those below CV1000 that are described as crystals in early amendments and I believe to be microwave types:

• CV101 to CV103 are described as crystals, and CV101 has prototype/substitute CV102.
• CV111 to CV113 are described as crystals, and CV111 has prototype/substitute CV253.
• CV162 is described as a crystal and has prototype/substitute 52094, which is unknown to me.
• CV165 is described as a crystal and has prototype/substitute CS36-A.
• CV226 is described as a crystal but has no prototype/substitute.
• CV241 is described as a crystal but has no prototype/substitute.
• CV246 and CV247 are described as crystals but have no prototype/substitute
• CV253 is described as a crystal and has prototypes/substitutes CS3-A (a BTH microwave diode) and CV111.
• CV291 is described as a crystal and has prototype/substitute CV291(BTH), whatever that means!
• CV351 is described as a crystal and has prototype/substitute CV1055 which seems to be an error as that is a valve.
• CV364 is described as a crystal and has prototype/substitute CV291.
• CV367 is described as a crystal and has prototypes/substitutes 1N21B (an American microwave diode), CS2-A, CS37-A and CV3923.
• CV727 is described as a 'crystal valve' and has prototypes/substitutes 1N21, 1N21B (American microwave diodes) and CV367.
• CV749 is described as a 'crystal valve' and has prototype/substitute 1N23A (an American microwave diode).

A few of these are sometimes for sale on eBay but old microwave diodes seem to command quite high prices. I cannot see why as I very much doubt that anyone still uses them.

Apart from the above, there are few semiconductors below CV2000:

• CV164 has as prototype the BTH point-contact diode CG3-E. I would like to obtain this.
• CV166 has as prototype the GEC point-contact diode GEX54/3. I would like to obtain this.

• CV425 is a miniature germanium point-contact diode with several 'prototypes and possible substitutes': the BTH type CG1-C, the GEC types GEX44 and GEX45/1, and the Mullard OA74. Like many of the CV types, it is commonly found sealed in a small packet containing a label bearing various military numbers, and often an old-fashioned designation such as the 'CRYSTAL VALVE' shown. The images show the orange-sleeved Mullard version, the sub-miniature (so-called 'form H') BTH version, and a so-called 'form E' BTH version. The latter two outlines are not the same as the BTH prototype CG1-C, which is the same shape as the CVX425 immediately below.

• This is an extremely rare CVX425 shown in a compound image. I speculate that the X denotes 'experimental': obviously the CV number has been allocated, but type approval has presumably not yet been gained. The small polythene packet contains a label printed with the BTH logo plus 'CRYSTAL VALVE', and the 'X' has been added by hand on the CV number field. The diode outline is BTH 'form C', correct for the CV425 prototype of CG1-C.

• CV442 has as prototype the GEC point-contact diodes GEX33 and GEX35, and the Mullard OA73. The image shows a typically grubby GEC example.

• CV448 is the commonest of the CV diodes. Interestingly it is marked as obsolete in the 1963 CV Register, although it is not present in the 1946 register. The 1963 document states that it has been replaced by CV7310 which has several 'prototypes and possible substitutes': the Mullard type OA91, the GEC type GEX23, and the Hughes type HG1006. However, I possess manufacturer data that state CV448 is equivalent to Mullard OA81, GEC GEX54 and BTH CG4-E and my image shows all three manufacturers' versions from bottom to top respectively.

• CV727 has as prototype the American microwave diode 1N21.
• CV1353 has as prototype the Mullard diode OA81. I would like to obtain this.
• CV1354 has as prototype the Mullard diode OA85. I would like to obtain this.
• CV1632 has as prototype the American UHF mixer diode 1N82.
• CV1785 has as prototype the American microwave diode 1N26.

Between CV2001 and CV5000 there are remarkably few interesting semiconductors:

• There are quite a lot of semiconductor devices with prototypes either another CV type or a VX type. I'm unsure whether these were manufactured in any quantity.
• There are quite a lot of American types, mostly low-numbered 1N-series diodes. These do not interest me.
• CV2154 and CV2155 are silicon microwave diodes with prototypes CS12B or SIM2 and CS12BR or SIM5.
• CV2226 is listed as a semiconductor diode, puzzlingly with multiple prototypes CV2355 to CV2357. These are in turn have prototypes CV7181 to CV7183, which in turn have VX prototypes!

• CV2258 has prototype CS4B, an early BTH silicon microwave diode. My example (compound image) was made at Rugby (the factory code S is assigned to AEI in the document above, because AEI took over BTH) and bears the War Department 'broad arrow' as well as the letters WD which are probably the date code for April 1965.

• I have CV2279 and CV2290 diodes in the same grubby GEC point-contact outline as the CV442 above. Neither has a GEC prototype listed in the Register but GEC documents equate them to GEX36 and GEX66 respectively.
• CV2310 is listed as a 'crystal' equivalent to the GEC diode GEX64.
• CV2355 to CV2357 are described under CV2226 above.
• CV2384 has as prototype the semiconductor diode type ZS10C but is superseded by CV4073 (see below). I have no data for this but ZS10A and ZS10B are Ferranti silicon junction diodes.
• CV2391 and CV2392 are diodes with VX prototypes.

• CV2389 is the lowest-numbered transistor in the CV series. My example was made by GEC. It has prototypes CXT1 and CV7005. The latter has prototypes OC7l, GET103 and 2N280. As one might expect, these are very early germanium PNP AF types dating from 1954. I believe the CXT series contains experimental GPO types but I have no documentation for it.

• I also have a device marked CVX2389 and I believe that X stands for 'experimental'. Perhaps GEC had not yet gained full registration of the CV number 2389. If anyone can explain this, please In any case it is clearly one of GEC's first transistors, employing the crude black-painted case that was only used for the commercial types GET3, GET4 and GET6. The packet bears the appropriately early date of August 1956.

• CV2400 is the second transistor in the series, with prototype CXT2. I don't have any CV2400 but I do have this CVX2400 that is clearly a Mullard type in a black SO-2 glass capsule. As the image shows, it has an unusual marking, NB/D. The meaning of the N is not known to me, but the /D confirms that this is a Mullard (Mitcham) device.

• CV2413 is a diode superseded by CV4074 (see below).
• CV3923 is a microwave diode with prototypes CS37A and others.

• CV4073 has as prototypes the CV2384 that it supersedes, the Ferranti ZS10 and the GEC SX641 silicon diodes. The GEC version is shown.

• CV4074 has as prototypes the CV2413 that it supersedes, the Ferranti ZS20 and the GEC SX642 silicon diodes. The Ferranti version is in the image.

• CV4077 has as prototypes the Ferranti ZS21 and the GEC SX643 silicon diodes.
• CV4078 has as prototypes the Ferranti ZS22 and the GEC SX644 silicon diodes.

I've seen it written, and it seems to be true, that it was decided at this point to split the series: the CV7nnn block would be semiconductor devices and the CV6nnn block would be everything else. Certainly there are no semiconductors in the CV6nnn block, which only extends as far as CV6234. The CV7nnn block is all semiconductors and goes as far as about CV7875, although the Register is confusing for higher-numbered CV78nn types, and there is a block of apparently unused numbers at the end. The split was not continued and CV8nnn, CV9nnn and CV10nnn types are a mix of semiconductors and thermionic/other devices.

The CV7nnn block is enormous so I cannot even list them all. It is by far the most common number range, for instance an eBay search returns almost only CV7nnn types. I shall refer only to those that are particularly interesting. The block starts with CV7001 (there are in fact no types numbered CVn000 in the entire series), and the first types are all low-power germanium PNP from the big three UK manufacturers: Mullard, GEC and STC.

• The Register gives CV7001 prototypes PO3, GET103, CV7001(Mullard) and ACY31, AF amplifier types. I'm not sure what the 'Mullard' in brackets means. My image shows STC, GEC and Texas Instruments (UK) examples, the last not being a prototype.
• The CV7002 is a medium-power AF type with prototypes PO4, GET116 and OC72(Mullard). I show two examples, one in classic GEC medium-power encapsulation, and one in a unique Mullard attempt to fit an SO-2 metal can into the GEC heatsink shape. Presumably Mullard took this drastic step because the device was registered by the military as having those dimensions. The Mullard device has a datecode of 6426, for week 26 of 1964, and the GEC example uses the code WM, for December 1965.
• The CV7003 is a switching transistor with prototypes OC44, GET874 and STC types TK31C and ASY55. I show three examples, one in classic GEC low-power encapsulation, one in a Mullard SO-2 glass can, and a third in a short SO-2 or TO-1 metal can made by Texas Instruments (UK). The last is again not listed in the Register, although all should be interchangeable in a circuit.
• The CV7004 is a another switching transistor with prototypes OC45, GET873, and the STC types TK30C and ASY54.
• CV7005 has prototypes OC71, GET103 and 2N280, AF amplifiers. The last is a renumbered OC71, made for the American market and sold by Amperex.
• CV7006 has prototypes OC72, GET103 and 2N281, AF output types. The last is a renumbered OC72, made for the American market and sold by Amperex.
• CV7007 has prototypes OC77 and 2N284A. The last is a renumbered OC77, made for the American market and sold by Amperex.
• CV7008 has prototype GET106.
• CV7009 has prototype GET110. My example in the image has a date code YG which apparently is July 1967.

The group continues and eventually there is a series CV7080 - CV7086 that contains only power transistors:

---

---

---

---