All PM valves were directly heated and all (we think, though possibly not the PM2BA) were originally made by the azide process (which was central to the Philips-Mullard link-up). From the mid-1930s onwards Mullard were forced to reinvest in oxide coating processes so PM types then still in current use were redesigned (in a somewhat piecemeal fashion) to suit manufacture by the oxide process. As far as we know the azide process was dead in the UK by 1940 and all PM types made subsequently (for replacement purposes) would have modern oxide-coated filaments.
Most, if not all, of the PM series were Mullard-made copies of existing Philips types. Philips-Mullard issued lists of available PM types which included some Philips types, redesignated as PM types, which might have been available as relabelled imported samples but which, presumably due to lack of demand, were never made in Britain or supplied in quantity. Thus the PM lists include some types which may, strictly speaking, never have existed! Conversely, we have samples of a two or three rare PM types which do not appear in any lists seen.
MORAL: Do not ask for a complete list of PM types and do not believe it if you have such a list.
The approximate age of Mullard-made PM valves can be determined by the style of base cap fitted to them and by the style of marking. Some Mullard-branded PM valves were in fact imported Philips- made valves. These too can be recognised by style of base cap and marking but I don't know exactly how these foreigners fit into the date sequence of the home-brewed ones.
The earliest PMs (PM3 and PM4) were fitted with base caps of similar shape and size to that shown in exhibit PM4 but with crude split pins of the type shown on exhibit Red Ring. None yet in the museum exhibits. The very earliest (imported?) samples were of this type and hearsay has it that 'British Made' referred only to the base caps and the paper strips gummed round them. These earliest samples also had 'PHILIPS MULLARD' and the type number (PM3) etched on the side if the glass envelope. We have, however, yet to see a PM4 as old as this.
The next batches, which I assume to be of British manufacture looked physically similar but were marked on the side with the Mullard logo and PM type number (PM3 or PM4). The wording on the paper strip seems to vary slightly from batch to batch but, since so many specimens have either lost their strips or have had a strip (not necessarily the original) glued back on, the exact wording on the strip is not always a reliable indicator of age.
The next physical change was alteration of the base pins to the sprung type shown in PM4. However, the vertical line up the side of the base cap and the letter A moulded in to indicate the anode pin were at this time carefully inlaid with white paint. Etched markings, including type number and Mullard logo, were still on the side of the glass. Around this time the number of PM types begins to proliferate and the etched type number is enlarged to make it more readable.
The next change was the discontinuation of the white inlay on the base cap. The etched markings remained on the side of the glass. The following exhibits fall into this group: PM1HF, PM1LF, PM2 and PM5X
The next stage was to redesign the etched markings to include a new Mullard logo, and to transfer these markings to the crown of the bulb except where a top cap was fitted. This group includes the PM22, PM3 and PM4
At this point things start to get complicated and it becomes more difficult to sort specimens into date order with confidence. The principal problem was the continuing proliferation of valve types; new style base caps may be designed and fitted to new types of valve while existing styles of cap may continue to be fitted to older types. Moreover, during the early 1930s there was a brief but unsuccessful flirtation with parallel-sided and unsprung base pins (see PM24A and PM24). In general, paper strips disappeared (though not overnight) and the height of the base caps diminished accordingly. Characteristic of this era are neater base caps, continuing reliance on the azide process with consequent internal blackening, absence of gold paint, and the style and placing of the bulb markings. Exhibits of this period include: PM24A, PM24, PM1HF and PM2DX
Gold spray paint appeared during the mid 1930s. The colour was originally a coppery gold but this was lightened in stages until by the late 1930s it had become a silvery gold, and finally (we think 1939) turned bright red. At first the gold paint seemed to be sprayed onto all types of valve except those which got hot, such as output valves. This was possibly as a sales gimmick as well as a means of hiding the unsightly results of the azide process. Exhibits of this period include: PM12A, PM1HL, PM2HL and PM2A an output valve
Once valves had been redesigned with oxide-coated filaments it was necessary to show customers how clean the valves were by leaving them uncoated except where an external metallised spray was functionally necessary. However, since customers had by this time been schooled to buy gold valves, it was necessary to leave at least a token amount of gold paint in place. Exhibits of this era include: PM22A, PM12M and PM12M
Finally, PM types made for replacement purposes during the post-war era were usually without any gold paint at all. Exhibits of post-war manufacture include: PM2B and PM24M
The identification of 'foreigners' remains, and we may not be completely correct in our identification. However, all Mullard valves having large, coarse silver codes stamped on the side of the base cap are of suspect origin. This form of marking was common practice on the Continent at the time, whereas British-made Mullard valves had such codes (if any) in small, fine print on the underside of the base cap. Alien suspects among the exhibits thus include: PM12A and PM1HL
Two final things before leaving the PMs.
The pentode was a Philips patent (around 1928) and Mullard, of course, made the first British type which was the PM24 (our PM24 is a later example). This was used, for example in the famous (or infamous) Philips 'Shoebox' all-electric AC mains radio sold in the UK around this time. The (4-pin + side terminal) PM24 was AC heated in this set and was used as the output valve.
The other thing concerns mystery exhibit which is captioned 'PM2'. The only Mullard- related feature of this specimen seems to be the paper label. The valve itself is a directly-heated oxide- coated triode of the later 1930s but it looks MOV-ish rather than Mullard and (compare L610) but some details of the anode box pressing are not quite 'pukka' for MOV and, more importantly, MOV did not at that time make any 'LF' types (see ink marking on pinch). The base cap (which may not be the original) is not MOV in style; it looks a bit like a Micromesh (STC) type but the valve itself does not.