Valve Questions

[Synchrodyne]

More on the EF91 in the attached Mullard advertisment from Wireless World 1949 September, including some information on noise figures.

Cheers,

[G6Tanuki]

Fascinating... as indeed has this whole thread been! For a designer, trying to strike the right balance between overall performance, reliability, cost and overall ability of a production-line to build the design - as well as acknowledging political constraints (for example being "strongly encouraged" to use only the range of valves or other components produced by a sister-company) must have been a struggle.

It's also most intriguing to compare the different directios different countries went in and how this affected design requirements: the US getting lots more stations and colo(u)r decades before the UK/Europe, but largely ignoring the use of UHF, and compared to Europe also being rather late into the "all transistor" TV (I'm guessing that their liking for larger CRTs than Europe meant they needed deflection stages with power-handling demands that were expensive to meet with semiconductors?).

Also, in the TV valve context, it's worth noting that AC/DC valve technique was rather uncommon in US TV sets: most of them used traditional big heavy transformers and 6.3V heater valves rather than our 100/300mA series-strings and direct mains rectification for HT. Half-wave-rectifying a 117VAC supply didn't offer enough HT 'oomph' to deflect a 26-inch colour tube.

The US also stuck to Octal-based 'tubes' for line-output stages whereas us Europeans went for all-glass bases (B10) a lot sooner.

Apart from the RCA Nuvistor, the one intriguing high-point of US TV valve development was the 12-pin-based "Compactron" range introduced by General Electric in 1960. Using Compactrons (some of which contained two pentodes and a triode) you could build a NTSC Colo(u)r TV with a surprisingly small valve-count.

More on Compactrons here: http://www.junkbox.com/electronics/C...besIndex.shtml

[n_r_muir]

Hi, I have been following this thread with interest. On reading the very interesting article on Compactrons I noticed the following statement

" the Compactron idea might have been developed further, with even more tubes in a single envelope, with perhaps resistors and capacitors to implement resistance coupling networks for amplifiers."

In the mid 1930's Telefunken did just that with their NF3, this had 3 valve/tubes including some resistors all in the one glass envelope.

[n_r_muir]

Hi oopps I got it wrong

It was Loewe who introduced the multiple valve

http://www.electricstuff.co.uk/loewe.html

[G6Tanuki]

I remember seeing one of the Loewe NF3 valves. A triumph of the German glassblower's crazed imagination! Who says "Integrated Circuits" are a new thing?

I believe it was created because at the time many countries taxed receivers on the basis of the number of valves. The NF3 counted for tax as one valve not three.

[Synchrodyne]

Originally Posted by turretslug

It does make me wonder what the gain/NF of Band III front-ends was like- but presumably, TV service areas weren't that large from powerful transmitters and beam aerials with reasonable gain were the order of the day. Similarly, UHF tuners with their PC88/PC86 line-ups- these valves with their multiple (grounded) grid connections and very snug construction looking like an attempt to get near to disc-seal type low-inductance connection but without the manufacturing and installation expense and complication. I gather that US UHF tuner technique was often no RF stage and diode mixer but this must have focussed the demands made of 1st IF amp performance.

By no means the whole answer, but some example noise and gain numbers are provided in the attached book excerpts.

The first (one sheet) is from “V.H.F. Television Tuners”, by D.H. Fisher (of Pye). I think the original was published by Heywood, but I have the American edition published by Philosophical Library in 1957. A couple of items that come out of the tabulation are that the 6AK5 was very good, and just how noisy, at Band III, was the PCF80.

The second (two sheets) is from “Radio and Television Engineers’ Reference Book”, 4th Edition, 1963, by Newnes.

Interesting is that valved VHF TV tuners could have quite high overall gain, around 50 dB in some cases. I guess that RF stage gain had to be kept high to minimize the effect of mixer noise, so that the overall gain more-or-less lay where it fell. In the solid state era, at least in US practice, 30 dB tuner overall gain seemed to be the general target. For example, that number was assumed by Motorola in the design of its MC1350 (or MC1352) plus MC1330 integrated circuit TV IF system in the late 1960s, which was about the first of its kind.

In noise terms, the PC97 (beam or guided grid triode) was no better than the PCC89 cascode, although it was 3 dB down in gain. So simplicity and lower cost were more likely the driving factors for its use.

Regarding the UHF valves as used in European UHF tuners, my understanding is that the PC86 was developed first for use in Germany, and was intended to cover both the RF amplifier and mixer roles. But better (lower noise?) RF amplifier performance was desired, hence the PC88 was developed with the focus on its performance solely as an RF amplifier.

Cheers,

[Synchrodyne]

And some addition noise and gain numbers are provided in a couple of pages from the 1955 Fewings and Fife (English Electric) paper, details provided in the first page.

Cheers,

[Synchrodyne]

Originally Posted by turretslug

and the apparently budget solution of a single ECC85 as RF stage and self-oscillating mixer was better and successful enough to be widely adopted.

Insofar as it used triodes, then from a noise viewpoint the ECC85 solution should have been better than any that used a pentode or pentodes. I imagine that the ideal approach, within the constraints of using consumer-type valves, would have been a cascode RF amplifier followed by a triode mixer. With the latter, there would be more flexibility in choosing the level of RF gain than with a noisier pentode mixer. But for the most part that would have been too costly for the setmakers.

Some reasonably authoritative background on the triode vs. pentode issue is provided by the attached book excerpts.

P.R. Keller provided a succinct summary in “V.H.F. Radio Manual”, Newnes, 1957, in the chapter “F.M. Broadcast Receivers”. The book also included a useful chapter on television tuners.

And J.D. Jones provided a commentary in “Frequency Modulation Receivers”, Heywood, 1956.

Overall, my take is that although pentodes were a bit noisier than triodes at Band II frequencies, the difference was not such as to preclude their use.

For the setmakers, the ECC85 solution was attractive on cost grounds, and probably also because ECC85-based front end units were available from component and assembly suppliers. And whilst it its use by UK setmakers was widespread, at least in the early days there were some variations.

Bush I think started with a pentode RF amplifier (EF80) and a pentode self-oscillating mixer (ECF80), but had migrated to the ECC85 by the time the VHF64 was designed.

Murphy I think went through a couple of iterations before settling on the ECC85. It started with a triode RF amplifier (6L34) and separate triode self-oscillating mixer (EC92). Next was the same but with a pentode RF amplifier (6F12 = EF91), then came the ECC85.

Pye had a bet each way. Its FenMan I used the ECC85, but the more expensive FenMan II had a pentode RF amplifier (EF80) and a pentode mixer with separate oscillator (ECF80). Its Mozart FM tuner had a pentode RF amplifier (EF80) with a pentode self-oscillating mixer (ECF80). Given Pye’s extensive involvement in VHF R/T equipment, one would be inclined to give some weight to its circuit choices.

There were more variations in the British hi-fi FM tuners, though.

Armstrong used the ECC85 front end from the start with its FM56, then through to the end of the valve era. It was also used in its FM-AM radiogram chassis, from the AF105 onwards.

Chapman started with the ECC85 in its initial FM81, but changed to pentodes in the FM81 Mk II and subsequent FM circuits. It used an EF85 RF amplifier (with or without agc) and an EF80 (or ECF80) self-oscillating mixer.

Jason generally favoured a pentode RF amplifier (Z77 or EF80), followed by a pentode mixer (Z77, EF80 or ECF80) either self-oscillating or separately excited. It did though use a cascode RF amplifier in its AM/FM and AM/FM2 models, although with what kind of mixer is unknown. But in its final FM tuner effort, the FMT4 of 1960, it changed to an ECC85 front end; this model was not available as a Jasonkit, though. It might be noted though that all Jasonkit FM and FM/AM models used pentode RF amplifiers and mixers, which from its literature one would infer were chosen for their relative stability. Its TV/FM sound tuners were different. The JTV had an ECC84 plus ECF80 combination following standard TV practice, whilst the JTV2 used an ECF80 with the triode as a grounded-grid RF amplifier and the pentode as a separately-excited mixer. A pentode mixer following a triode RF amplifier was breaking the “rules” somewhat at Band II and Band III. The Jasonkit versions of the JTV and JTV2 were supplied with pre-aligned front ends.

Leak started with a pentode RF amplifier (EF80) and pentode separately-excited mixer (ECF80) in its original Troughline, but changed to a cascode RF amplifier (ECC84) and triode separately-excited mixer (ECC85) with the Troughline II. Leak Troughline experts would no doubt be able to advise whether there was a significant difference between the two when it came to noisiness with very weak signals. The cascode valve was changed to the frame-grid ECC88 for the Troughline Stereo. It is interesting that the sharp cutoff ECC88 was chosen, not the remote cutoff ECC89.

Lowther as far as I know used pentode RF amplifiers and pentode mixers throughout, from Mark I to Mark V, although there is some interpolation in that assumption. But its initial effort, actually a VHF-AM-FM tuner, probably predated the availability of the ECC84 and ECC85, so it would not have had too many options.

Quad used a pentode RF amplifier (initially a 6BJ6, later a 6BH6) with a separately-excited triode mixer (ECC81). But although the production version of its FM tuner dated from 1955, the prototype appeared in 1952, definitely before the ECC84 and ECC85 were available.

Radford was late to the party, with its mid-1960s FMT1. This used an ECC85 front end ahead of a relatively sophisticated four-stage IF strip, which thus might be taken as a vote for the ECC85. But it is also possible that Radford chose to use a third-party front end, in which case the ECC85 could well have been a forced choice. I doubt for example that one could have easily found an off-the-shelf cascode FM front end.

So the empirical evidence of manufacturer choice is very mixed. It does suggest that the in situ differences between pentodes and triodes were not overwhelming. It is clear though that cascode RF amplifiers were quite scarce.

One should perhaps also look at other UK valved FM tuners, such as those from Rogers, and the RCA New Orthophonic. But I have not seen schematics for these.

It might also be observed that those British TV/FM combination receivers in which the TV tuner unit was also used as the FM front end were towards the top of the totem pole in terms of potential at least. They had cascode (or guided grid triode) RF amplifiers, albeit with pentode mixers. And later iterations benefitted from frame grid valves in both positions.

And whilst frame grid valves were quickly adopted for TV (and TV/FM) receivers, I don’t recall that there was a European frame-grid successor to the ECC85. In the American series there was the 6JK8, which had a frame-grid triode for the RF amplifier function, but a conventional triode for the mixer/oscillator section. The irony was that I don’t think there was an American valve quite like the ECC85. Rather the 12AT7 (ECC81) was used for single-valve front ends.

Cheers,

[Synchrodyne]

As previously discussed, the RCA 6X8 is generally regarded as having been the first TV VHF triode-pentode frequency changer valve, and it was closely followed by the Tung-Sol 6U8. Both of these were released in 1951. The first European developed valve of this type looks to have been the Philips ECF80/PCF80 of 1953. The 6U8 was also taken up in Europe, and had the Pro-Electron designation ECF82. Whilst one might have expected Brimar to adopt the 6U8 as such, which valve maker actually obtained the ECF82 designation is unknown.

But a “wild card” here in terms of precedence was the Mazda 10C2, a TV VHF frequency changer triode-pentode on the Rimlock base. It appears to date from 1952, but I cannot find any “hard” confirmation of this. If so, it would have been quite late for a new Rimlock release. And also, as a TV valve with a 100 mA heater, it would not have been well-aligned with emerging practice, which was for TV valves with 300 mA heaters. So it just might have been released earlier than 1952.

Thus the 10C2 can be viewed as having been another Mazda oddity, along with the 6F33 and 6H1 previously reviewed in this thread. And if the 1952 release date is confirmed, it was also leaning towards being anachronistic.

On the other hand it may be noted that Mazda claimed to the first TV VHF frequency changer triode-pentode with remote cutoff characteristics for the pentode, namely the 30C17 of circa 1961.

Cheers,

[SteveCG]

Mention has been made of the Mullard PC97. My only experience of this was its use in a Ferguson 12 inch portable 405 line only TV (I cannot remember the exact model). I was impressed by its performance on the upper UK Band III TV channels. Was there a tuner that used a PC200 - or is my memory paying tricks on me all these years later?

[turretslug]

Something of a tsunami of research and posting there, Synchrodyne- that made for an absorbing read! So, valves were still capable of turning in creditable performance at Band III- good RF gain being achievable, even if noise was becoming significant. Even so, not disastrously bad generally compared to Band II and Band I, one and two octaves down. The sophisticated and elegant approaches seen in turret tuner development must have helped here in keeping stray inductance and capacitance right down when conventional wisdom might have held that switched band-changing at these frequencies was asking for trouble.

It's evident that UHF was another matter, though, for valves, even though the likes of the PC86 and PC88 stretched conventional design and based (i.e. pinned axial plug-in) development as far as possible. Little wonder that, having invested in low-noise, high gain RF amp in the VHF tuner section, this often ended up being used to give the UHF tuner a helping hand as 1st IF.

It occurred to me that military aviation communication would have needed good front-end performance to 400MHz- the 6AK5 has been mentioned already as a significantly good performer, possibly this was the motivation in quoting the use of the 6AG5 at 400MHz, if only on a "what else is there?" reasoning. Presumably there was some precision detailing in the construction of the 6AK5 that ruled it out for significant consumer usage on economic grounds.

[kalee20]

That is a pretty comprehensive run-through of Band 2 receiver front-ends, Synchrodyne.

A couple missing: the Eddystone 820 used the 6AM6 (EF91) as RF amplifier and 12AT7 (ECC81) as mixer with separate oscillator. Presumably with separate RF amplifier, oscillator radiation from the aerial would not have been a problem as the ECC81 does not have the intersection screen that the ECC85 does. This is not that different to the Quad configuration, apart from choice of RF valve.

Battery receivers: The few British sets seem to use the DF97 pentode connected as triodes as self-oscillating mixers, and no RF amplifier. Without any buffering, oscillator radiation across the band must have been a nightmare even allowing for the null-point bridge injection technique! The Continental radios on the other hand often use the DC90 triode, albeit in the same sort of circuit. So why was the DF97 developed when the, presumably cheaper triode, would have worked? And also, why no RF amplifier, unless it was in the interests of battery economy? Yet the Ever Ready Sky Monarch has no fewer than THREE stages of IF amplification. It would surely be better in the interests of noise to redispose one on the IF amplifiers as an RF amplifier preceding the mixer, or was the DF97 out of its depth as an RF amplifier? Yet with its directly-heated cathode, there is at least access to both ends of the cathode, maybe not quite as good as the EF80 twin cathode connections for minimising cathode lead inductance, but surely useable nevertheless, to separate input and output circuitry!

[G8HQP Dave]

To get good performance at high frequencies you need precise element positioning. It is difficult to put a grid near a battery valve filament, as the filament is not sufficiently rigid. I suspect that is why good VHF performance is not really possible for a battery valve. You could make the filament more rigid, but then it would need more heating power.

Another issue is circuit impedance. Battery valves run at lower currents and higher impedances than normal valves, but circuit stray capacitance will be much the same for both. Capacitance can be tuned out but that means more good quality inductors.

[Synchrodyne]

An unusual feature of the DF97 was that it designed to work, inter alia, as a dual-control pentode, that is, as a “multiplicative” mixer with signal on g1 and oscillator on g3. As discussed earlier in this thread, British/European-origin dual-control pentodes were a scarce species, the only other miniature identified as such being the Mazda 6F33. Also, the DF97 was remote cutoff, unusual I think for dual-control pentodes in general.

Originally Posted by Synchrodyne

The irony was that I don’t think there was an American valve quite like the ECC85. Rather the 12AT7 (ECC81) was used for single-valve front ends.

In fact it was not so. In the American series there was at least the 6DT8, which was broadly comparable to the ECC85. In fact the 6DT8 might well have been more-or-less the 12AT7 with an internal screen and sans the split heater. The “6D” number suggests that it dated from the later 1950s. The ECC85 goes back to 1953 or 1954 and was registered as the 6AQ8 in 1955. Also, it is possible that the 6BQ7, and other similar cascode valves, were used in the single-valve FM front end role. The 6BQ7 was symmetrical, and both triodes could be used in either the grounded grid or grounded cathode mode.

The early miniature VHF pentodes, EF42, EF80 and EF91 were also used in FM receiver IF strips. I surmise that this was so because although the early miniature HF pentodes, such as the EF41, 6F15 and W77 (EF92) were specified to work at up to 30 MHz and perhaps beyond, they may have been less than ideal as relatively wideband 10.7 MHz FM IF amplifiers. This group of HF pentodes had slopes in the 2 to 2.5 mA/V range, generally comparable with those of their non-miniature predecessors. Philips Rimlock literature nominated the EF42, not the EF41 for the FM IF amplifier application.

In contrast, the first American miniature HF pentodes intended for domestic receiver applications, the 6BA6 (remote cutoff) and 6AU6 (sharp cutoff), were, according to RCA, designed with higher slopes to better suit them for FM as well as AM applications. Whilst they were stretched at 100 MHz, they were evidently fine in wideband 10.7 MHz IF strips. The 6BA6 and 6AU6 remained the norm for FM/AM and FM-only IF strips in American practice right through until the end of the valve era.

In its initial range of FM/AM receiver valves, Philips nominated the EF85 for the combination FM/AM IF amplifier job. The EF85 was from the TV series, being the remote cutoff counterpart to the EF80. But as already noted, it was not ideal for the purpose, and so the EF89 was developed. This was an improved EF41, and was also proximate to the 6BA6, perhaps slightly better than it. Not only did the EF89 displace the EF85 in FM/AM receiver IF strips, it was also used in FM-only IF strips, typically in combination with an EF80 as the final limiter, for which sharp cutoff valves were usually preferred.

One example was the Pye Mozart FM tuner, which had an EF89 1st IF stage, an EF80 amplifier/limiter and an EF80 limiter. Another was the 1960s Radford FMT1, which had three EF89s, the 2nd and 3rd as amplifier/limiters, followed by an EF80. So one might infer that the EF89 was preferred over the EF85 “VHF” valve in the FM IF role, and was seen as being suitable for amplifier/limiter applications.

The Heathkit FM-4U, from the early 1960s I think, had an EF89, EF89, EF80 IF line-up, following an ECC85-based front end. The 2nd EF89 was configured as an amplifier/limiter. Now the FM-4U might well have been the UK version of the American FM-4, so a comparison is worthwhile. The FM-4 had a similar front end, two-gang permeability tuned, but using a 6DT8 rather than an ECC85, and also with afc via a capacity diode, something that the FM-4U lacked. The FM-4 IF strip had a 3 x 6AU6 valve line-up, configured respectively as amplifier, amplifier/limiter and limiter. Assuming that this was the prototype, then one may analyze the “translation” to the UK valve line-up. Although the 6AU6 was certainly available in Europe (and had Pro-Electron designation EF94), there was not a comparable valve in the European series. It would appear that the EF89 was the closest match where amplification was required, and quite possibly this was notwithstanding that it was of the remote cutoff type, rather than because of it. On the other hand, for the final limiter, sharp cutoff was required, and so it was necessary to use a “VHF” valve, namely the EF80. The FM-4 used a 6BN8 double-diode-triode as ratio demodulator and cathode follower output. This kind of double-diode-triode was not found in the European series, so the FM-4U had an EB91 ratio demodulator and omitted the cathode follower output. But it did have a magic eye (EM84), something that the FM-4 lacked.

Given how often the 6AU6 was found in a diverse range of roles, it is perhaps surprising that there was not a sharp cutoff HF (in contradistinction to VHF) pentode in the mainstream European range. But where such was required, the options seem to be either use a VHF sharp cutoff pentode or if possible an otherwise suitable HF remote cutoff pentode. That may explain why the EF91/Z77/6AM6 was also found in diverse roles that on the face of it would not have required a VHF pentode, such as calibration crystal oscillators in HF receivers. For example, Eddystone practice varied amongst its receivers, and both the 6AU6 and EF91 were used for this purpose.

There was though the Mazda 6F11, in the Rimlock series, and evidently the sharp cutoff counterpart to the remote cutoff 6F15. The 6F11 had a slope of 2.2 mA/V, very close to the 2.3 mA/V of the 6F15. As best I can ascertain, the 6F11 did not have a Pro-Electron designation. I think that it may be added to the list of Mazda oddities, along with the 6F33, 6H1 and 10C2. The 6F18 was Mazda’s offering in the EF89 class. It had occurred to me that this too might have had a sharp cutoff counterpart, but if so, I cannot find any trace of it.

Cheers,

[Synchrodyne]

Regarding American use of one-valve FM front ends and its effect on valve developments, it would appear that unlike the European case, this type of FM front end did not come into widespread use in the USA until the later 1950s, so in this case American practice trailed European practice.

At least this is the impression I get from an article “Don’t Dodge the One-Tube FM Front End” in Electronics World (EW) magazine for 1959 August, p.52ff. That the species was something of a European import may be gleaned from the fact that the specific version discussed was as used in a Telefunken receiver, and it was also noted as being included in Grundig equipment. At that time General Instrument Corporation was reported as making a similar unit in the USA.

An article in Radio-Electronics magazine for 1960 March, p.38ff, describes a new miniature single-valve FM front end from Standard Coil, using a 6AQ8 or 6DT8 valve. One might say that Standard Coil’s entry into this field meant that it had really “arrived”.

In respect of valves, the Telefunken unit was described as using a “foreign” twin triode, the ECC85, local equivalent 6AQ8. But it was also noted in the EW article that TV twin-triode cascode valves could be and had been used for the purpose, including the 6BQ7, 6BZ7 and 6DT8. Actually the 6DT8 was purpose-designed for FM front ends, although it might well have been used as a cascode in TV front ends, as well. Anyway, it is evident that the 6DT8 was extant by mid-1959, and it is a reasonable inference that it did not have an American predecessor that was a specialized multifunction FM front end valve. So the American branch of this family tree starts with the 6DT8.

A successor to the 6DT8 already mentioned upthread was the 6JK8, whose main advance was the use of a frame-grid triode for the RF amplifier. Another successor, along a different vector, was the 6GY8 triple triode, designed to do the RF amplifier, autodyne mixer and afc reactance valve jobs in one envelope. It had a different pinout to the 6EZ8 triple triode already mentioned. The 6GY8 had separate cathode connections for each section, although one was also tied to the grid of another triode and also to one side of the heater. The 6EZ8 had two of the cathodes commoned and then tied to one side of the heater. Suggested uses for it were mixer, oscillator and afc, and RF amplifier, mixer and oscillator.

From these developments one may get the impression that the inclusion of afc was important for even lower-cost American FM receivers, and that the 6GY8 was intended to accommodate this without stepping outside of the one-valve front end concept. Interesting here was that there was still a requirement for afc using reactance valves even though by the late 1950s, suitable variable capacity diodes were available. The renaissance of afc (aft) in American TV receiver practice is probably marked by the Westinghouse system of 1958, as described in RE 1958 February, p.56ff. Here a diode was used, possibly because it would not have been easy to add a reactance valve to a typical VHF TV tuner. And I think that this set the pattern.

I don’t know for sure, but possibly reactance valves inherently provided higher afc loop gain than diodes, thus perhaps making them preferable where the afc control voltage source was a ratio demodulator or wideband ratio demodulator with relatively low DC output per unit frequency deviation. The Westinghouse TV afc system was quite complex, and the method of deriving the control voltage probably provided reasonable gain. Also, a reactance valve would work with a zero-centred control voltage on its grid, whereas a diode would require an offset afc control swing, all above or below zero. Practical Television magazine for 1962 December, p.123ff, in an article about intercarrier sound, included a block schematic of what was said to be a typical “Continental” TV receiver, with diode-based afc for both its VHF and UHF tuners. The afc control voltage was derived from a circuit consisting of half a PCF80 as a 4th IF amplifier (presumably to centre the IF response over 38.9 MHz as well as provide additional gain), a two-diode (OA79) discriminator and the other half of the PCF80 as a DC amplifier (presumably providing the required offset as well as the DC gain). That lends some credence to the notion that with diode afc, some DC gain, as well as offsetting was needed following the discriminator.

Whatever were the reasons to retain reactance valve afc, they were cogent enough to develop of ad hoc triple triodes for FM applications in the early variable capacity diode era, and at a time when the inexorable move to transistors was under way. It was evidently an application where valve technology was still more cost effective than the solid state. Circa 1960, an afc diode driven by a transistor DC amplifier (as used by Jason in its FMT4 tuner) might have been enough more costly than an extra triode in an existing valve envelope when it came to consumer receivers as distinct from hi-fi equipment.

The above-mentioned magazine articles are all available at AmericanRadioHistory.com; http://www.americanradiohistory.com/index.htm.

Cheers,

[Synchrodyne]

From the thread “GEC BC5243 Missing Valves”, https://www.vintage-radio.net/forum/...d.php?t=116851

Originally Posted by G6Tanuki

Yep - 9BW6 is a 300mA-heater one. I'm thinking US TV sets here. No doubt there's a 12BW6 and an 18BW6 as well, just as there are 12.6V and 25.2V versions of the 807/6L6/6146 etc.

The 6BW6 was as far as I know a Brimar own development, not an American valve. Brimar did kind of thing this when European setmaker requirements were not met from the standard American range. From circa 1950 anyway, it registered its own valves in the American series, though, which is apt to cause confusion. The American setmakers at the time seemed content to use either the 6AQ5 or the 6V6-GT.

And the 9BW6, for 300 mA series heater chains, appears to have been the only heater variant of the 6BW6. Perhaps surprisingly, there was not a 150 mA version, even though Brimar had followed American practice in offering 150 mA valves for AC-DC radio receivers rather than the 100 mA type commonly used in Europe. Another Brimar own-development to address European preferences was the 12AH8 triode-heptode. This had a split heater, so could be used in both AC (6.3 V) and AC-DC (150 mA) applications.

Early in 1953 though, RCA introduced the 12V6-GT, with 12.6 V heater. This was intended for car radio applications, and I’d guess its introduction more-or-less corresponded with the period when the US auto makers were abandoning 6-volt electrical systems in favour of 12 volts. In another example, the 12BF6 had been released at the end of 1951. Of course, many of the 150 mA heater small-signal valves, such as the 12AU6, 12BA6 and 12BE6 happened to have 12.6 volt heaters, so much of the required 12.6-volt range was already in place.

Then towards the end of 1955, RCA introduced the 12AB5, which was a noval-based version of the 12V6-GT, and also intended for car radio applications. Thus one might view the 12AB5 as being the same as would have been a notional 12BW6.

The 12L6-GT was part of a major 1954 second-half industry release of valves suitable for 600 mA series chains in TV receivers. The choice of 600 mA apparently allowed the use of a single-string string with the American 115-volt (nominal) power supply. The new 600 mA valves also had controlled warm-up time to minimize overvoltage due to differential warm-up times. The initial release included one or two octal types as well as many miniatures, both B7G and noval. Many 6.3-volt TV valves were reissued in 600 mA form, and as well, the native versions were often also reissued with controlled warm-up heaters, usually distinguished by an “A” or “B” suffix. The 12L6-GT was, heater aside, similar to the 25L6-GT (300 mA) and 50L6-GT (150 mA), but not apparently similar to the 6L6-G.

By the end of 1956 there was also a series of TV valves with 450 mA heaters. Quite what the rationale for this was I don’t know, and I don’t think that it was ever as extensive as the 600 mA series. On the other hand, some existing TV valves with 6.3 V, 450 mA heaters, such as the 6U8 and 6X8 mixer-oscillator triode-pentodes, fitted right into it. Interestingly, in mid-1956 RCA had released the 5CG8 (600 mA) and 6CG8 (6.3 V) triode-pentodes as improved versions of the 5X8 and 6X8 respectively, in particular having duplicated cathode pinouts. Although the 6CG8 was 6.3 V, 450 mA, no mention was made of its suitability for 450 mA heater strings. Thus one might infer that the timing of the 6CG8 was on the cusp of the inception of 450 mA series strings.

That though was not the final word in American series-string valve heater currents. Surprising (to me, anyway), quite late on, and as reported in Radio-Electronics 1960 December, was a set of 100 mA valves for AC-DC radio receivers. It included the 18FX6 heptode, confirming the American love-affair with the “pentagrid converter”.

Cheers,

[G8HQP Dave]

Originally Posted by Synchrodyne

And the 9BW6, for 300 mA series heater chains, appears to have been the only heater variant of the 6BW6.

I believe I have some 19BW6 in my valve collection - 100mA heater.

[Synchrodyne]

Interesting that; possibly the 19AW6 was made by other than Brimar, given that it [Brimar] was nominally at least in the 150 mA camp not the 100 mA camp when it came to valves for AC-DC radio receivers.

As well as the 12V6-GT and 12BF6 mentioned in my previous post, the release of 12.6-volt heater car radio valves in this period also included the 12AQ5 and 12X4. That both the 12V6-GT and the 12AQ5 were included is interesting. Quite possibly some car radio makers had been using the more robust 6V6-GT output valve, along with miniature signal valves, because they were not too happy with the relatively highly-stressed 6AQ5. With the conversion from 6- to 12-volts, drop-in replacements were required, so the valve-makers needed to cater for those in the 6V6-GT camp as well as those using the 6AQ5.

Then one might assume that the 12AB5 was developed to meet the needs of car radio makers who wanted a miniature output valve that matched the 6V6-GT, but which was more robust then the 12AQ5. Hence the use of the larger envelope associated with the noval base. The 12AB5 appeared about a year before the first crop of American 12-volt HT radio valves, which also had 12.6-volt heaters.

One may wonder why a separate 12.6-volt heater series was developed for 12-volt car radios, as opposed to adopting the European practice of using series-connected pairs of 6.3-volt valves. Possibly it was a combination of happenstance, in that part of the series already existed in the form of 150 mA valves that happened to have 12.6-volt heaters, and the relatively large scope of the American car radio market. The European approach did require some alignment of heater currents, though. The EF41 and EL42 were both 200 mA, and the ECH42 and EBC41 were both 230 mA. The EZ41 was 400 mA, but it could have been used in series with a pair of 200 mA valves in parallel. The symmetry was broken somewhat when the ECH81 (300 mA) arrived, but then this could have been paired with the existent EBF80, and later the EBF89. This was the pattern carried over to the 12-volt HT valve set, which included the ECH83 and EBF83. The other members of the initial series, EF97 and EF98, were also 300 mA.

More on the multiplicity of American heater variations, the 5V6-GTand 5AQ5 were also issued for use in 600 mA TV heater strings. Whilst their parent versions were 450 mA, evidently “A”-suffixed versions (controlled heater warm-up times) were issued for use in 450 mA TV heater strings.

Perhaps the 6V6 would win the prize for the having most heater and envelope variations?

Cheers,

[G8HQP Dave]

Originally Posted by G8HQP Dave

I believe I have some 19BW6 in my valve collection - 100mA heater.

Sorry, my mistake - 19BW6 is 150mA heater.

Originally Posted by Synchrodyne

Perhaps the 6V6 would win the prize for the having most heater and envelope variations?

The 6Q7/EBC90 etc. would not be far behind?

[G6Tanuki]

Perhaps the 6V6 would win the prize for the having most heater and envelope variations?[/QUOTE]

I'm eternally surprised that nobody ever did a 6V6-with-top-cap-anode.