Valve Questions

[Synchrodyne]

Originally Posted by G8HQP Dave

Heathkit made a habit of using sharp cutoff valves where everyone else would use a remote cutoff type. Perhaps an individual designer with a bee in his bonnet? Of course, if AGC is not applied then the sharp cutoff is better but Heathkit often applied AGC to them.

In this case the Heathkit Mohawk did not have AGC applied to the two 6CS6 mixers. In fact AGC was applied only to the 6BZ6 RF amplifier and the 6BA6 1st IF amplifier. But the same agc line was used for both, one, a 6BZ6 of the semi-remote cutoff type and the other a 6BA6 of the (full) remote cutoff type. But Heathkit was not alone in mixing valve types on the same AGC line. For example the Eddystone 830/7 had its sharp cutoff 6AK5 1st mixer coupled to the same AGC line as were the remote cutoff ECC189 RF amplifier, ECH81 2nd mixer, and 6BA6 1st and 2nd IF amplifiers. That seems quite contrary to TV practice, where the mixing of sharp and remote cutoff valves required careful grading of the respective AGC voltages and appropriate selection of the delays, rather than simple coupling to the one agc line. Anyway, that’s a really separate topic.

Back to valves per se, the 6BZ6 was developed for agc’d TV IF applications circa 1954. I am not sure, but it might have been the first receiving miniature of the semi-remote cutoff type. The concept was not new, though, and goes back at least to the 6SG7 and 7H7. Of those, which had high slopes for their time, it was said that they had a cutoff as remote as was consistent with maintaining a good slope-to-anode current ratio, which I imagine meant not as remote as conventional remote cutoff valves. It would appear that RCA did not deem it necessary to include a semi-remote cutoff pentode in its initial post-WWI miniature range. American TV receiver practice post-WWII sad been to use AGC with sharp cutoff valves, such as the 6AG5 and its successors. The argument for applying AGC to sharp cutoff valves had been that cross-modulation effects were of lower significance for TV than for AM radio, so the technique was worth using because the lower magnitude AGC voltage required for sharp cutoff valves reduced the need for AGC system gain.

But the early RF cascode valves, such as the 6BQ7 and 6BK7, although comprised of sharp cutoff triode units, had somewhat remote cutoff characteristics when used in typical series-cascode circuits. So perhaps the 6BZ6 was intended to match the cascode valve curves, requiring about the same magnitude of AGC voltage. Also, the advent of colour TV with its in-band subcarrier may have placed more emphasis on cross-modulation performance, indicating a need to move to remote cutoff or at least semi-remote cutoff valves for agc’d stages. Prior to the arrival of the 6BZ6, improved sharp cutoff valves had been introduced for the TV IF application. The 6BC5 had been a somewhat improved 6AG5, and the 6CB6 was associated with the move to “high” IFs.

The 6BZ6 and 12BZ6 were also liked by some of the HF receiver makers in the RF amplifierposition, as shown by the subject Heathkit example. In respect of its NC-300 amateur receiver, National said that for the frequencies covered, it was not necessary to use a triode input stage, such as the cascode, to achieve a low noise figure. Laboratory tests showed that the 6BZ6 pentode r.f. amplifier in the NC-300 yielded a low noise figure as a result of the careful design of the input transformer. Typical results were 4 db at 20 metres and 5 db at 10 metres.

Cheers,

[Synchrodyne]

The agc question and suitability of valves therefor gets even more complicated. I turns out that there was the 6CF6, which was described thus in a GE data sheet: “The 6CF6 is a miniature sharp-cutoff pentode designed especially for use as a gain-controlled intermediate frequency amplifier in television receivers. Except for the plate current cutoff characteristic, the 6CG6 is identical to the 6BC6; and like that tube features high transconductance and low interelectrode capacitances.”

An RCA data sheet said more-or-less the same, albeit with some quantification: “This type is identical with miniature type 6CB6 except that the grid-No.1 bias (approx.) for a plate current of 35 microamperes is -6.5 volts.” For the 6CB6, RCA quoted a grid bias of -8 volts for an anode current of 10 µA. Nominal anode current was 9.5 mA.

So it looks as if the 6CF6 had a faster drop in anode current than the 6CB6 as the grid bias became more negative. Thus more gain reduction was available before the valve reached the sharp bend in the curve just before cutoff.

One may surmise that the 6CF6 was the response to an apparent need for TV IF valves with better behaviour (lower cross-modulation) than the 6AG5 and 6CB6 when operating under high (magnitude) agc bias conditions, whilst not departing from the sharp cutoff concept with its relatively modest agc voltage requirement.

The 6CF6 appears to have predated the 6BZ6, although not by much. Possibly the 6CF6 was not enough better than the 6CB6, hence the need for the next step. Or perhaps the 6CB6 and 6BZ6 were independent and quasi-parallel solutions to the putative problem.

Cheers,

[Radio Wrangler]

It's alright judging these valves in terms of the gain reduction they can achieve, but that can also be done with a short cut-off valve. The reason for the longer curves with their distant cutoff is so that the gain range can be achieved while distorting a typical level signal less, by having less curvature being swept by the signal. The harmonics generated by the curve will be far away in frequency (except for wideband applications like audio gain control like tape recorder automatic level circuits) but off-order components to the equation of the curve will also create odd-order intermodulation products where there are two or more signals present.

Odd order intermods produce both distant and nearby frequency components, and it's the nearby ones which cause trouble. By the time the miniature valves were being developed, most bands were getting pretty full.

David

[G8HQP Dave]

Yes, the ideal characteristic for a wide range AGC controlled device is exponential (sometimes called logarithmic!). This means that signal handling (i.e. linearity) does not deteriorate as gain is reduced. Good remote cutoff valves achieve a reasonable approximation to this.

An exponential device is characterised by a voltage, which sets how quickly the gain (and current) changes with bias (and signal level). BJTs are extremely good approximations to exponential, but unfortunately the relevant voltage is far too small (10's of mV) so they are awful at conventional AGC. A remote cutoff pentode will have several volts so can handle larger signals. A useful figure-of-merit is Ia/gm - for a perfect exponential device this gives the voltage, and remains constant with gain reduction.

Note that one problem with using the active amplifying devices for gain control is that linearity does not improve as gain is reduced for larger signals - the best you can have is that linearity doesn't get worse. If you need improved signal handling for lower gain then you have to use things like PIN diode attenuators instead.

[Synchrodyne]

Against that background, the 6CF6 may be seen as not an attempt to produce an excellent agc valve, but as an effort to obtain some improvement over the 6CB6 whilst retaining its sharp cutoff characteristic. It’s the kind of development for which its designers might have written an article or paper about how that was done, inclusive of a rationale for staying with sharp cutoff in the face of both the conventional wisdom and the implied evidence that the 6CB6 (and like valves) were not good enough.

Originally Posted by G8HQP Dave

A useful figure-of-merit is Ia/gm - for a perfect exponential device this gives the voltage, and remains constant with gain reduction.

Aha, reasonably constant Ia/gm might have been a design basis for the 6BZ6, with the trade-off being some sacrifice in the remoteness of the cutoff. If so, the 6BZ6 might have had better linearity retention with gain reduction than say the 6BA6 or 6BJ6, which could explain its attractiveness to the HF receiver makers for use in the RF amplifier position – provided at least that it was not expected to work with the huge agc voltage magnitudes that a 6BA6 could handle.

Cheers,

[Synchrodyne]

Mentioned previously was that the initial post-WWII release of American miniature receiving valves, 6BA6, 6BE6, etc. was intended for use in FM and FM-AM receivers as well as AM-only receivers. In particular the use of the 6BE6/12BE6 heptode as a mixer-oscillator for FM seems counter-intuitive when one takes as the norm early British FM receiver practice, by which time triodes and pentodes were used for this job.

Apparently though one of the reasons for so doing might have been to facilitate the use of the same valve set for both the FM and AM front ends. This comes out in an article in Radio & Television News for 1948 November, p.62ff; “An A.C.-D.C. Receiver for AM and FM”, written by two RCA staffers (http://www.americanradiohistory.com/...Page_Guide.htm).

This circuit used a 6BJ6 as RF amplifier for both FM and AM, and a 12BE6 as mixer-oscillator for both FM and AM. Another 6BJ6 served as 1st FM IF amplifier and AM IF amplifier, and yet another as 2nd FM IF amplifier. A 6AQ6 did the AM demodulation and agc rectification job as well as providing the AF voltage amplifier, a 12AL5 served as FM ratio demodulator and a 35B5 provided the audio output. The rectifier was a 35W4.

Note that the 6BJ6 rather than the 6BA6 was chosen for the RF and IF amplifiers. When it was released in 1947, the 6BJ6 was described as being particularly suitable for mobile equipment where low heater current was an important consideration. In this specific application, RCA said: “For the r.f. and i.f. stages type 6BJ6, a remote cut-off pentode with six-volt 150 milliampere heater, is used. Although similar to type 6BA6 in many respects, the 6BJ6 has a higher input resistance and a lower input capacitance than the 6BA6. In addition, the difference in basing results in lower feedback admittance for the 6BJ6 at high frequencies.” So even though the 6BA6 was designed with FM applications in mind, in short order a valve that was better for that purpose was released. The choice of a 6BJ6 for the non-agc’d 2nd FM IF stage, rather than say a 12AU6 or 6BH6 that might have been expected in this position may have been to avoid proliferation of valve types. Also, in the case of the 6BH6, although it was the sharp cutoff counterpart to the 6BJ6, it was I think a Tung-Sol not an RCA development. RCA may well have preferred to use only valves that it had developed. That might also explain the use of the 6AQ6 and the 12AL5 rather than say the GE-developed 19T8 in place of both.

On the face of it, it seems odd that whilst RCA chose to use an improved pentode, it stayed with the “original” 12BE6 heptode, rather than using an improved version. I think that this may be explained by the fact that using an all-miniature valve line-up was another objective, and this would have outruled the 6SB7. And the design was probably started before the 6BA7 was released.

An interesting sidebar is that the design followed American AC-DC practice in having a non-live chassis and a “B- lead”, which evidently required extensive and careful decoupling to avoid problems at FM frequencies.

RCA retained the possibility of FM-AM front-end commonality at the next step, as well. Mentioned previously in this thread was that its 1951 6X8 triode-pentode VHF TV mixer-oscillator was also suitable for use for FM and AM frequency changing. I had dismissed the AM case as unlikely – why use a triode-pentode rather than a heptode or triode-heptode - but it makes more sense in the light of the FM-AM commonality requirement. And I have since found this commentary from RCA in respect of the 6X8: “Miniature type used as combined oscillator and mixer tube in television receivers utilizing an intermediate frequency in the order of 40 megacycles per second. In such service, the 6X8 gives performance comparable to that obtainable with a 6AG5 mixer and an oscillator consisting of one unit of a type 6J6. When used in an AM/FM receiver, the triode unit is used as an oscillator for both sections. In the AM section, the pentode unit is used as a high-gain pentode mixer; in the FM section, the pentode unit is used either as a pentode mixer or as a triode-connected mixer depending on signal-to-noise considerations.”

One might say that whereas the 6BE6 and 6BA7 were good AM mixer-oscillators and passable FM mixer-oscillators, the 6X8 was a very good FM mixer-oscillator and a passable AM mixer-oscillator.

That RCA was serious about the FM-AM application was shown by the fact that in 1952 it released a 19X8 variant, with 150 mA heater, and so clearly an AC-DC radio valve, and described as such: “RCA-19X8 is a multi-unit tube of the 9-pin miniature type containing a medium-mu triode and a sharp-cutoff pentode in one envelope. It is designed primarily for use as a combined oscillator and mixer tube in "transformerless" AM/FM receivers. The 19X8 has a 150-milliampere heater which permits series-string heater operation with other tubes having 150-milliamperes heaters.”

A curiosity is that RCA made exactly the same statement in respect of the 6U8 triode-pentode as it did for the 6X8, as quoted above. The 6U8 was a Tung-Sol development that closely followed the 6X8. It differed in having separate cathodes, and an internal cathode connection for the pentode section, whereas the 6X8 had a common cathode and an independent suppressor grid. As far as I know, Tung-Sol did not mention the FM-AM application, and there was not a “19U8” variant. So RCA may have been alone in offering the 6U8 - or indeed any TV mixer-oscillator triode-pentode - for FM-AM radio applications. Whilst Philips/Mullard presented the ECF80/PCF80 as having multiple applications, I don't think that FM-AM receivers were mentioned.

Cheers,

[Radio Wrangler]

The information which remains of all these devices consists of the data sheets and advertisements. The datasheets say what the device IS but don't really explain what was going through their designer's head. The adverts and press releases were written by marketing departments and the quality of their information varied widely, sometimes some insight was imparted, other times you could question whether the guy who wrote it was even living on the same planet as the design team. Finding technical papers written by the designers would be a great help in understanding what was going on and what the intent was.

It's a very unfortunate effect of variable-slope techniques for applying AGC that when there are big signals around and the slope is backed-off, so is the anode (/collector/drain) current, thus reducing the available power at the output. So the dynamic range of a set is reduced in the presence of larger signals, and is at its best when it isn't needed so much.

The same applies to long-tailed pair balanced amplifiers when the tail bias current is varied to change gain. And the same happens in Gilbert-cell analogue multiplier trees when the voltage difference in the bases of the upper pairs are varied to divert the collector currents of the lower devices into dump outputs. And the same happens in dual gate mosfets. Whenever the signal current is backed off by AGC, so is the quiescent current.

As said by Dave, above, PIN diodes are the stars here. Unfortunately I can't always use PINs because of the DC power consumption. PIN diodes also have a parameter called 'Carrier Lifetime' which sets a frequency above which they operate mostly as current-controlled variable resistors, below which they operate as diodes and start rectifying signals. This makes them only useful in the several MHz and upwards region.

A lot of Japanese amateur radio transceivers used plain diodes to switch front-end filters and to switch crystal filters in the IF. They are tens of pence cheaper than PIN types, and there are no moving parts. The ICOM IC765 is a set with a good reputation for performance in its era (There are plenty of reviews on the net and if you filter out the ones with 'awsome' or 'sucks' in their text to leave the more considered reviews, the reputation seems to have some basis). It uses plain silicon diodes, not PINs, and relies on the DC bias being a lot larger than most signals. Ulrich Rohde wrote an article on the benefit of re-stuffing the set with PIN diodes. They offer a useful increase in the signal handling per milliamp of DC consumed.

Icom later decided to go the whole hog and the recent IC7800 and 7700 are stuffed full of miniature relays. Hundreds of the things. Even the front-end preselector is tuned with relay-switched Ls and Cs, and the same goes for the auto ATU. You want to hear the unified clonk when you switch the power on or off. Fortunately they are all oriented in various directions, otherwise the moving mass would make the whole set jump to the left or right on your bench

In a very modern set, with its channel-defining filter implemented digitally, it must handle large, adjacent signals reaching right down its structure to the ADC and beyond.

Yipes! I'm sitting in the radio shack, radios off, and I just heard a sharp snap as the HF atu flashed over internally and there was a peal of thunder. I'm switching things off and isolating the Apple. 2 seconds is too close for comfort. Bye.

David

[Radio Wrangler]

It seems to have passed over. I'm in a low risk area for lightning, but maybe I need to plan some proper protection though no-one else round here seems to bother.

David

[G8HQP Dave]

Originally Posted by Synchrodyne

Aha, reasonably constant Ia/gm might have been a design basis for the 6BZ6, with the trade-off being some sacrifice in the remoteness of the cutoff. If so, the 6BZ6 might have had better linearity retention with gain reduction than say the 6BA6 or 6BJ6, which could explain its attractiveness to the HF receiver makers for use in the RF amplifier position – provided at least that it was not expected to work with the huge agc voltage magnitudes that a 6BA6 could handle.

Yes, hopefully later remote and semi-remote cutoff valves had nearer to ideal characteristics than the earlier ones. However, I suspect that the 6BZ6 was widely used by (amateur) HF receiver manufacturers for two reasons:
1. they didn't realise how much front end noise comes from the tuned circuit rather than the valve,
2. everyone else was using 6BZ6 and saying so in their marketing.
Some receivers used a 6BZ6 but ran it at a low current - presumably to avoid overloading the mixer. Much better (technically) to use a lower gain device but commercially bad when your potential customers 'know' that a 6BZ6 is a superior device.

The fact that a 6BA6 can handle huge AGC voltages is a hint that it can also handle quite large signal voltages too. There was a craze for 'upgrading' a receiver by swapping the RF valve for a 'hotter' one (often 6BZ6). The most likely result was a tiny improvement in noise figure and a big reduction in dynamic range and AGC performance. The extra 'band noise' (assumed to be due to better sensitivity) would often have been mainly extra intermodulation generated in a mixer which is now struggling to cope with the extra gain in front of it.

[Synchrodyne]

Originally Posted by G8HQP Dave

However, I suspect that the 6BZ6 was widely used by (amateur) HF receiver manufacturers for two reasons:
1. they didn't realise how much front end noise comes from the tuned circuit rather than the valve,
2. everyone else was using 6BZ6 and saying so in their marketing.
Some receivers used a 6BZ6 but ran it at a low current - presumably to avoid overloading the mixer. Much better (technically) to use a lower gain device but commercially bad when your potential customers 'know' that a 6BZ6 is a superior device.

That was something of a paradox. The typical customers for HF/amateur receivers were more knowledgeable than most of those for domestic receivers. Thus for the latter, you were free to choose an appropriate RF amplifier valve – be it 6BA6, 6BD6 or other - on its in situ merits. But for the former, judgement might well be made on the perceived virtues of the valve itself rather than its in situ performance. That is apparent from an article in Popular Electronics for 1960 March, “Noise in R.F. Amplifier Tubes”, p.95ff. (See: http://www.americanradiohistory.com/...nics-Guide.htm.)

The author of that article was persuaded to try a 6DE6 (sharp cutoff), a reputedly “hot” valve, in place of the 6BZ6 in his Heathkit Mohawk receiver. The result – hardly unexpected – was that the receiver then suffered from front-end overload.

A much earlier article, in Radio News 1947 April, p.47ff, described, for home construction a high-fidelity AM receiver of the TRF type. (See: http://www.americanradiohistory.com/...Page_Guide.htm). In that case the 6BD6 was chosen for both of the RF stages. The 6BD6 was the lesser-known and lower slope companion to the 6BA6, slope-wise similar to the say the W77 and EF41. The comment was made: “We could choose several other higher gain tubes, but they are unnecessary and their tendency toward oscillation is too great. In the interest of maximum sensitivity, we might have chosen the 6AK5 which has a mutual conductance of twice that of the 6BD6 and, by using a series connected screen resistor, imparted to it semiremote cutoff characteristics.” One has the sense that the author was heading off expected questions about the choice of an apparently “unexciting” valve choice.

Cheers,

[Synchrodyne]

Originally Posted by Radio Wrangler

It's a very unfortunate effect of variable-slope techniques for applying AGC that when there are big signals around and the slope is backed-off, so is the anode (/collector/drain) current, thus reducing the available power at the output. So the dynamic range of a set is reduced in the presence of larger signals, and is at its best when it isn't needed so much.

But it would appear that back in the valve era, designers had little choice but to use conventional variable slope agc, at least in domestic equipment. The problem of reduced anode current at reduced gain would still seem to be there even if unconventional techniques were used, such as applying agc bias to grid 3 of a dual-control pentode or a heptode, or using a sheet beam valve (e.g. 6AR8) with the agc bias swinging the electron stream away from the signal anode.

Returning to the question of the slope and cutoff characteristics of American TV IF valves, and empirical look thereat, a short article in “Radio-Electronics” for 1961 February 1961, p.61ff, about substitution of IF valves, (http://www.americanradiohistory.com/...aster_Page.htm) included a graph showing the slopes of then-popular TV IF amplifier valves, namely the 6BZ6, 6DC6, 6DE6, 6CB6, 6CF6 and 6BC5. I have attached a copy of the pertinent page.

This showed quite a diversity of curves, and even significant spreads within the semi-remote cutoff (6BZ6 and 6DC6) and sharp cutoff (others) categories. Both categories were used in agc’d vision IF amplifiers, and the choice between them, and of individual valve type, seemed to be very much up to the setmakers.

It turns out that the 6CF6 (sharp cutoff) and 6DC6 (semi-remote cutoff) were actually selected versions of the 6CF6, the selections being based upon cutoff properties. So the 6CF6 must have had quite a production spread.

The 6DE6 might also have been a selected 6CF6. Of it, RCA said:

“Designed especially for use in the gain-controlled video i.f. stages of television receivers using a 40-mc i.f., it is also well suited for use as an r.f. amplifier in v.h.f. TV tuners.

“The 6DE6 features a controlled grid-1 voltage of -5.5 for a transconductance of 600 micromhos minimum. This cutoff characteristic makes an a.g.c. amplifier unnecessary in certain TV receiver designs and minimizes over-load distortion and cross-modulation effects in if, stages.”

At about the same time, it was saying of the 3BZ6, the 600 mA series-string version of the 6BZ6:

“...intended for use particularly in the gain-controlled picture i-f stages of television receivers. The semiremote cutoff of the 3BZ6 reduces cross-modulation effects in the picture i-f stages, and minimizes distortion resulting from high signal levels and automatic gain control time delay.”

So RCA – and I think the US valve industry generally - was backing both the sharp cutoff and semi-remote cutoff valve approaches to obtaining good agc performance, notwithstanding that theory favoured the latter of the two.

“RCA Engineer” for 1955 October-November, p.18ff, carried an article “Some Problems Associated with Television Circuits”. (http://www.americanradiohistory.com/..._Issue_Key.htm) One of the issues addressed was getting the RF agc magnitude and delay correct relative to the IF agc. It did not discuss individual valve types, but in a general sense covered various combinations of sharp and remote cutoff valves, without favouring any of these. A copy of the page with the graphical treatment of these combinations is attached. The article did include the comment: “The largest production variable that affects r-f gain is variation in tube cutoff.”

History seemed to repeat itself, though. In 1963 the 6JH6 was introduced for agc’d TV IF stages. From an article in Radio-Electronics for 1963 December:

“Similar to the 6BZ6 in most ways, it has a tightly controlled cutoff point intended to provide headache-free interchangeability with others of the same type. According to RCA engineers, there were problems with certain 6BZ6's in the agc'ed first i.f. stage, which is critical about its bias voltage. The new 6JH6 is said to introduce less cross-modulation distortion in areas where channels 6 and 8 are strong."

Cheers,

[Synchrodyne]

Originally Posted by Synchrodyne

I haven’t fathomed the specific utility of the 6BF6 double diode-triode, with its 16 mu triode, as compared with the 6AT6, with its 70 mu triode.

That is answered in this RCA advertisement for its car radio miniature valves in “Tele-Tech” for 1947 February, attached, in which was included the comment:

“It is expected that the 6AT6 will be used in sets having a single 6AQ5 output tube, while the 6BF6 will be preferred as a driver for two 6AQ5's in push-pull, class AB operation.”

This configuration must have been specific to car radios, as there was not initially a 12BF6 for AC-DC 150 mA series-string home receivers. Rather the 12BF6 was added late in 1952 for 12-volt car radio applications.

It does seem that there were radio double diode-triodes for all conceivable applications.

At the end of 1949, Sylvania introduced another series that paralleled the existing 6AT6, 6AV6 and 6BF6, except that they had high-perveance rather than medium-perveance diodes. They were:

6BT6 and 12BT6, with 70 mu triodes
6BK6, 12BK6 and 26BK6, with 100 mu triodes
6BU6 and 12BU6 with 16 mu triodes

Actually, the 12BU6 was ahead of the 12BF6.

The 26BK6 had a 26.5 volt, 70 mA heater, and I imagine was intended primarily for “12-cell storage battery applications”. At least that is how RCA described the 26A6, 26C6 and 26D6, all of which had 26.5 V, 70 mA heaters and which were otherwise similar to the 6BA6, 6AT6 and 6BE6 respectively.

And then there were quite a few TV (and FM) double diode-triodes, with various triode types and which usually had three separate cathodes (designations ending in “8”), although sometimes the diodes shared a cathode (designations ending in “7”).

Cheers,

[Synchrodyne]

Originally Posted by Synchrodyne

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.

That question is, I think, answered by the attached GE advertisement from “Tele-Tech” for 1956 June. The key wording was:

“Series-string tubes for television were pioneered by General Electric, so that designers could match cost-saving with reliable TV performance. Over 50 G-E 600-ma series-string types with uniform warm-up time already are available for use in larger models of television receivers.

“Now General Electric targets the needs of the fast-growing market for small second sets and portables with new 300-ma and 450-ma series-string tubes. Power requirements are lower, and less heat is generated. As a result, cabinets can be smaller and lighter than ever before.”

It would appear that existing valves with 6.3 V, 300 mA and 450 mA heaters were reissued with controlled warm-up heaters and typically given “A” suffixes, unless that suffix was already used to denote other changes.

A typical 300 mA case was:

6AU6 6.3V ( 300 mA) original
12AU6 150 mA series-string version, also later included in the 12.6 V car radio series
3AU6 600 mA (3.15 V) series-string version with controlled heater warm-up
4AU6 450 mA (4.2 V) series-string version with controlled heater warm-up
6AU6A 300 mA (6.3 V) series-string version with controlled heater warm-up, also replacing the 6AU6

And 450 mA:

6U8 6.3 V (450 mA) original
5U8 600 mA (4.7 V) series-string version with controlled heater warm-up
6U8A 450 mA (4.2 V) series-string version with controlled heater warm-up, also replacing the 6U8
9U8A 300 mA (9.45 V) series-string version with controlled heater warm-up

I imagine that the 9U8A was so-named to differentiate it from the 9U8, extant in Europe, and also known as the PCF82.

Cheers,

[Synchrodyne]

Originally Posted by Synchrodyne

What is interesting is that the initial release included a full set of AM radio AC/DC valves inclusive of audio output and half-wave rectifier types, whereas the corresponding AC series lacked the latter two types, although it did include an FM limiter that the AC/DC series lacked. The 12AW6 and 12AL5 were added to complete the AC/DC set. Why the 12AW6 and not a 12AU6 is not clear, and in any event a 12AU6 was released a couple or so years later.

It turns out that the 12AW6 was effectively the 150 mA series-string version of the 6AG5. The 12AW6/6AG5 were offered as an alternative to the 12BA6/6BA6 for FM receiver RF amplifiers. The 12AW6 was essentially the same as the 6AG5 except for its heater, but differed in having a separate pinout for its suppressor grid, for which it gave up the second cathode pinout. Probably for this reason it had a separate designation, rather than say “12AG5”. Before the 12AU6 was released, the 12AW6 possibly also served as an FM limiter in AC-DC receivers.

RCA evidently preferred the separate suppressor grid pinout for RF pentodes. Of its 6CB6 of 1950, so equipped, it said that the use of a separate suppressor grid pinout allowed the use of an unbypassed cathode resistor to reduce changes of input capacitance and input conductance with changes in bias. Subsequent RCA RF pentodes followed the same pattern. Also, its 6X8 triode-pentode frequency changer of 1951 traded off separate cathode pinouts for a separate pentode suppressor pinout, whereas Tung-Sol’s 6U8 had separate cathodes and an internally connected suppressor.

From circa 1954, the US valve-makers had introduced several multipurpose TV triode-pentodes (along with other multiple types), in part to help reduce valve envelope count in colour TV receivers. These followed prior use of the 6U8, and to a lesser extent the 6X8 in other-tan-frequency changer roles, and it would appear that they incorporated diverse triodes tailored to TV applications. The pentode sections could be used in vision IF stages. Typically these triode-pentodes followed the 6U8 pattern with separate cathodes and internally connected suppressors, one assumes to obtain maximum versatility of deployment for the two sections. The RCA 6AZ8 of 1956 had separate cathodes, and its suppressor shared a pinout with one side of the heater. Its pentode was of the semi-remote cutoff type, making it suitable for use as an agc’d vision IF amplifier. So RCA was quite adamant about directly earthing the suppressor of agc’d pentodes.

Cheers,

[Synchrodyne]

Returning to the origins of the noval miniatures and their apparent low-key introduction, some additional information is available from the AWV Radiotronics magazines, available here: http://tubedata.milbert.com/other/AW...ics/index.html.

The 1947 January issue noted the arrival of the 12AU7, with the following comment: “...is a companion type to the miniature series. The valve features a small glass envelope with integral button 9-pin base,...”

So RCA had quietly slipped in what was the first noval-based valve as a companion to the miniature (B7G) series, and wasn’t actually describing it as a miniature.

More information was given in the following issue, 1947 March. It included a dimensional outline of what was described as the “Small Button Noval 9-Pin Base”, copy attached.

Apparent is that RCA was using the “noval” descriptor from very early on.

Release of the 12AX7 was recorded in AWV Radiotronics 1948 January. Of it was said: “It utilizes the small-button noval 9-pin base and a glass bulb slightly larger than used on the regular miniatures.”

Thus in RCA’s view, it was not a regular miniature, but that wording allowed the possibility that it might be an irregular miniature.

By then though, as recorded in post #380, GE had declared that its 9-pin button-based valves (which included the 12AU7) were miniatures, so if RCA’s original intention was to keep the term “miniature” strictly for the B7G type, then the battle was already lost.

Incidentally, the release advice for the 12AX7 also included the 6AV6 and 12AV6, which shared the 100 mu triode.

It looks as if RCA had capitulated on the descriptor by later in 1948. AVW Radiotronics for 1948 September announced the 6BA7 and 12BA7, which were described as “being of the miniature 9-pin type”.

So the advent of the noval valves was in fact relatively low-key. And initially anyway, the noval base was used only where more than 7 pins were required. American applications where 7 pins would have sufficed but the extra heat dissipation capacity of the larger noval envelope was required seem not to have been too common. One of the earliest such that I can trace is the 6S4 TV field output triode.

Although RCA was the first to release a noval valve, namely the 12AU7, GE was probably the first to reserve a designation for such a valve. It is a reasonable assumption that the 12AT7 reservation preceded that for the 12AU7. Possibly then the noval was a GE original idea.

Cheers,

[Synchrodyne]

Originally Posted by Synchrodyne

Interesting is that the 6BE6 miniature pentagrid was also said to have been suitable for FM applications, but the inference is that the 6SB7 was more suitable. As mentioned upthread, I suspect that the FM suitability for the 6BE6 may have been predicated on the older 42 to 50 MHz FM band.

Well, my suspicions were wrong. The 6BE6 was designed with 100 MHz applications in mind. That comes out in RCA Application Note AN-121, which was reprinted in AWV Radiotronics for 1948 August, and was entitled “Use of the 6BA6 and 6BE6 Miniature Tubes in FM Receivers”. The opening paragraph of AN-121 read:

“The expanding requirements of modern receiver design have resulted in the development of several new tubes which may be used to advantage in both FM and standard broadcast bands. Two of these, the 6BA6 and 6BE6, are new miniature tubes particularly suitable for the rf and converter positions of dual-purpose FM/AM receivers.”

The subsequent worked examples referred only to the 88 to 108 MHz FM band.

So the 6BE6 (and 6BA6) were designed to provide adequate performance at 100 MHz. Nevertheless, very soon after the 6BA6 was released, RCA also offered the 6SB7 as a heptode frequency changer with better performance at 100 MHz, specifically higher conversion conductance. And a couple or so years later, the 6SB7 was followed by its noval variant, the 6BA7.

An example of the progression in front-end valve choices is given by some of the Zenith chassis of the late 1940s and early 1950s.

The Zenith 7E01 AM/FM chassis of circa 1948 valve line-up was: 12BA6, 12BE6, 2 x 12BA6, 12AU6, 19T8, 35B5. The 12AU6 was the FM limiter; the other signal valves were used on both AM and FM.

The 7F03 FM-only chassis of about the same time was: 12BA6, 12AT7, 2 x 12BA6, 12AU6, 19T8, 35B5. On the face of it, the fact that this chassis was FM-only has allowed the use of the 12AT7 double triode oscillator-mixer as presumably a better choice at 100 MHz.

But then the 7K01 AM/FM chassis of the early 1950s was: 6BJ6, 12AT7, 2 x 12BA6, 12AU6, 19T8, 35C5. The 12AT7 was used as frequency changer on AM as well as FM. So here Zenith evidently allowed the FM performance requirements to take precedence when it came to the choice of frequency changer valve. Possibly the double-triode AM frequency changer – which one would assume had higher backwards oscillator leakage than a 12BE6 – could be accepted because it was preceded by an RF stage, which would help block the oscillator signal from getting to the aerial input. The change from 12BA6 to 6BJ6 as RF amplifier might have been to obtain some improvement at 100 MHz.

Zenith had opted to jump direct from the 6BE6 to the 12AT7 for better FM performance, apparently skipping the 6BA7.

As a sidebar observation, the AM sections of the 7E01 and 7K01 chassis, intended for use in compact and relatively simple receivers, had not only an RF stage, but also two IF stages, so they were quite well equipped as domestic receivers went.

But the use of a double-triode for both AM and FM was not new. It was found for example in the Hallicrafters SX-42 of 1946. This was intended to be both a broadcast and a communications receiver with coverage from 540 kHz to 110 MHz, thus inclusive of both the “old” and “new” US FM bands. On all bands it had a two-stage RF amplifier, each stage using a 6AG5, followed by a 7F8 double-triode oscillator-mixer. The 7F8 was a Loctal valve designed for VHF applications. Unlike say the 6J6, it had separate cathode pinouts. (The SX-42 predated the arrival of the 12AT7, which otherwise might have been a contender.)

Cheers,

[jacob roschy]

Originally Posted by Synchrodyne

Something that I have never come across is a paper or article that addresses in detail the introduction of noval-based miniature valves, and the reasons for the specific choice of pinout number and dimensions. As compared with the 7-pin miniatures and the Rimlocks, both of which had learned subject articles published, the novals seemed to have had a rather quiet introduction.

The first noval, the 12AU7, was mentioned almost in passing in the RCA article “Miniature Tubes in War and Peace”. (See RCA Review 1947 June, p.331ff, recently available at: http://www.americanradiohistory.com/..._Issue_Key.htm.) This essentially covered the inception and early development of the 7-pin miniatures. Therein the noval form was presented simply as a logical development with a greater pinout count and larger envelope size. Which of course it was, but no elaboration was provided. In hindsight one might say that the parameters were well-chosen.

It could have been that introduction of the noval form was done fairly quickly, and that it was not part of the original post-WWI American plan for receiving valves. That seems to have been based upon the general use of 7-pin miniatures for small-signal valves, the use of octals for “power” valves that required larger envelopes than the 7-pin miniature format allowed, and also the use of octals for small-signal valves that required more than seven pinouts.
.........
The 6T8 (and 19T8) triple diode-triode was also an early noval release. It was not quite the same as the 6S8GT, in that it had a 70 mu rather than a 100 mu triode, but it was generally regarded as the miniature successor to the 6S8GT.

So one may see that novalization of small-signal receiving valves (those that wouldn’t fit on a B7G base) proceeded quite quickly, notwithstanding the comparable octal releases of but two or three years previously.

The 2C51 appears to be one of the earliest Noval (B9A) tubes ever, since it was registered in November 25, 1946, in the Electron Tube Registration List, allegedly developed by the Bell Labs; Murray Hill, NJ. USA

[turretslug]

That's an interesting case- it's almost like a 6J6 but with a bit more design option flexibility added, courtesy of the extra pins. The twin triode must have been appealing to designers with its compactness and likelihood of characteristic similarities but the common cathode of the 6J6 would have been a low-inductance boon in some applications and a hindrance in others

[bikerhifinut]

Incidentally, the release advice for the 12AX7 also included the 6AV6 and 12AV6, which shared the 100 mu triode.

Hope this isn't off topic.

Does that Mean a 6AV6 could be used in an application where one half of an ECC83/12AX7 is needed but not the second triode?
I was thinking about the first Valve in some valve Amplifiers where one would be building a pair of monoblocs (or a single mono). I guess the side anodes are strapped to the cathodes if used as a pure Audio valve?

Mods please feel free to move or delete if I have veered off thread.

Andy.

[Synchrodyne]

Originally Posted by bikerhifinut

Does that Mean a 6AV6 could be used in an application where one half of an ECC83/12AX7 is needed but not the second triode?

I was thinking about the first Valve in some valve Amplifiers where one would be building a pair of monoblocs (or a single mono). I guess the side anodes are strapped to the cathodes if used as a pure Audio valve?

Basically, yes, and I think it has been done. I can’t think of any examples right now, but I’ll look to see if I can find one. Certainly one can find quite a few cases where a 6AT6 or similar has been used this way, with the diode anodes strapped to the cathode, where just a single triode was required. Although on the face of it wasteful, it ensured that standard, abundant and probably relatively cheap receiving valves were used, and overall was probably more economic than would have been the production of a 100 mu single triode for what was a relatively small market.

Cheers,