Valve Questions

[Leon Crampin]

Although to be as fair as is possible to Philips, neither the DL92 (3S4) nor the 6V6 was designed by them. It wouldn't have hurt them very much to make their data sheets accurate though.

Many battery output valves were beam tetrodes - their increased efficiency over a pentode was of course well worth having.

Leon.

[julie_m]

Originally Posted by Leon Crampin

Philips ..... were extremely sniffy about beam tetrodes. They made beam tetrodes (even a 6V6) and always described them in the data books as pentodes.

Which they are, really, sort of. Just with a funny-shaped third grid

[neon indicator]

Not as funny a shape as the miniature 1.2V 1Jnnb series Russian "pentode" valves that appear to be entirely a bunch of rods.

[G8HQP Dave]

I think I saw it argued somewhere that there is a continuum between pure pentode and pure beam tetrode construction. It all depends on what produces the low potential between the screen grid and anode: electrodes in the electron stream connected to the cathode (pentode), or bunching of electrons due to beaming partly caused by electrodes just outside the electron stream (also connected to the cathode - beam tetrode). I don't see any reason why you can't have aligned grids (like a beam valve) in a pentode, but I don't know if this was ever done.

[kalee20]

Originally Posted by ajs_derby

Which they are, really, sort of. Just with a funny-shaped third grid

Not really - in a pentode, the suppressor grid repels secondary-emission electrons back to the anode. In a beam tetrode, there is a potential minimum caused by a particular electrode spacing and aligning the control and screen grids. The so-called beam forming plates play no part in this, and in theory should not be necessary at all. The reason that they are, is that the grid support side rods muck up the geometry, the electron distribution is wrong in these area, so these plates are put in to blank off electron flow at the troublesome areas.

As for the DL92 being a Mullard second-source, not of their design, I'm wondering why if they had the ability to make these, and pay appropriate royalties, why their DL96 last-generation battery valve with miniscule heater power, wasn't designed as a BT? It's a true pentode!

[Heatercathodeshort]

Originally Posted by kalee20

The KCF80 - is this number correct? I had a search in the National Valve museum, as it would be a 2V jobbie in B9A, but it came up blank!

Sorry! The valve should read KCF30. It is actually a Mazda valve badged Mullard. J.

[Synchrodyne]

There seems to have been some variation in beam tetrode terminology. For example, Osram sometimes referred to the KT66 as a “power tetrode”, and then went on to say that the KT66 was designed as a beam tetrode with aligned grids. Later, GEC used the term “beam pentode” for the KT series of valves. That appears in the GEC book “An Approach to Audio Frequency Amplifier Design”. The term “kinkless tetrode” which led to the “KT” designation faded from the scene quite early on.

A common feature – I think – of beam tetrodes is that the beam plates are internally connected to the cathode. On the other hand output pentodes vary as to whether or not the suppressor grid was internally connected to the cathode or had its own pinout. For example the EL84 had an internal connection whereas in the EL34 the suppressor grid had its own pinout. I am not sure whether it was an important difference at AF. For RF applications it was considered good practice to have the suppressor grid brought out to its own connection.

Cheers,

[neon indicator]

Beam plates connected to cathode is really a Pentode? "Grids" don't have to be wires, meshes etc. They are any suitable piece of metal that's between the Cathode and Anode?

It's just that most valves use sparse wire coils or open meshes as grids. Some "grids" are a single plate with a hole or a pair of rods.

[neon indicator]

Originally Posted by Synchrodyne

For RF applications it was considered good practice to have the suppressor grid brought out to its own connection.

I suspect that for common grid RF amp, or any AF amp with a lot of negative feed back via Cathode resistor NOT decoupled, you run into problems with an internally connected suppressor grid. Logically in such applications the G3 needs to be on a separate pin to allow flexibility in AC impedance and DC bias.

I found experimenting with "common grid" mode on the sub-miniature Russian Pentodes I had to use it as a triode, using G2 as anode and earthing the anode (G3 then is just extra capacitance between Cathode and Ground and Cathode and G2, now used as anode, so oscillation at 100MHz on the 60MHz rated valve is then possible with no external capacitors, just L on G2 as anode).
(since it's 15V to 48V HT experiments and 0.1mA to 1mA HT current, and the grids and anode are all similar rods it's not likely to run into dissipation problems).

Looking at recent Audio amps with so called "triode" connected Pentodes I can't buy that they are really working like Triodes. To me they just look like badly biased Pentodes to get a particular desired distortion.

[Synchrodyne]

The attached book excerpt provides some background on the preference for a separate suppressor grid pin for RF pentodes.

Virtually all of the major types of indirectly-heated miniature RF pentodes used in consumer equipment had separate suppressor grid connections, including the 6BA6/EF93, EF42, EF80, EF85, EF89, EF183 and EF184. A strange exception was the EF41, which had the suppressor grid connected internally to the cathode. The derivative EAF41 was similarly configured, but the EAF42 differed from the EAF41 in having a separate suppressor grid connection. But ostensibly this was done simply to allow the suppressor grid to be used as an AGC delay diode, given that there was only one discrete diode in the EAF41/EAF42. Still, the EBF80, which was essentially an EF41 with two diodes and therefore not needing the ersatz second diode of the EAF42, retained the separate suppressor grid connection introduced with the EAF42, suggesting that this arrangement was considered preferable per se. The oddity of the EF41 appears compounded when one considers that its contemporaries the EF42 (RF) and EF40 (AF) did have separate suppressor grid connections. In the case of the high-slope EF42, also intended for use as a TV IF amplifier, perhaps the separate suppressor grid connection was considered more necessary than with the medium-slope EF41.

The battery RF pentodes such as DF91 and DF96 had their suppressor grids connected internally to their filaments. These typically had low slopes. And internal connection to the cathode was the norm for frequency changer heptodes and for the hexode/heptode portions of triode-hexodes/heptodes. Conversion slopes for these were usually modest. So maybe slope was a factor in deciding the suppressor grid connection?

Cheers,

[neon indicator]

Yes, they say for RF amp don't connect suppressor Grid to Cathode. I suspected that.

If the cathode to ground is very good RF wise, (low inductance, high capacitance, or simply well grounded), then up to a certain RF level you can use suppressor on cathode. Hence I suppose some IF amps. With any significant RF impedance at cathode (with suppressor to cathode) you have a nice oscillator as the extract and my earlier post suggests. Higher "slope" certainly makes oscillation more likely.

Thanks Synchrodyne.
In PSP it scaled up by 200% and adjusted white/black levels to be very readable.

Which book is it, is the rest good and is it available as PDF or hardcopy?

[Synchrodyne]

The book is:

"Television Receiving Eqipment"
W.T. Cocking
Fourth Edition
Wireless World, 1957

It is a well-written book. As the fourth edition of an original first published in 1940, I'd say that it's mostly of historic interest, and probably better-suited to the interested layperson like me than to those who have a professional connection with electronics. By way of comparison, it's at about the same level as the (probably) better-known Spreadbury books. I can scan and post the contents page to give a better idea as to what it covers.

I would guess that it is available second-hand from ABE et al. I bought my copy at Hay-on-Wye in 1995.

Cheers,

[neon indicator]

Thanks

No need to scan the contents.
Amazon has various versions from $4 to $20 and the last version maybe 1985.
I'll maybe treat myself to 1947 & 1985 versions some day.

[Synchrodyne]

I have found another brief reference to the RF pentode suppressor grid connection question, attached. It is from “Television Engineering Handbook; Donald G. Fink, Editor; McGraw-Hill, 1957; LCC 55-11564. Basically it confirms the preference for external grounding of the suppressor rather than internal connection.

Originally Posted by Ed_Dinning

H Gents, I'm not sure if the later line output valves (PL519 etc) were used that much in Hi Fi amps; PA, instrument, etc possibly.

An examination of the Vg/Ia curves shows them to be pretty non-linear, exactly what is needed for valve pulse operation.

It is only latterly that the technique of applying the signal to G2 and the bias to G1 has been developed, where Vg2/ Ia is pretty linear.

Not too many curves around that show this though, and there is the drawback of a lower driving impedance and the signal needing to have a high standing DC.

Maybe the PL519 etc., were OK for Class B PA amplifiers?

Although the PL519 had 35 W anode dissipation, the magnoval envelope size doesn’t appear to be much different to that of say an EL34. So maybe for Class A high quality audio applications, around 25 W anode dissipation would have been a reasonable upper limit in the magnoval envelope. But that power class valve was already well-subscribed with the established octal types, hence there was no real need to develop new magnovals.

Originally Posted by Heatercathodeshort

I believe the only B8A valve with a top cap is the UL44. It is a line output valve used in early post war Philips television receivers including the projection models [600A etc] Looks a bit like a PL81.

Eventually it occurred to me to look through Kerkhof & Werner, and sure enough, I found the UL44. K&W has a description and circuit of a (presumed Philips, as the book was published by Philips) 625-line projection TV receiver, with no less than 31 valves, and a real mix at that: 9 x EF80, 2 x EB91, 1 x PL83, 1 x EQ80, 1 x EL42, 3 x ECL80, 1 x PL81, 1 x PY80, 1 x DF96, 2 x PY82, 3 x UY41, 1 x UBC41, 2 x UL44 and 3 x EY51. So there were “D”, “E”, “P” and “U” series, and B7G, B8A, B9A and “special” (EY51) bases. The circuit was transformerless, with both 100 mA and 300 mA heater chains. The CRT was in the 300 mA chain, as were the 200 mA valves (EQ80 and EL42), which had parallel resistors. The DF96 was part of the CRT protection circuit, used because of its low thermal inertia; it would stop very quickly upon loss of filament volts. IFs were 18 MHz (sound) and 23.5 MHz (vision); the receiver at interest would seem to have predated the arrival of the CCIR standard 33.4/38.9 MHz IFs.

Also described in K&W is a UK 5-channel direct-view receiver whose valve complement is 6 x UF42, 2 x UL46, 1 x UB41, 1 x UAF42, 1 x UL41, 2 x UCH42, 1 x UL44, 1 x UY41, 2 x EY51 and 1 x PZ30. It was also transformerless, with 3 100 mA heater chains that joined together at the CRT.

That example would be an ideal “QED” for the adoption of 300 mA heater chains (at any rate where the transformerless technique was somehow justified) and so the “P” series valves. It appears that these arrived after the B7G and Rimlock series, but just before European adoption of the B9A base. So there were one or two early octal-based examples, but the combination of 300 mA heaters and the B9A base became “mainstream” with the Mullard World Series of television valves in 1950.

So “U” (100 mA) and “P” (300 mA) series valves were essentially European? American practice for AC/DC radio receivers was to use 150 mA heater valves? I am not sure when transformerless TV receivers were adopted in the USA, but the examples shown in Fink (1957) all have power transformers and 6.3 volt heaters.

K&W also has an example of an American 12-channel turret tuner using a 6AG5 (EF96) pentode RF amplifier and a 6J6 (ECC91) frequency changer. Given that the IFs were 21.25 MHz (sound) and 25.75 MHz (vision), the triode mixer was logical. But looking up the 6AG5 to see how it differed from the 6AK5, I discovered that it is described both a beam tetrode and a pentode (see: http://www.r-type.org/exhib/aaa0409.htm), which takes us back to the very recent subject discussion. K&W seem to have been writing just before the advent of the cascode amplifier, which is not mentioned despite there being a detailed treatment of triode VHF amplifiers. Similarly they may have been just ahead of the adoption of the standard 41.25/45.75 MHz IFs in the USA. As an interesting sidebar, whilst the July, 1950 adoption of the 625-line standard is mentioned, nothing is said about the prior Russian 625-line standard – “politics” of the era, perhaps.

Cheers,

[neon indicator]

Thanks, Synchrodyne, great stuff.

I guess cable Broadband Modems in Europe are using 44MHz as carry over from availability of 44MHz SAW for US TV IFs.

[Synchrodyne]

Re the “P series 300 mA series heater chain valves, the timing of the larger scale advent of these actually helps makes more sense of the Mullard activities in terms of its move away from Rimlock valves, and to the noval types.

In general the “P” series involved many new valves and not just different heater versions of established valves. Not only that, but several of these required noval bases in order to have sufficient pinouts. So it would have made sense to standardize on the noval base for the TV World Series, apart from one or two special cases, hence the more-or-less “clean sweep” change in 1950.

Where appropriate “E” series valves happened to have 300mA heaters, they were adopted into the TV World Series as such. “E” seemed to take precedence over “P” for designation purposes. It is reasonable to assume that “E” series valves with 300 mA heaters introduced from circa 1950, or even a little before, were designed with heater-cathode insulation appropriate for series chain use, but I don’t know if that had always been the case.

On the other hand the timeline for changes to the radio valve range was more relaxed.

With radio, it was the advent of FM that was the evident spur to move from Rimlock to noval bases, starting in 1953. The “first wave” of noval-based radio valves were those aimed at FM-AM receivers, including the ECC85 (new type), ECH81 (improved replacement for the ECH42), EABC80 (new type), EL84 (higher output replacement for the EL41) and EZ80 (direct replacement for the EZ40). The EF85 was also adopted from the TV series, and the EBF80, previously an honorary member of the Rimlock series, was by virtue of its noval base also in the first radio group, although not featured in the Mullard literature. Similarly the EF80 from the TV series was often used in FM-AM radio receivers, but not featured in the radio series.

On the audio side, the initial World Series arrived about the same time as the noval-based radio valves, perhaps because the EL84 was shared. It included the EF86 (direct replacement for the EF40), three double-triodes, ECC81, ECC82 and ECC83, none of which were quite the same as the ECC40, the EL84, and the EZ80. Of the noval double-triodes, the ECC81 at least, and probably the ECC82 were drawn from the TV series. The ECC83 was the primary member and “new”, and one assumes better-suited to anticipated future requirements than would have been a simple noval rework of the ECC40. But Brimar had been offering the essentially similar 12AT7, 12AU7 and 12AX7 double triodes in the UK since 1950, so the Mullard valves were not really new.

The “second wave” of a couple of years or thereabouts later included the EF89 (improved replacement for the EF41), EBF89, EBC81 (direct replacement for the EBC41) and EL85 (direct replacement for the EL42) on the radio side. The EZ41 does not seem to have had an EZ8x replacement, but on the other hand, with the higher capacity EZ81 became available in this time frame. On the audio side the EL34 (octal) was added and the GZ34 (octal) might also have been from this time period.

Whilst the EBF89 bore the same relationship to the EBF80 as did the EF89 to the EF41, it became an alternative, not a replacement, and EBF80 was used through to the end of the valve era. Evidently Mullard saw its EBF80 and EBF89 as adequately covering the applications previously addressed by the EAF42. Although, as recorded previously in this thread, the EAF81 existed, it wasn’t part of the advertised Mullard radio range. (Neither were either the EF81 or ECH80, which were simply noval-based versions of the EF41 and ECH42 respectively.)

So it is possible to impute some greater logic to Mullard’s activities in the Rimlock-to-noval transition era.

Cheers,

[Synchrodyne]

The subject of tuning indicators (magic eyes) has come up in a couple of recent threads (Leak Troughline I Question at: https://www.vintage-radio.net/forum/...ad.php?t=70457
and Were there any all valve FM stereo radios? at: https://www.vintage-radio.net/forum/...ad.php?t=69972.) and has prompted a few thoughts and questions.

The Rimlock valve series did not include any tuning indicators. At the time Philips/Mullard said that owing to the small diameter, deflection would be difficult to gauge and so almost ineffective. Thus the EM34 (octal) was recommended for use with the Rimlock radio valves.

Clearly there was a rethink about tuning indicator size limitations, resulting firstly in the subminiature DM70 in 1953, basically a battery valve but also used in some mains-powered receivers. Then later in 1954 the noval-based EM80 arrived, followed later by the EM81 and EM84. The EM80 was in the “second wave” of Mullard’s noval radio valve introductions.

All well and good, but Mullard did not really address the requirements of FM receivers, where the requirement was for an accurate centre-channel indication for minimum demodulator distortion. In respect of the EM80 it said: “The EM80 is suitable for use in a.c. mains-operated receivers. It will also give a reliable indication of the correct tuning point in FM receivers where the bandpass response is slightly peaked.” But I don’t think that all FM receiver builders necessarily wanted to adjust their IF bandpass responses just for this purpose.

In the USA, an FM-oriented tuning indicator had been introduced in the late 1940s, namely the two-bar 6AL7GT (octal base). This could provide both centre-channel and signal strength indications, using discriminator DC for former and limiter grid DC for the latter. So the precedent was there. In Europe the two-bar EM83, introduced in 1956 as best I can determine, did more-or-less the same job, and was also used as a level indicator for stereo tape recorders. See: http://www.radiomuseum.org/tubes/tube_em83.html.

Somehow Mullard must have missed this need/opportunity and the FM tuner and receiver makers were left to fend for themselves. Thus Leak used 50 Hz injection with a conventional magic eye to provide to provide accurate centre-channel as well as signal strength indication. I think that Lowther did the same with some of its FM tuners. Quad used a twin neon indicator driven by a double-triode phase inverter, but this provided centre-channel indication only, not signal strength. Jason also used this on its evidently short-lived AM/FM.2 tuner, which also had a conventional magic eye. GEC used an EM80 driven by a B309 double-triode. Another option, although I don’t know if it was used by any other UK maker, was to use diodes to obtain the negative of the modulus of the discriminator DC output and add this to the limiter grid DC for feeding to a conventional magic eye.

But all of these methods seem more complicated than would the use of a two-bar magic eye. That said, I have not seen numbers for two-bar magic eye precision, and perhaps it didn’t match say the twin neon (and later twin-lamp) indicators. Although for its FM4, Quad switched from the twin lamp system used on the FM3 to what was in effect a stepped led implementation of the two-bar magic eye.

Brimar is reported (Wireless World October, 1961) to have introduced the two-bar EMM802 as a tape recorder level indicator. Presumably this could also have been used as an FM tuning indicator, but by then new valved FM tuner and receiver designs would have been scarce.

None of the UK valve makers seems to have offered the EMM803 or something similar that simultaneously was a signal strength indicator and stereo beacon. Perhaps that was because valved stereo decoders were not used by UK makers. But even so, an EMM803 would not have been out of place in a hybrid receiver. Of course, the EMM803 was not a two-bar FM tuning indicator. So perhaps the ideal would have been an EMMM803?

Cheers,

[Synchrodyne]

On a different tack, looking at the schematics for two American all-valve FM stereo radio receivers (which activity came about because of the thread: Were there any all valve FM stereo radios? at: https://www.vintage-radio.net/forum/...ad.php?t=69972) shows up some differences in 1960s American practice of the period, including the use of some valve types that I think would be rare in British practice.

The GE T1000 has a two-valve FM front end, consisting of a 6EW6 RF amplifier and a 6EZ8 mixer/oscillator/AFC stage. The 6EW6 is a sharp cutoff high-transconductance TV IF pentode. I suspect it might be of the frame grid type, but cannot confirm this. Broadly speaking, its use as an FM RF amplifier would be akin to a UK manufacturer using an EF184 (instead of an EF80) in the same role, which I doubt rarely happened, if at all.

The 6EZ8 is a triple-triode. It fits on a B9A base because two of the cathodes share a pin with one side of the heater. In the GE application, the fully independent triode is the reactance valve. It might be, more-or-less, a triple-diode version of the ECC81, and if so, purpose-designed for the application in which GE used it, bearing in mind that the ECC81 was used as an FM mixer/oscillator combination.

No other unusual valves are found in the GE circuit, but the AM side is obviously very much an economy design, in contrast to the FM side, which as well as a two-valve front end, has a three-valve IF strip, 6AU6A, 6BA6 and 6AU6A, the last-mentioned being the limiter. The 6AU6A FM 1st IF stage also serves as the AM self-oscillating mixer, there being no RF stage. I suppose the use of a pentode AM mixer seems a bit strange until one remembers that some of the more elaborate HF communications receivers of the period did the same, albeit not self-oscillating. The 6BA6 is a common FM and AM IF stage, but the AM is also fed to the 6AU6A limiter stage, which is then acting as a grid leak detector. Thus the basic FM valve line-up is neither added to nor modified to cater for AM.

The Zenith MJ1035W used a single-valve FM front-end, using the 6JK8. This is described as being purpose-designed for FM front ends, with dissimilar triodes, and with a strap frame grid RF amplifier side. I am not sure how or even if “strap frame grid” differs from “frame grid”. But in a general sense, the 6JK8 is the frame grid era development of the ECC85 concept. I don’t think that there was a comparable UK/European development, the ECC85 serving in the single-valve FM function throughout the valve era.

Zenith also used a three-valve FM IF Strip, 2 x 6BA6 and a 6AU6A limiter. Both 6BA6s were used as AM IF amplifiers, and there was a separate AM front end with 6BA6 RF amplifier (good) and 6BE6 self-oscillating mixer (perhaps not so good). Then a 6BC7 triple-diode was used as balanced FM ratio demodulator and AM demodulator (no AM AGC delay as far as I can see, not so good). The 6BC7 seems to date from 1950, so was probably developed primarily for use in FM-AM receivers. Another and later American triple-diode was the 6BJ7, said to be a triple version of the 6AL5 (EB91) and developed primarily for DC restoration purposes in colour TV receivers.

The Zenith decoder includes a 6BN8 double-diode-triode, the triode serving as matrix amplifier and the diode pair as multiplex demodulator. The 6BN8 has separate cathodes each for the two diodes and the triode. The triode has a mu of 70, so to a first approximation it looks like an EBC41/81 with separated cathodes. Such could also be used as a combination FM demodulator (including balanced ratio demodulator) and 1st AF stage, or perhaps with the triode as a cathode follower for an FM tuner, although it is described as having been developed for television applications. Another American TV double-diode-triode with separate cathodes was the 6BJ8, its triode having a mu of 20. Much earlier in this thread I suggested that had Mullard chosen introduced this kind of double-diode-triode, it might have done well to use the “EAA” designation to differentiate it from the common-cathode “EBC” type. I suspect that Zenith’s use of the 6BN8 in its stereo decoder simply reflected its availability and suitability for its purpose, and that stereo decoder applications were not originally envisaged.

Anyway, in summary this little excursion, which started out as a search for all-valve FM stereo radio receivers, has thrown up some transatlantic differences, and has one wondering why some of these American developments were (apparently) not used in the UK (and Europe).

High slope frame grid pentode as FM RF amplifier – maybe lesser need for fringe reception in the UK and anyway a preference for single-valve front ends.

Dissimilar double-triode FM front end with frame grid RF stage – again maybe lesser need for extra RF gain. (Although it should be noted that Leak used a frame grid cascode double triode ECC88 in the Troughline Stereo, and late UK TV-FM receivers would have had frame-grid front ends, so in that case FM got the improved valves by the back door, so as to speak.)

Triple-triode FM mixer/oscillator/AFC – less perceived need for AFC, at least in table radio receivers and the single-valve front end preference.

Triple-diode – this would have allowed circuits to get beyond the limitations of the EABC80, but there were other pathways. Possibly it was colour TV that made the tripe diode relatively commonplace in the US and so acceptable for other applications. Still, a triple diode would have been beneficial for high-quality AM receivers. One diode for demodulation, one for AGC, and one for AGC delay. Separating the last two functions allows the avoidance of differential distortion. On the other hand the same benefit can be obtained by using an AGC IF side chain with an EBF80 (or EBF89) as was done by Chapman, which has the added possibility of reducing modulation rise without trading off AGC performance.

Double-diode-triode with separate cathodes – this would have been more useful than the EABC80 for FM-only receivers and also for FM-AM receivers, in the latter case using say an EBF89 in the IF chain to cover AM demodulation and AFC.

Was it that the setmakers were not interested in these possibilities or that the valve makers didn’t want to offer them in order not have too many types in their ranges? Philips/Mullard dominance in the valve business probably came into play, but did Brimar quietly offer any of the above-mentioned American valve types?

Cheers,

[Synchrodyne]

Quite early in this thread, and again in the most recent posting, the subject of double diode triodes (DDTs) was reviewed, and it was commented noted that a variant with separate cathodes for each unit, i.e. three cathodes in all, might have had greater utility for FM and FM-AM receivers than the standard European/British layout with a common cathode for all three sections, for example as was used in the EBC41 and EBC81. It was noted that such separate cathode DDTs did exist in the American series, as evidenced by the 6BJ8 and 6BN8, both apparently intended for TV line timebase applications.

A further look shows that as well as these, there was a 6FM8, designed expressly for use as an FM demodulator and first audio stage. The triode was 70 µ, so I should think comparable to the triode in the EBC41/81/90. One could visualize that an FM-AM receiver valve line-up could have been say ECC85, ECH81, EBF89 (or EBF80), 6FM8, EL84 instead of the more common ECC85, ECH81, EF89 (or EF85), EABC80, EL84. The valve count was the same, but the alternative offered improved functionality in that delayed AGC on the AM side was possible, as was a balanced FM ratio demodulator. Evidently the American designers saw it that way, but for reasons that are probably now not discernible, Philips/Mullard and others in Europe were not convinced.

Another American three-cathode DDT was the 6BV8, primarily intended for use as a TV synchronous demodulator and chrominance amplifier, but also said to be suitable for use as an FM demodulator and audio amplifier.

As well as the three-cathode DDTs, the American series also included two-cathode DDTs, in which both diodes shared a common cathode and the triode had its own cathode. Examples are the 6CN7 and 12BR7, both developed for TV applications, although the triode in the 6CN7 was said to be the same as that in the 6T8 (EABC80). In an AM receiver, this kind of DDT would have allowed triode biasing to be independent of that for the diodes (AGV delay).

Again unlike the European case, the American valve series also included diode pentodes (DPs) and double diodes pentodes (DDPs) intended primarily for TV applications, and which differed from the single-cathode types developed for radio receiver applications. I stumbled across the 6AS8 DP when looking at the schematic for the Sony CVM-2300U receiver/monitor, where it was used as the 3rd vision IF amplifier and vision demodulator. There was also the 6BY8, perhaps intended for a broader range of uses, that was effectively a combination of the ubiquitous 6AU6 (EF94) with half of a 6AL5 (EB91). Both the 6AS8 and the 6BY8 each had separate cathode sections for the diode and pentode sections.

American DDPs included the 6BT8, 6BW8 and 6ET7. These each had two cathodes, one shared by the diode pair and one for the pentode. The diode pairs were primarily for use as horizontal phase detectors; the pentodes had various uses.

Cheers,

[Leon Crampin]

A very interesting outline of the evolution of FM receiver design.

Two things which spring to mind are that, in the UK at least, by the time FM was introduced there was little demand for receivers with good AM performance and so a simple demodulator stage such as the EABC80 would suffice. Few people would notice the loss of delayed AM AGC. For any set maker who wished to incorporate this, an EBF89 in the IF stage would do the job without increasing the valve count.

Remember also that, at about the time the EABC80 was introduced in Europe, germanium diodes were a commercial reality. Although relatively few FM sets used them, an "economy" set was easily possible by using for example, an ECL82 in the output stage.

My Pye Piper P117 is such a set - FM/LW/MW using only 4 valves in total - ECC85, EBF89, ECL82 and EZ80. The FM detector uses Ge diodes. It works quite well, but must have been a very difficult set to manufacture and align as there is a lot of RF switching and reflexing. (They didn't make another set like this again).

I would guess that as Philips had the stranglehold on valve design by then (Mazda and Osram had given up) they saw the OA79 and friends as the way ahead, at least for demodulation.

Leon.