Valves in commercial applications for which they were not specifically designed.

[Synchrodyne]

Regarding the ECC81, ECC82, ECC83 double triodes and their originally intended uses, I think that one needs to start with their American prototypes, respectively the 12AT7, 12AU7 and 12AX7.

The 12AT7 was a GE development. It was described as being designed for use as a grounded-grid RF amplifier and a frequency converter at frequencies up to 300 MHz. GE advertised it as being primarily for use as a converter in FM and TV receivers. RCA described it similarly, but added that it was also suitable for AF applications.

The 12AU7 was an RCA development. It was described as being useful in a diversified range of applications including multivibrators, synchronising amplifiers, oscillators, mixers and numerous industrial control devices.

The 12AX7 was also an RCA development, and was also described as being useful in diversified applications including phase inverters, multivibrators and industrial control devices where high voltage gain was a requirement. Strangely, the AF application was not emphasized in the early data, even though it used the same 100 mu triode as the 6AV6, which had been released about six months earlier.

Thus all three were presented as multipurpose valves from the start, probably leaving not too many applications that were clearly out-of-scope. The previously mentioned use of the 12AT7 as an MF mixer-oscillator was probably not envisaged by GE, but may be seen as an extrapolation from its primary role and not a step aside.

Being multipurpose, how they were represented was often contextual, as the attached Mullard advertisements from 1953 indicate.


Cheers,

[Synchrodyne]

The 7360 sheet beam tube was a late 1959 RCA release, described thus:

“The 7360 is a new beam-deflection valve of the 9-pin miniature type designed specifically for use in single- and double-sideband, suppressed-carrier communications equipment. It is capable of operating at frequencies up to 100 Mc. The 7360 makes possible the design of simplified, low-cost circuits such as product detectors, balanced modulators capable of providing a carrier suppression in the order of 60 db, and balanced mixers providing oscillator-signal suppression of at least 40 db. This valve is also well suited for use in low-distortion audio-fader circuits, remote switching of studio and high-fidelity equipment, and in other applications in which isolation of the control voltage from the signal is an important design requirement.”

This gives the impression that RCA might have seen its [the 7360] primary task as SSB generation and demodulation, and perhaps use as the final signal mixer in those professional receivers that converted from a low IF to a very low final IF before final demodulation (for example as was done in the various SSB and ISB adaptors for the Racal RA17). Whether its use in receiver front ends as a first signal mixer was envisaged is unknown, but the mention of its 100 MHz capability does at least include that possibility.

The 7360 does look very much as if it were an improved, commercial/industrial version of the GE 6AR8 sheet beam tube of 1954, which was developed primarily for colour television chroma synchronous demodulation, with claimed advantages over heptodes and dual-control pentodes in this role. In the data sheet GE did mention that the 6AR8 was suitable use in the burst gate of colour TV receivers and for a variety of other switching and gating applications.


Cheers,

[Synchrodyne]

Returning to the ECC189 and its use in HF receiver RF amplifiers, I suspect that as turretslug has suggested, its variable transconductance was a factor in its specific selection. With the further extended grid base that series connection provided, the ECC189 was evidently operable with the same range of RF agc voltage that was used for variable transconductance pentode RF amplifiers such as the 6BA6. The first generation of cascode valves, such as the 6BQ7 and ECC84, were essentially of the sharp cutoff type, with series operation providing sufficient grid base extension for TV applications where applied agc voltages (as distinct from generated agc voltages) did not usually go larger in magnitude than around -15 V. I haven’t seen the in situ curves for the ECC189 in a series configuration, but those for the ECC84 look to be noticeably smoother than those for the smoothest-curve variable transconductance pentodes, such as the 6BA6 and EF89. So one might expect the ECC189, with some of its grid base extension due to the series connection spreading, to have a somewhat smoother curve than say the 6BA6. If so, that might have been the basis for Eddystone’s claim that its ECC189-based RF amplifier was better than a pentode in respect of blocking and cross-modulation, as well as having lower noise at upper HF.

Whilst one might ascribe Eddystone’s’ use of the ECC189 RF amplifier to “bragging rights” in the case of receivers sold into individual user markets, such as the 940 and EA12, that seems less likely in the case of the 880 series, sold mostly to government end-users. So in the context of expectations of the time (early 1960s) it is reasonable to infer that Eddystone saw its cascode RF amplifier as offering worthwhile improvements, even if when viewed from a post-RA1772 perspective it is seen as a misguided step.

More-or-less concurrent with the ECC189 was the sharp cutoff ECC88, normally operated with fixed bias on the upper unit grid to preclude grid base extension, and intended mostly for negative modulation TV receivers where the use of sharp cutoff valves was often preferred. Conceivably that could also have been used as an HF receiver RF amplifier, in the same manner that a minority of makers used sharp cutoff pentodes. Maybe series connection would have been used to obtain some spreading of the grid base; I think with sharp cutoff pentodes something similar was done through the screen bias arrangements. But Eddystone was not in the sharp cutoff 1st RF valve camp, which in respect of general-purpose receivers, was a minority anyway.

The ECC88 was described as a “high-slope, low-noise double triode for use as a cascode amplifier in television tuners”. But it has been used as a cascode AF amplifier, which I doubt was an originally intended use. The ECC83 has also been so-used, and I doubt that in this case the cascode connection had been envisaged when it was first designed as the 12AX7, nor when Philips/Mullard took it up as the ECC83.


Cheers,

[turretslug]

The Eddystone 880 was supposedly intended as a stealthy monitoring receiver with minimal detectable spurious radiation (at least, its cost, size and weight would have meant a fairly small club of privileged users....)- could it have been that a cascode 1st RF stage would ensure a little extra isolation between front end and aerial input? Having made a genuinely useful accomodation of the cascode ECC189 in one set, Eddystone might have felt a degree of marketing pressure to incorporate it in more modest mass-market sets- there might even have been an element of "the same technology as the spooks' top-notch £xxxx 880" type blurb.

I gather that cascode ECC88 audio amplifier circuits have become popular because of the circuit's capability of high amplitude output swing with good linearity, this being useful because of the urge to employ output valves from an ever-expanding field of original other uses, some of which have low inherent gain- such as the screen-grid driven PL509-type architecture and the low-mu 6080 family of series regulators. In fact, under the original thread heading, one could include "just about anything with more than a handful of watts anode dissipation as AF PA"!

[FERNSEH]

In 1957 Pye employed the ECC88 in the tuner of the model PTV. The PTV was a high performance 14" CRT portable TV receiver. In the same year Pye introduced another version of the CW17F fringe area TV receiver. The tuner was modified to accept the PCC88. The 1958 model CTL58F also employed PCC88. In 1959 Pye gave up using the PCC88 and for a year or so returned to using the PCC84. From the beginning of 1961 all Pye TV tuners were equipped with the PCC89. For a quick fix increase of tuner gain many TV manufactures opted for the Mazda 30L15. This valve became a favorite with TV repairmen as a means of squeezing a little more gain from tuners which were equipped with the PCC84.
http://www.r-type.org/exhib/aaa0244.htm

DFWB.

[G8HQP Dave]

An ideal variable-mu valve has an exponential grid characteristic. This means that gain is proportional to anode current. It also means that signal handling ability remains the same at any gain setting; a straight valve gets worse at low gain. No valve is ideal, so the signal handling varies somewhat at different gain settings - often being slightly better at currents just below the normal quiescent current. However, it is unlikely that a valve could have markedly better characteristics at any gain setting other than at normal quiescent current. So you can still estimate the big signal handling by dividing transconductance into current. On this measure ECC189 is poor for HF (as is the US favourite 6BZ6) because of high gain but not unusually high current. Smooth AGC response merely preserves the basic performance, it cannot significantly enhance it. I therefore find it hard to believe that the ECC189 is better than a pentode. Better than some pentodes maybe: the 6BA6 has quite a lumpy AGC performance as it appears to have just two grid pitches. The EF89 is much smoother, but seems to have rarely been used as an RF amp.

So I suspect that the ECC189 may be slightly better for noise at the upper end of the HF band (a VHF valve will have been designed with low grid noise - which matters up there) and not too much worse for signal handling at lower frequencies - which in any case was dominated by the mixer (typically ECH81 for Eddystone).

[Synchrodyne]

The use of cascode RF amplifiers was mentioned by Pat Hawker in “Technical Topics” (TT) for 1959 September. Inter alia he said: “...very few h.f. communications receiver designs, apart from a couple in The Radio Handbook, have featured this circuit. For the home constructor, however, some very useful points are that the cascode h.f. stage has such a reserve of sensitivity that, unlike in v.h.f., neutralizing of the first stage is unnecessary; that cross modulation effects are very low; and that even a lash-up amplifier will give a performance that would need careful layout and construction with a conventional r.f. pentode.” Whether that was basis his own research or based upon the work of others is not clear, although as he usually attributed third party sources, I’d say that it reflected his own thinking.

One may wonder if this was wherefrom Eddystone got the idea to use a cascode 1st RF amplifier in the second iteration of its 880 receiver. As far as can work out, Eddystone used the ECC189 in an unneutralized series-cascode circuit. But it does look to be another case where the idea for an unusual valve application came from the amateur domain, and then moved into the commercial field. As far as I can work out, Eddystone used the ECC189 in an unneutralized series-cascode circuit.

That same issue of TT also included an SSB balanced product demodulator based upon another TV valve, the 6BU8, intended for use as a combined noise-gated sync separator and noise-gated AGC rectifier. But that was an amateur design, and unless it moved into the commercial domain, would not qualify for inclusion here.

TT for 1963 December included a cascode mixer circuit used as part of a commercial SSB transmitter, namely the B&W 6100. This used a 12AU7 in a series cascode circuit, with oscillator input to the lower stage grid, and signal input to the lower stage cathode. The mixer function itself was certainly within the design ambit of the 12AU7, but I’d judge that the series cascode connection was without, and so qualifies for inclusion here. The valve makers, having developed special valves for cascode applications, were anyway probably reluctant to suggest the use of general-purpose double triodes. Also, I’d say that valve cascode mixers were probably very scarce overall.


Cheers,

[Radio Wrangler]

The input impedance of the upper valve in a cascode, as it is working in grounded-grid mode is relatively low compared to other nodes in the circuit, but it is not low enough for voltage swing to be trivial in operation. Consequently, there will be an amount of signal voltage on the upper valve cathode, and so a device for cascode operation either needs to have low heater-cathode capacitance to avoid spoiling cascode operation, or else it needs to have a fully separate heater which can be RF-floated using a common-mode choke in the heater supply.

Maybe one or other of these features is the main difference between valves intended for cascode service and the more mundane ones?

David

[G8HQP Dave]

Pat Hawker sometimes summarised what he had gleaned from his reading without necessarily giving full references. Given the wide range of things he read he was not always in a position to accurately judge the truth of it all. Sometimes he had hunches, which often proved to be correct such as his doubts about some compact antenna claims. I suspect that his glowing report on HF cascodes was based on what others had told him, not his own work.

The main difference between triodes intended for cascode use and other triodes is that cascode requires good performance at low anode voltage. Another difference may be that smooth (or linear) transconductance is more useful than constant mu, as mu plays little role in setting the gain of a cascode. HF cascodes can be simple because parasitic capacitances are small. VHF cascodes may need neutralizing (mainly to reduce noise) and an inductor between the sections to deal with input and output capacitances.

[G6Tanuki]

One other 'misuse' of valves I recall was in a large industrial signal-generator (by which I mean a device that generated quite a few tens of Watts at frequencies from about 0.2Hz to 50KHz for use in vibration testing).

This used a quartet of EL34 in the voltage-regulated HT supply. The cunning bit being that the EL34s were used (in pairs) as grid-controlled-rectifiers in place of the more-usual couple-of-GZ34 you'd expect to find as HT rectifiers in this kind of equipment. In a way this was quite a nice trick - you needed to rectify the HT and also regulate it so why not combine both activities in the same thermionic devices?

[There were another *eight* EL34s used in pairs in a classic 'bridge' DC-coupled output stage to drive the vibration transducer (which could be either a speaker-voice-coil style thing or a piezoelectric/ magnetostrictive 'slab' you bolted to the thing-to-be-tested). ]

[Synchrodyne]

The use of the EL34 as a grid-controlled rectifier reminds me of the classic single-valve AGC rectifier/amplifier circuit used in negative modulation TV receivers. This could be seen as a grid-controlled shunt rectifier whose “AC” input for rectification was in the form of line flyback pulses, and whose grid control voltage was the video signal, the gating action of the line flyback pulses sampling the video signal only at peak sync or back porch level. I’d guess that in its earliest implementations, it used available pentodes or triodes that were not specified this role, following which the valve makers specifically addressed the application, often in their multipurpose TV triode-pentodes. Early examples included the 6AN8, launched as a multipurpose valve, and the PCF80, launched as a VHF TV frequency changer valve and then recast as a multipurpose valve with many additional applications.

Re the cascode and in particular cascode valves, I re-read the RCA article on the 6BQ7 (*), which was the first of its kind. Included was the statement: “The 6BQ7 is primarily designed to provide the characteristics needed for good performance in the various driven-grounded-grid circuits. The following detailed description of the 6BQ7 correlates its design features and electrical characteristics with the specific requirements of driven-grounded-grid circuits.”

Note that “driven-grounded-grid” was RCA’s initial description for what was later known as the series cascode circuit.

Nothing specific was said about the cathode-heater capacitance of the upper unit, but the layout including pinouts was chosen to minimize coupling between the two units. Anyway, one might say that the new circuit (a development of the Wallman cascode) and a new valve (a purpose-configured double triode) were developed in a complementary fashion.

RCA did comment upon the cross-modulation aspect, as follows:

“Figure 7a gives the plate family of characteristic curves for the 6BQ7. The tube has a sharp-cutoff characteristic which results in low input loading, although the cross-modulation is thereby increased to a degree comparable to that obtained with the pentodes. When the series-connected direct-coupled circuit is used, the overall plate characteristic curve for the two tubes is that shown in Figure 7b. The cutoff is extended by a factor of two, without adverse effect on the input loading.

“Because the curve more nearly approaches a square-law characteristic, the theoretical requisite for absence of cross-modulation, this type of interference is greatly reduced. Interference measurements indicate that cross-modulation with the direct-coupled circuit is one-eighth that with the capacitively coupled circuit, an improvement which agrees well with theoretical calculation.”

That outcome was specific to the 6BQ7 and its RCA series-cascode circuit, but it might be that it was wrongly extrapolated by others as applying generally to cascode circuits. Even Pat Hawker could have fallen into that trap, as his TT article included examples of both series- and shunt-cascode circuits, no distinction being made between the two as far as the ostensible cross-modulation benefit was concerned, even though RCA had said that there was a big difference.

(*) RCA Review 1951 March page 03ff. (http://www.americanradiohistory.com/...w-1951-Mar.pdf)


Cheers,

[Synchrodyne]

Another example that comes to mind was Drake’s use of a 6BY6 heptode as a self-oscillating 3rd mixer in its 1-A SSB receiver. The other two mixers used the 6BE6. Apparently the 6BY6 was chosen for the 3rd mixer because of its high conversion conductance.

The 6BY6 was introduced by RCA late 1953/early 1954, primarily for use as a noise-gated sync separator. It thus had sharp cutoff characteristics, and was intended to operate with low screen and anode voltages. RCA also offered it for use as a chroma synchronous demodulator in colour TV receivers, although I don’t think that this was ever mentioned on the data sheet. Presumably it was better in this role than the 6BE6 and 6BA7, both of which were at least proposed, if not actually used, and lower cost than the 6AS6, which did see some use. So its use as a radio receiver mixer was not outside of its basic functional capability, although its operation in the self-excited mode was a departure. Conventional heptode self-oscillating mixers such as the 6BE6 and 6BA7 included constructional features that were intended to aid oscillator performance and minimize coupling from the signal input. It seems unlikely that the 6BY6, with its different mission, included such features.

Both the 6BY6 and Sylvania’s similar 6SC6 of 1953 may well have been responses to Zenith’s use of the 6BE6 as a noise-gated sync separator in its then-recent TV chassis. Zenith had developed its circuit around the 6BN6 gated-beam valve, which was apparently ideal for the purpose, but then looked at lower cost options including the 6SA6 and the 6BE6, choosing the latter as providing adequate performance. The 6BN6 – really an ersatz dual-control pentode - had been developed as a better FM limiter, and in particular as an FM limiter-discriminator combination that realized, in self-limiting form, the space charge-coupled FM quadrature discriminator proposed earlier by others and using conventional heptode mixer valves. But the 6BN6 was also offered as a sync separation from the start. It would appear that the higher cost of the 6BN6 could be justified for the FM application, where it could replace two or sometimes three separate stages, but not as a sync separator where it simply replaced the customary sync separator valve.

The 6CS6 appears to have been more used as a sync separator than the 6BY6, but I have not seen any mention of its used as a chroma demodulator. It was adopted in Europe as the EH90. Philips described it simply as a dual-control heptode. But then the EH90 turned up in Mullard advertising from late 1963 where it was offered for use as a locked-oscillator quadrature FM sound demodulator in dual-standard TV receivers. (The noise-gated sync separator application was covered by the ECH84.) I am not sure that that application ever found its way into the data sheet, though, but it was validated by Mullard’s advertising. This form of FM demodulator was originally developed by Philco, who also had Sylvania develop a special valve for it, the FM1000 sharp cutoff heptode. Anyway, the FM demodulator represented a valve-maker authorized 6CS6/EH90 application in which it was self-oscillating, and where the oscillation depended upon space charge coupling from the signal input. Given the basic mission similarity between the 6SC6 and the 6BY6, one might expect that the 6BY6 could also have been used as an FM demodulator. So, that the 6BY6 might have had a strong preference for oscillating at the same frequency as the signal input made it an even stranger choice for a self-oscillating mixer where the signal and oscillator frequencies were of course not the same. Perhaps it did not work as well as Drake had envisaged, as in the 2-B receiver the 3rd mixer was a 6BE6. But that model also used the 6BE6 as a self-oscillating SSB product demodulator, a role for which the 6BY6 might have been well-suited.


Cheers,

[ukcol]

Originally Posted by Synchrodyne

Early examples included the 6AN8, launched as a multipurpose valve, and the PCF80, launched as a VHF TV frequency changer valve and then recast as a multipurpose valve with many additional applications.

There must be many examples of valves having their possible applications expanded by the manufacturer.

[G6Tanuki]

Originally Posted by Synchrodyne

The use of the EL34 as a grid-controlled rectifier reminds me of the classic single-valve AGC rectifier/amplifier circuit used in negative modulation TV receivers. This could be seen as a grid-controlled shunt rectifier whose “AC” input for rectification was in the form of line flyback pulses, and whose grid control voltage was the video signal, the gating action of the line flyback pulses sampling the video signal only at peak sync or back porch level.

I hadn't ever thought of it in those terms - I always saw it as more of a "sample-and-hold" circuit - but your interpretation is just as valid.

Pye certainly made a big thing about using this "Automatic Picture Control" black-level-management feature in their 1952 'V4' TV.

http://www.pyemuseum.org/images/cons...%28550p%29.jpg

[Leon Crampin]

I guess every valve (other than diodes) under test in an AVO valve tester is operated as a grid controlled rectifier.

Leon.

[G6Tanuki]

A good illustration of the working of the AGC-clamp logic of TVs such as the Pye V4 is shown here:

http://www.thevalvepage.com/tv/pye/v4/v4.htm

and patent details here: https://www.google.com/patents/US2875277


A couple of years later Pye released the V14 which had a slew of 'issues' despite using the Mullard PCF80 in an apparently manufacturer-approved circuit:

http://www.thevalvepage.com/tv/pye/p...cf80/pcf80.htm

[ricard]

I've seen an EABC80 and EM80 used in a tape recorder, even though these are typical radio valves. In this case the manufacturer's main product line at the time was radios, so they probably had good prices on those valves and tried to use them as much as possible. With the EABC80, the main purpose was of course the triode, and then one of the diodes was in fact use as a rectifier in the recording level meter circuit, however the others were unused. The EM80 is an unusual choice for a tape recorder as it has no overload indication capability (overlapping fields) like the EM81 or EM87. At the time of this particular machine's design (Luxor MP22) I don't even think the EM81 was available (and the EM87 even further in the future), and whereas they could have used something like the EM71 I'm thinking the EM80 was already known and good enough for the purpose.

So, perhaps a bit in the gray area of the topic as valves were actually not being used for anything other than their intended purpose, but still in a bit of an unusual context.

[kalee20]

Originally Posted by Synchrodyne

Another example was an early 1950s Philips projection TV receiver... This used a DF96 battery pentode (triode-strapped) in the picture tube timebase failure protective circuit, chosen because of the very low thermal inertia of its heater, which was fed from an additional winding on the field output transformer. Thus, termination of the heater supply would shut down the valve very quickly. When cutoff, this valve allowed a high negative bias to be placed on the tube grid, thus suppressing the beam. It was biased for cutoff, but normally held off from this condition by positive rectified line flyback pulses.

Another interesting application - and there can't be many uses of the DF96 where the filament is fed by AC!

I'm thinking, why not rectify an output from the field output transformer and just use that? Either in a metal rectifier or an indirectly heated, (longer-life)thermionic one? It would have come to the same thing, because if the timebase failed there'd be nothing to rectify. Though I suppose Philips liked to do things differently...! Doubtless if it had been Hewlett-Packard, not Philips, they'd have used a light bulb fed by the field transformer, shining on a CdS photocell!

[Giulio Maiocco]

Hello everybody

I have an early FM radio which uses a 6SN7 as a FM oscillator and mixer, up to 102MHz. You have a big trouble ahead if the 6SN7 in the set is bad, as it requires a FIVRE made one, with a brown base, otherwise this circuit simply won't work or is impossible to realign.

Cheers

Giulio Maiocco

[Synchrodyne]

Originally Posted by kalee20

Another interesting application - and there can't be many uses of the DF96 where the filament is fed by AC!

I'm thinking, why not rectify an output from the field output transformer and just use that? Either in a metal rectifier or an indirectly heated, (longer-life)thermionic one? It would have come to the same thing, because if the timebase failed there'd be nothing to rectify.

Perhaps because Philips saw the need as being for a circuit that was very quickly and independently responsive to a failure of either the line timebase or the field timebase?

Originally Posted by Synchrodyne

Philips also used the ECH81, and then ECH83 for the same purpose [noise-gated sync separation] in some of its European TV receivers, but I don’t think that it was ever mentioned in their respective data sheets.

Well, the Philips datasheet for the ECH81 had the opening statement: “Triode-heptode for various purposes in F.M., FM/AM, A.M. and television receivers”. So the various TV receiver applications were covered by this. Whether Philips had envisaged the sync separator application from the start is unknown; possibly what it had in mind in respect of TV applications was the line timebase coincidence detector for which the ECH42 had been used. For the ECH83 though the opening statement was: “Triode-heptode primarily for use as a mixer in car-radio sets.” The TV application was not included.

Originally Posted by Synchrodyne

So if not “from DC-to-daylight”, the ECC85 was used from DC to VHF, although it was typically presented as being primarily an FM front-end valve.

The Philips datasheet for the ECC85 said: “Double triode for use as R.F. amplifier and selfoscillating mixer”. Whilst that aligned with FM front-end application, it was not specifically limited to that.

The same statement was used for the UCC85, which, heater aside, was very similar to the ECC85, although I cannot find confirmation that they were exactly alike in functional terms. But the PCC85, heater aside, was the same as the UCC85. However, in the datasheet, Philips described it thus: “Double triode for use as oscillator and mixer in television receivers, up to 200 Mc/s (not for split carrier)”.

It was though used as an FM front end valve, for which it was obviously suitable in a technical sense, although that application was outside of the datasheet. The only TV front end application I have found for the PCC85 was in a Philips multistandard (Belgian) receiver, which was of the split sound type on all systems. So in that respect it was being used contrarily to datasheet advice.


Cheers,