23rd Aug 2016, 12:04 am | #461 |
Nonode
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,944
|
Re: Valve Questions
My understanding is that the 6J8 (Sylvania) and 6K8 (RCA), both released in 1938, were developed to address some of the shortcomings of the 6A8, including oscillator pulling and conversion gain fall off at higehr frequencies, say above 10 to 15 MHz.
The 6J8 triode-heptode essentially combined the functions of the 6L7 mixer heptode and a separate oscillator triode, for use where the better performance of the latter was desired but a single valve envelope was also a requirement. The 6K8 triode-hexode, as said, used a different approach that in part was intended to eliminate drawback of the 6J8 type, and that was the limited cathode area that could be devoted to the oscillator section. As a result the 6K8 was said to be usable for FM at 100 MHz. It did not have a suppressor grid, but apparently had fairly wide screen grid-to-hexode anode spacing that allowed for a low-potential virtual suppressor grid between those two electrodes. No long after the 6J8 and 6K8 came the 6SA7 (RCA, I think), which was an improved type of pentagrid that addressed the 6A8 problems in different ways. And post-WWII it was the 6SA7 type that prevailed in US practice, being used for the 6BE6 (out of pin-count necessity) and in improved form, for 100 MHz operation, as the 6SB7/6BA7. Although for 100 MHz operation it was quickly displaced by the 12AT7 and then the 6U8 and 6X8. European triode-hexodes and triode-heptodes, before the ECH81, generally had the triode grid connected internally to the hexode/heptode 3rd grid. The ECH81 deleted that internal connection, and that probably made it more versatile and increased its appeal to commercial and military users. Even with the triode unused, once it was established, it seemed to be used in place of the 6BE6/EK90 as an externally fed mixer. One may see that change in some of the Eddystone and Marconi receivers. Nevertheless, even in receivers where it used the ECH81 as signal mixers, Eddystone preferred the 6BE6 for use as a self-oscillating SSB demodulator, so that horse was apparently still best for that specific course, one might say. The Mazda 6C9, with a 450 mA heater, seemed to have adopted a “brute force” approach when it came to assigning adequate cathode area to its oscillator section; most other triode-hexodes/heptodes had 300 mA or even smaller heaters. That should have made it better at VHF. The initial TV triode-pentodes, the 6Z8 and 6U8, both also had 450 mA heaters. The previously posted Mullard advertisement for the ECH81, with industrial/military orientation, mentioned 100 MHz capability, although I don’t think that that was featured in domestic receiver advertisements and literature. Early Osram data for the X79 also mentioned 100 MHz capability. Cheers, |
23rd Aug 2016, 10:22 am | #462 |
Octode
Join Date: Nov 2005
Location: Surrey, UK.
Posts: 1,874
|
Re: Valve Questions
It's worth pointing out also that the much earlier Philips ECH21 matched the ECH81 both in terms of conversion conductance and pin-out versatility, allowing independent use of the heptode and the triode. This was of course well exploited by Philips in their 3 + R superhets.
The B8G base was a historical loser, but it sported some excellent devices in its short term of popularity. The ring seal construction would have given good VHF performance also. Leon. |
23rd Aug 2016, 5:48 pm | #463 |
Dekatron
Join Date: Nov 2011
Location: Surrey, UK.
Posts: 4,398
|
Re: Valve Questions
I'm not too sure of the motivation behind B8G superseding B9G- perhaps more than the simple slight reduction in diameter?- but the cathode spigot of later ECH21s was a distinctive consequence. There are some excellent valves on this briefly extant base encompassing low noise, high gain, frequency capability at least into low-, if not mid-VHF and appreciable power capability types. Perhaps a handicap was that its footprint was still much the same as IO and similar 'thirties types. If I may be permitted to allow art to intrude on the discussion, they're also surely one of the more appealing looking series.
The B8a types also strike me as featuring some excellent devices. The internal construction of the ECH42 is notably neat, compact and direct. Whilst the ECH81 represented refinement with the separated triode and heptode oscillator injection grids, heptode (as opposed to hexode) construction and apparent detail optimisation for FM IF amp applications, I remain to be convinced that it was a huge jump ahead of the ECH42. |
23rd Aug 2016, 8:05 pm | #464 |
Dekatron
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 14,007
|
Re: Valve Questions
Must admit, I always liked the B9G base as an improvement on the old Octal. Apart from the extra pin over the Octal, it gave significantly shorter lead-outs - important for VHF applications - and one more pin so you could reduce unwanted inductances (multiple cathode-connections as in valves like the EF54).
I particularly like the B9G screw-down retaining-ring/outer-can-earthing arrangements too: see inside a WWII "Gee" RF-unit for an example - the same was also used for the QV04/7 and TT15 VHF valves in some of the first commercial Pye two-way mobiles. Some mabufacturers rather messed-around with the outer base diameter of B9G bottles though... see the TT12/CV524 http://www.r-type.org/exhib/aaa0976.htm |
1st Nov 2016, 10:27 pm | #465 | ||
Nonode
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,944
|
Re: Valve Questions
Quote:
Another treatment of the subject was provided in RCA Engineer for 1956 December – 1957 January. (See: http://www.americanradiohistory.com/...956-12-01.pdf; “Series-String Applications of Electron Tubes (Commercial and Military), page 32ff.) From that one infers that quite a bit of research was done when the US TV industry opted for widespread use of series-string heaters. The Tele-Tech article mentions the use of 18.9 volts as a heater supply, and thus three 6.3-volt valves in series across it. A quick search found the circuit details for a couple of early 1950s UK aviation units, namely the Marconi AD108 HF receiver and the STC STR9-X VHF transceiver. The STC unit had 21 valves, mostly B7G miniatures and a minority of B9G transmitting types. All had 6.3-volt heaters, with heater currents variously of 150, 200, 300 and 600 mA. These were connected in a complex series-parallel arrangement, with three parallel groups of varying counts in series, across an 18.9-volt supply obtained from the battery (quoted in this case as 27 volts nominal) via a carbon-pile regulator. HT and grid bias were supplied by a motor-generator unit with 27-volt input and two outputs. The Marconi AD108 had nine valves, all 6.3-volt B7G miniatures, with heater currents of 200 and 300 mA. There were three series-strings of three valve heaters each, each string fed in parallel from a 19-volt controlled source. Two of the strings comprised 200 mA valves, with the third being 300 mA. Two data points only, but it looks at least possible that three 6.3-volt valves in series fed from 18.9 volts (evidently sometimes rounded to 19 volts) might have been something of a norm. That was probably as high a voltage was feasible when fed from a nominal 28-volt source via an electromechanical regulator. The next increment, to accommodate four 6.3-volt valves in series, would have taken the required heater supply voltage to 25.2 volts. Possibly this would not have allowed sufficient regulating “headroom”. Nevertheless, the RCA article referred to using four 6.3-volt valves in series across a 28 volt supply. As an aside, but coming out of the foregoing, electrical systems that were based upon 12-cell lead-acid batteries seem to have had a diversity of nominal voltage monikers, including 24-, 26-, 27- and 28-volts. To some extent these have been industry dependent, and I think that 24-volts was/is the norm for is automotive, off-highway and railroad applications, with 28-volts used in the aviation world. Returning to valves, evident is that even before series-strings became commonplace in domestic TV receiver practice, the valve makers had good reason to keep to what might be called “cardinal” heater current values for 6.3-volt valves, primarily 300 mA, but with variations to 200, 450 and 600 mA, and occasionally down to 150 mA. This facilitated their use in series-strings where required, particularly in commercial and military equipment. When the European norm for AC/DC radio receiving valves was 200 mA, some effort was made, where practicable, to have 6.3-volt, 200 mA valves that could be used in both AC and AC/DC receivers. A legacy of that was that the 200 mA number was carried over into the era when 100 mA became the AC/DC norm in Europe, as for example in valves such as the EF41, W77 and EF89. When Philips adopted 300 mA series-string valves for its standard TV receiving valves in 1949-50, there appears to have been a concerted effort to have as many as reasonably possible fit the 6.3-volt, 300 mA “mould”, the EF80 being a prime example. Two-valve series-strings were used in European 12-volt car radios, and whilst a common heater current was not required, pairwise matching was desirable, which probably explains why the ECH42 and EBC41 were both 230 mA (although it was referred to as 225 mA in a late 1940s Mullard publication). Cheers, |
||
5th Nov 2016, 1:00 am | #466 | |
Nonode
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,944
|
Re: Valve Questions
Quote:
Some valve bases seemed to have multiple envelope diameters as a matter of course, whereas for others variations were rare or even non-existent. For example the novals I think were uniform until quite late on, when the Sylvania 9-T9 series arrived in 1960. Cheers, |
|
5th Nov 2016, 12:21 pm | #467 |
Dekatron
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 14,007
|
Re: Valve Questions
The military seem to have been prepared to accept some odd setups: for example some "Larkspur" series of radios had 1.4V directly-heated-filament valves fed from 12V via individual dropper-resistors (R209 as an example) and if you want truly odd look at the 62-set: it uses a mix of directly-heated 2-volt valves (ARP12/VP23, AR8 and PEN25) and 6.3V indirectly-heated ones (ECH35, EF50, CV309). Mazda-octals, real-octals and B9G all in the same box....
It was unwise to connect these kind of radios to a "float-charged" 12V battery supply (the little 'Charging Sets 300W' could easily push a 12V battery to 16V if you weren't careful - with significant impact on valve life). |