26th Apr 2010, 1:19 pm | #41 |
Hexode
Join Date: Jul 2008
Location: Toulouse, France
Posts: 406
|
Re: Valve Questions
Hi !
Bear in mind that German sets of that era where marked with a tax on the valve number. Including the rectifier. This forced manufacturer to an early adoption of solid state rectifiers and combined valves. There is a Grundig AM/FM set with only 3 valves ! (I can't get my hans on the specs right now). |
26th Apr 2010, 1:44 pm | #42 |
Rest in Peace
Join Date: Sep 2008
Location: Solihull, West Midlands, UK.
Posts: 4,872
|
Re: Valve Questions
Apart from the screen and heater, the ECC85 is very similar to the ECC81:
ECC81: gm at 10mA=5.5mA/V, mu=60, heater 300mA ECC85: gm at 10mA=5.9mA/V, mu=57, heater 435mA I wonder, is the increased heater current an attempt to prolong cathode life at higher anode currents or is there better heater-cathode insulation to reduce stray capacitance effects on the oscillator? Something I feel we miss nowadays is an actual description by someone who knows why a particular valve was designed the way it was. I imagine a direct frame-grid successor to ECC85 was not developed because if you want higher performance you wouldn't do it that way. Instead, ECC88 cascode or EC900 triode would be followed by ECF80 or ECF86 mixer. |
26th Apr 2010, 3:49 pm | #43 | |
Dekatron
Join Date: Mar 2005
Location: Gateshead, Tyne and Wear, UK.
Posts: 7,444
|
Re: Valve Questions
Quote:
The set was lively enough on AM but not so good on FM. DFWB. |
|
26th Apr 2010, 4:54 pm | #44 |
Tetrode
Join Date: Jun 2008
Location: Preston, Lancashire, UK.
Posts: 50
|
Re: Valve Questions
Hello,
can any help with following. I am interested in old valve amplifiers especially like the Williamson amplifier of 1940s.i like reading about the circuits and the valves used. I am after information oN THE RANK GAUMONT DUO-SONIC KALEE valve amplifiers of the 1950s , such as the valve line up -----wattage out put from these kalee amplifiers, also did the rank gaumont make there own amplifiers kind regards. cinema1. |
27th Apr 2010, 2:18 am | #45 | |||||
Nonode
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,944
|
Re: Valve Questions
Quote:
It seems though that the general-purpose but VHF-capable ECC81 concept was developed in two directions to meet specific receiver requirements. One was the ECC85, optimized for single-valve FM front ends, and the other was the ECC84, optimized for TV tuner RF amplifier use, where high and stable gain was needed to offset the noisiness of pentode mixers, the latter less desirable at Band III but needed for Band I once standard IFs moved up to being just under the bottom end of that band. The ECC81 (12AT7) was used as a TV mixer-oscillator in Band I only days, and with the lower IFs, for example in the Cyldon TV.5 5-channel tuner. I think it was also used as a single-valve front end in some early FM receivers, perhaps conceived before the ECC85 became available. An interesting application was in the Eddystone 770U UHF communications receiver. One ECC81 was used as a shunt cascode 50 MHz 1st IF amplifier, and another as the mixer-oscillator for the 2nd frequency conversion. Quote:
Quote:
Quote:
Quote:
Cheers, |
|||||
27th Apr 2010, 11:32 am | #46 |
Octode
Join Date: Nov 2005
Location: Surrey, UK.
Posts: 1,870
|
Re: Valve Questions
An interesting analysis. With reference to the "P" series valves, TV sets with band II were very much in vogue at one time and hence spawned the PCC85, PABC80 and even (although I've never seen one in a set) the Brimar PM84.
With reference to the last paragraph, a circuit of this set would be interesting if anyone has it. It could easily have been a 3 valve set - guess who didn't make Se rectifiers... Leon. |
27th Apr 2010, 11:47 am | #47 | |
Rest in Peace
Join Date: Sep 2008
Location: Solihull, West Midlands, UK.
Posts: 4,872
|
Re: Valve Questions
Quote:
The ECC88 probably represents not only the historical limit of 'normal' low-noise valve development but also near enough the technical limit. To get lower noise, referred to the input, you need to increase gain but this reduces signal handling. For some specialised purposes you can live with a restricted dynamic range, but not domestic broadcast reception. |
|
27th Apr 2010, 1:17 pm | #48 | |
Dekatron
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,079
|
Re: Valve Questions
Quote:
|
|
27th Apr 2010, 1:22 pm | #49 | |
Dekatron
Join Date: Feb 2007
Location: Leominster, Herefordshire, UK.
Posts: 16,535
|
Re: Valve Questions
Quote:
This sounds like a question for a new thread.... it seems to have been ignored here. I wouldn't be surprised to find someone on this forum who knew what you're after though...... Oops.... replied to a thread that had been on screen for about 20 minutes!
__________________
....__________ ....|____||__|__\_____ .=.| _---\__|__|_---_|. .........O..Chris....O Last edited by Herald1360; 27th Apr 2010 at 1:24 pm. Reason: Beaten to the draw! |
|
28th Apr 2010, 11:17 am | #50 | |
Nonode
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,944
|
Re: Valve Questions
Quote:
Cheers, |
|
3rd May 2010, 4:49 am | #51 |
Nonode
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,944
|
Re: Valve Questions
The concurrent thread "ECH81 vs 12AH8", at:
https://www.vintage-radio.net/forum/...ad.php?t=53612 tends to confirm the ubiquity and dominance of the ECH81, as discussed earlier here. Also, the receiver at interest, the KB LG40FM, seems to be an example that used an ECC81 rather than an ECC85 for its FM front end. Perhaps it was designed before the ECH81and ECC85 became available. Cheers, |
3rd May 2010, 7:46 am | #52 |
Octode
Join Date: Jan 2003
Location: Ware, Herts. UK.
Posts: 1,082
|
Re: Valve Questions
KB used the ECC81 in quite a few of their FM equipped sets. I have a FB10FM that uses an ECC81, this set was introduced in 1955 by which time the ECC85 would have been available.
I think that this use of the ECC81 may be due to KB's connection to the American ITT corporation. It's possible that that KB's FM tuners were originally American designs. John |
3rd May 2010, 8:32 am | #53 | ||
Nonode
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,944
|
Re: Valve Questions
That makes me think that the ECC85 was a European development, and that perhaps the American version, the 6AQ8, came later. Radiomuseum.org dates the 6AQ8 to August, 1955, but the ECC85 to October, 1953 (Telefunken). Maybe the Germans were more concerned with optimizing overall performance and minimizing oscillator radiation than the Americans, hence the latter found the 12AT7/ECC81 acceptable until someone else had done the development work for the improved version?
In saying this, though, it is not clear to me how valve technology IP was transferred amongst manufacturers. E.g. did the developer of a particular valve routinely cross-license it to other manufacturers? Or was there a lot of reverse-engineering taking place? Returning to KB, did it not also favour the 6BE6/EK90 pentagrid frequency changer in its AM-only designs? That would be another indication of American influence. Quote:
Still, it wasn’t until FM tuner front end solid state design settled down to the dual gate mosfet RF amplifier and dual gate mosfet mixer combination that the best valve performance (e.g. ECC88 plus ECF86) was matched in both noise and dynamic range terms. “Intermediate” designs using pnp germanium or npn silicon bipolar transistors seem to have been regarded as inferior, particularly in dynamic range terms, and many hi-fi manufacturers skipped this stage of development, retaining valve designs until mosfet-based front ends became practicable. Quote:
For Band I-only TV tuners, pentodes seemed to have been seen as adequate. For example, the Cyldon TV.5 had a pentode RF amplifier with double triode mixer-oscillator, whereas its contemporary TV.12, covering Bands I and III, had a cascode RF amplifier and triode-pentode frequency changer. Evidently Cyldon did not see that the cascode RF amplifier was justified at Band I frequencies. On the other hand, the triode noise advantage probably became evident to a small extent somewhere in the HF band, and so where very low noise RF amplification was advantageous, the cascode RF amplifier was used in HF communications receivers from the later 1950s, notably in several Eddystone modes. But an interesting earlier case was the Dynatron T139 tuner (circa 1954), which used a broad band earthed grid triode RF pre-amplifier (in fact a Z77 strapped as a triode) on the HF bands above about 10 MHz and on all bandspread bands. Use of pentode RF amplifiers at higher-than-Band II frequencies seem to have been rare, although the Eddystone 770R communications receiver, which tuned up to 165 MHz, used pentode RF amplifier and a pentode mixer, in both cases the 6AK5/EF95, which I understand was one of the lowest noise VHF pentodes. Cheers, |
||
3rd May 2010, 12:05 pm | #54 |
Rest in Peace
Join Date: Sep 2008
Location: Solihull, West Midlands, UK.
Posts: 4,872
|
Re: Valve Questions
At first sight it seems puzzling that people continued to use pentodes at VHF, but it could be some combination of:
- familiarity: the cascode was not invented until the late 1940's and engineers are mostly a conservative lot! - AGC: true variable-mu triodes were late in arriving so signal handling could be compromised, and for domestic purposes it might seem wasteful to triode-connect a perfectly good pentode and then add an extra triode for a cascode (although this is how the first cascode was done with a EF95/6AK5 and ECC91/6J6). - valve noise was still a bit of a mystery in the 1950s **, and although the experts were getting to understand it this might not have trickled down fully to practitioners. Given a pentode mixer, a pentode RF amp is just about good enough at Band I. Our modern ideas about intermodulation and using only as little gain as necessary before the main IF filter were not around back then; Gain was Good!! A pentode mixer also probably helped isolate the local oscillator from the incoming signals, better than a triode mixer. In domestic equipment you couldn't afford an oscillator buffer stage. A rule of thumb I always remember is that at 100MHz antenna noise is about the same as thermal noise, so you need a receiver noise figure of around 3dB. Below that external nose rises roughly proportional to frequency, so you need 10dB noise figure at 30MHz and 20dB at 10MHz. Above 100MHz the noise reduces, but more slowly, so across the UHF range the noise power is about half thermal - an antenna sees a hot earth (300K) and a cold sky (3K?) in roughly equal proportions. Broad-brush perhaps, but it gives a rough idea of what we need to aim at. ** I have a Radio Research report published by HMSO, dated 1951, in which things we take for granted in terms of valve noise are presented as though they are fairly new ideas. Many of the classic papers only appeared in the 1940s. |
9th Jun 2010, 7:35 am | #55 | |
Nonode
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,944
|
Re: Valve Questions
Quote:
Regarding AGC, valves like the ECC84 – which I think is vari-mu - seem to have a somewhat shorter grid base than say the EF85 pentode. Whether that reflects the greater difficulty of designing vari-mu triodes I don’t know. But possibly it might be related to the way that TV receiver AGC systems need to operate. I have a 1957 Mullard article on black-level AGC for positive modulation TV systems. In this example, minimum gain occurs with -10 V AGC voltage on the PCC84 RF amplifier, -13.5 V on the RF85 1st IF amplifier and – 5V on the EF80 2nd IF amplifier. The RF AGC is delayed -3.5 V relative to the IF AGC, and that applied to the EF80 is proportioned relative to that on the RF85. Given the RF AGC delay, it would seem that a shorter grid base for the RF amplifier – whether by valve design or limitations, or by biasing and HT voltage choices – would be preferable. Another interesting comment in Fink is that at the time of writing, there was no commercial use of the cascode RF amplifier at UHF, there being no suitable double triodes available (I imagine interstage shielding could be difficult, as well as getting enough pinouts for the grids), whilst use of two separate triodes would be uneconomic. But somewhere else – in connection with UK early TV UHF tuner practice, which seems to have been more exacting than that in Europe or North America at the time – I recall reading that at the higher frequencies, the earthed cathode input stage would confer but modest benefit, unlike the case at VHF, and so was not justified on performance grounds anyway. Cheers, * The referenced book is: Fink, Donald D., Editor-in-Chief Television Engineering Handbook McGraw-Hill, 1957 LCC 55-11564 (too early for SBN or ISBN) The cascode amplifier article referenced therein (which I haven’t traced) is given as: H. Wallman, A,B. Macnee & C.P. Gadsen, Low-Noise Amplifier, Proc. IRE, vol. 36, p.700, 1948. |
|
9th Jun 2010, 10:13 am | #56 |
Rest in Peace
Join Date: Sep 2008
Location: Solihull, West Midlands, UK.
Posts: 4,872
|
Re: Valve Questions
The AGC balance in a TV would need to be right because there is not so much difference in signal levels between 'not enough' and 'too much' - this is because TV is wideband. The RF amp would have delayed AGC in order to keep up the S/N ratio on weak signals. The 2nd IF amp would have reduced AGC partly because the EF80 is not a variable-mu valve and partly because it needs to have enough anode current to generate a good strong signal for the video detector diode.
The cascode doesn't work at UHF because the first stage, grounded-cathode, doesn't work at UHF. Even a specially designed triode like EC900 can only drop Cag down to around 0.5pF so it is only good up to a few hundred MHz. Beyond that grounded-grid is the only option for 'normal' valves. Alternatives were more exotic like travelling wave tubes or parametric amplifiers. That is why for many years UHF receivers went straight to a crystal mixer, followed by a low noise IF amplifier. Even at slightly lower frequencies something similar was done - the EAC91 was intended as a high VHF front-end (triode oscillator and diode mixer) with no RF stage. As a general rule, a given technology can oscillate and mix at a higher frequency than it can amplify with sufficienty low noise. I have a pdf of the Wallman article, although I can't remember where I obtained it. PM me if you are interested. Last edited by G8HQP Dave; 9th Jun 2010 at 10:16 am. Reason: add note on Wallman pdf |
9th Jun 2010, 12:38 pm | #57 |
Octode
Join Date: Apr 2007
Location: Hampshire, UK.
Posts: 1,576
|
Re: Valve Questions
This is the most interesting thread for ages!
One set that seems to have been missed out of the discussions is the Pye FenMan II, which from memory uses and EF80 and and ECF80 in it's FM front end. Pye by this stage were very much aware of the ECC85 (it is used in the FenMan I for example) but yet chose this non-standard arrangement for their larger model. What could the reason have been? I suspect it may have been "to use as many valves as is possible", this being the common theme of the rest of the design... |
12th Jun 2010, 12:55 am | #58 |
Nonode
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,944
|
Re: Valve Questions
G8HQP Dave:
Thanks for your comments. In respect of the cascode at UHF, you have restored my previous understanding. I thought that the comment in Fink was a bit odd, but I did not want to lightly set aside information from such a well-respected author. And thanks for the clarification on TV AGC systems. Interestingly, the Mullard circuit I referred to had a 3-stage IF, the 3rd stage being another EF80 without AGC. Given that the RF AGC needs to be delayed, then the total voltage available will be reduced by the delay voltage, which in turn indicates that whatever type of vari-mu valve is used for the RF amplifier, ideally it would have a slightly shorter grid base than the valve used for the 1st IF amplifier. Maybe there was some serendipity, if it was the case that the difficulty in designing vari-mu triodes meant that they couldn’t easily reach the same relatively wide grid bases that were available with vari-mu pentodes. Perhaps also the earliest double-triodes developed for TV cascode use in the USA were not really of the vari-mu type, in which case one might expect that the AGC voltage to them was both delayed and proportioned (or perhaps proportioned and then delayed in that order) to accommodate their relatively short grid bases. Studio263: Perhaps the Pye Fenman II design was about valve count. But Pye might also have been using some of the circuitry it had developed for the hi-fi equipment market, such as for its Mozart FM tuner. I don’t know how the chronology of the Fenman II aligns with that of the Mozart, but according to a 1959 WW advertisement – with some deduction in aligning the stated valve line-up with the stated functions, the Mozart had an EF80 RF amplifier with an ECF80 serving as frequency changer and AFC reactance valve. I would expect that the pentode in the ECF80 was configured as a self-oscillating frequency changer, with the triode as the reactance valve, but conceivably it could have been the other way around. The IF strip had an EF89 1st IF, followed by two EF80 limiters, and a crystal diode Foster-Seeley discriminator. I think that this circuit was reasonably representative of British FM tuner practice of the period, inclusive of the preference for the pentode RF amplifier, which refers back to earlier commentary in this thread. But then Pye, given its TV experience, should have had no difficulty in using an ECC84-based cascode RF amplifier had it chosen so to do. If the Fenman II did not have AFC, then it could have been that the ECF80 was configured as a mixer with separate oscillator, per TV VHF tuner practice. Cheers, |
13th Jun 2010, 1:53 pm | #59 | |
Rest in Peace
Join Date: Sep 2008
Location: Solihull, West Midlands, UK.
Posts: 4,872
|
Re: Valve Questions
Quote:
The grid base of vari-mu RF pentodes covers a wider range than vari-mu triodes mainly because the gain covers a wider range too. You can get a rough figure of merit by dividing the normal anode current by the normal gm (to get a voltage = projected cutoff if valve was linear), then dividing that into the actual cutoff voltage (or 1% gm point). I have not done a survey, but I suspect you will find that triodes and pentodes then look more alike. There are no low gain vari-mu triodes, corresponding to pentodes like EF92 or EF93, because the intended application for triodes was VHF where high gain is needed for low noise. In the 1960's when GEC wanted to develop a receiver for the Royal Navy, they used a cascode ECC82 as the RF amplifier. I don't know if they applied AGC. I reckon a low gain version of ECC189, if it existed, would have been better although ECC82 is quite close to being a square-law device so gives low 3rd order IM. As an alternative they could have tried a triode-connected EF92 as the first stage in the cascode. |
|
15th Jul 2010, 7:56 am | #60 |
Nonode
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,944
|
Re: Valve Questions
In keeping with the recent cascode theme of this thread, and as the subject of Mazda valves has come up in the thread “replacement Valves”; https://www.vintage-radio.net/forum/...ad.php?t=56282, attached are a couple of Mazda advertisements from Wireless World October, 1961.
I would imagine that the 30L17 was pretty much at the tail end of cascode valve development, at least for consumer applications, and would have had better all-round performance than the PCC84/ECC84, including in respect of AGC capabilities. The 30C17 triode pentode TV frequency changer was unusual in that the pentode was designed for AGC. Since the quoted gain reduction point is when conversion conductance is 10% of its full value, it looks as if the AGC action was milder than for the 30L17, where mutual conductance is quoted at 1% of full value point. Presumably the 30C17, when used, would be connected to the (delayed) RF AGC line along with the RF amplifier. Did the 30C17 have any PCF or ECF counterparts? Or was it a single-sourced item? Cheers, |