UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > General Vintage Technology > Components and Circuits

Notices

Components and Circuits For discussions about component types, alternatives and availability, circuit configurations and modifications etc. Discussions here should be of a general nature and not about specific sets.

Closed Thread
 
Thread Tools
Old 28th Oct 2017, 10:43 am   #1
G6Tanuki
Dekatron
 
G6Tanuki's Avatar
 
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 13,951
Default Infinite Impedance detectors.

The thread about the Armstrong circuit using a non-AGC-controlled triode stage feeding the diode detector has reminded me of one underlying question I've had about the design of AM detectors in 'upmarket' receivers: why so few of them use the 'infinite impedance' detector?

The II detector looks at first like a cathode-follower but it's not Class-A biased and so it detects. The two big features are (1) it provides an essentially infinite impedance to the preceding stage (hence the name) and (2) it also has what amounts to 100% audio negative-feedback.

(1) means that the secondary-winding of the IFT driving it can be designed for higher loaded Q and so deliver a higher IF voltage to the grid of the triode, meaning higher potential audio output.

(2) means low distortion.

The only downside of it that I can see is that it doesn't provide AGC voltage - but I *have* seen a published design using a pentode in which the cathode/control-grid/screen-grid form the triode for the II detector, and the amplified IF signal appearing at the anode is then rectified using a diode to provide AGC: I guess this is modelled after the pentode "electron-coupled oscillator" where the screen-grid serves as the oscialltor anode?

So - why wasn't the II detector more popular?
G6Tanuki is offline  
Old 28th Oct 2017, 12:16 pm   #2
ukcol
Rest in Peace
 
ukcol's Avatar
 
Join Date: May 2003
Location: Harlaxton, Lincolnshire, UK.
Posts: 3,944
Default Re: Infinite Impedance detectors.

I cannot suggest why the infinite impedance detector wasn't more popular except to guess that it may have been for reasons of cost.

However going very slightly off topic if I may. Many, including me, may not be particularly familiar with the II detector, (probably precisely because of its lack of popularity) so here is an excellent explanation of how the II detector works from Al.

Quote:
Originally Posted by Skywave;
The conventional valve-based infinite impedance detector is a derivation of the anode-bend detector. However, in the infinite impedance detector, the output is taken from the cathode, unlike the anode bend detector, where the output is taken from the anode. Hence, the infinite impedance detector is a form of a cathode follower (CF). Therefore the input Z is extremely high owing to the 100% -ve feedback and as with a CF, the anode needs to be decoupled: in this case, to RF and AF. The resultant signal at the cathode will contain RF and AF components, so a filter at the cathode is necessary to remove the RF components. It is usual to add more filtering than a simple capacitor across the cathode resistor, as in your example***. Note that there is a critical minimum value for that capacitor. If its capacity is too small, the input to the valve will appear to be a negative resistance when 'seen' by the IFT and the whole stage will probably become unstable. The output Z of a CF is approximately equal to 1/gm, where gm = the mutual conductance of the valve, and in the infinite impedance detector, the value of the cathode resistor should be fairly high compared with 1/gm: something like 20 kΩ to 47 kΩ is not uncommon.
Since the infinite impedance detector is a form of a CF, its voltage gain will be less than unity, but on account of its very high input Z it will present a very minimal damping on the tuned cct. of the last IFT, and as such, will provide a greater output from that IFT compared to the output that would be obtained if that same IFT fed a simple diode detector.
*** Al is referring to the HBR 16 receiver here.

Last edited by ukcol; 28th Oct 2017 at 12:30 pm.
ukcol is offline  
Old 28th Oct 2017, 12:59 pm   #3
ukcol
Rest in Peace
 
ukcol's Avatar
 
Join Date: May 2003
Location: Harlaxton, Lincolnshire, UK.
Posts: 3,944
Default Re: Infinite Impedance detectors.

Circuit of the II detector used in the HBR 16 receiver.
Attached Thumbnails
Click image for larger version

Name:	II detector cct.jpg
Views:	575
Size:	21.0 KB
ID:	151522  
ukcol is offline  
Old 30th Oct 2017, 8:55 pm   #4
G6Tanuki
Dekatron
 
G6Tanuki's Avatar
 
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 13,951
Default Re: Infinite Impedance detectors.

I wonder perhaps if the infinite-impedance detector was hampered in its acceptance in 'high quality' receivers because the high unloaded 'Q' of the tuned-circuit feeding it would have provided too-sharp a response for 'hi-fi' audio reception unless the IFT was deliberately damped (thereby losing the benefit of the II detector!) ?

Equally, the lack of 'free' AGC from a classic diode detector would have made it less popular, probably needing an extra valve compared to the usual Double-Diode-Triode-and-Beam-Tetrode audio lineup used by pretty much every broadcast-radio from about 1935 onwards.
G6Tanuki is offline  
Old 30th Oct 2017, 10:52 pm   #5
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: Infinite Impedance detectors.

Yes, the undesirability of overly high Q in the final IFT could have been a factor mitigating against the use of the infinite impedance demodulator in high quality receivers. If one looks at British hi-fi AM tuners of the 1950s or so, then the diode demodulator looks to have been the majority choice.

Even so, a couple of examples with infinite impedance demodulators come to mind. One was the Rogers RD Junior of 1952. This had a 12AH8 mixer-oscillator, 9D6 IF amplifier and 6C4 demodulator. Amplified and delayed AGC was accomplished through a 6AM6 sidechain IF amplifier and a crystal diode. IF bandwidth was variable, with 7, 15 and 23 kHz positions being available. So, the final IFT must have been reasonably well damped in the widest position. One may wonder why not a diode-pentode for the AGC circuit, but perhaps it was because Brimar did not offer one at the time.

The other was the Armstrong AM44 of circa 1955. It was sometimes described as having both an infinite impedance demodulator and a cathode follower output, but I suspect that the apparent two stages were one and the same. There was three-position switched variable bandwidth, although I do not know the actual numbers. I suspect though that the widest position would have been 20 kHz or so.

The inherent low impedance output of the infinite impedance demodulator was a potential advantage in a tuner that might be used with a variety of amplifiers of differing input impedances. With diode demodulators, the load impedance they looked into had to be sufficient so as not to materially lower the AC-to-DC load ratio and so introduce distortion. A separate cathode follower stage after a diode was always a possibility, but this required a triode or a pentode with an “uncombined” cathode, precluding the use of standard European double diode-triodes or double diode-pentodes, which had “combined” cathodes. So an additional valve, such as a 6C4, would have been needed anyway.

Whatever was the reason for Armstrong’s choice of the infinite impedance demodulator for the AM44, it reverted to a diode demodulator – in this case driven by a class A triode – for its ST3 of 1958.


Cheers,
Synchrodyne is offline  
Old 30th Oct 2017, 11:32 pm   #6
ukcol
Rest in Peace
 
ukcol's Avatar
 
Join Date: May 2003
Location: Harlaxton, Lincolnshire, UK.
Posts: 3,944
Default Re: Infinite Impedance detectors.

I can see that an IF transformer designed to feed a low impedance load (i.e. a conventional diode detector) would be unsuitable for use with a high impedance load for the reasons already given. However, is there a need to design a special IF transformer to feed a high impedance detector? The set will already have a transformer designed to feed a high impedance load and that transformer will already have the appropriate Q and bandpass characteristics. I'm referring, of course, to the first IF transformer.

Perhaps I'm missing something vital here?
ukcol is offline  
Old 30th Oct 2017, 11:58 pm   #7
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: Infinite Impedance detectors.

Usually with high-quality AM receivers, the IFTs were designed to work as a team, so each had Q, coupling factor and bandwidth(s) chosen to that end. Sometimes a critically coupled IFT was used to fill in the response centre-dip from an overcoupled IFT, either before or after it. Each maker seemed to adopt a somewhat different approach. Most accessible for study are the Quad AMII and Chapman S6 circuits, both well covered in the literature.

Possibly repetition of the 1st IFT ahead of an infinite impedance demodulator would work, but it would be unlikely to be "right" in terms of overall bandwidth, etc.

Cheers,
Synchrodyne is offline  
Old 31st Oct 2017, 9:57 am   #8
ukcol
Rest in Peace
 
ukcol's Avatar
 
Join Date: May 2003
Location: Harlaxton, Lincolnshire, UK.
Posts: 3,944
Default Re: Infinite Impedance detectors.

Please bare with me as I still have not quite got it and am still missing something.

Say you had a quality AM receiver with just two IF transformers each of which each of which had a -3dB bandwidth of +/- 9kHz. Why would the overall bandwidth come out wrong? I suspect a mathematical answer is coming.
ukcol is offline  
Old 31st Oct 2017, 1:02 pm   #9
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Arrow Re: Infinite Impedance detectors.

In a phrase, cumulative frequency-sensitive attenuation.

Al.
Skywave is offline  
Old 31st Oct 2017, 2:12 pm   #10
G8HQP Dave
Rest in Peace
 
Join Date: Sep 2008
Location: Solihull, West Midlands, UK.
Posts: 4,872
Default Re: Infinite Impedance detectors.

It is conceivable that the first IF transformer might have a special primary, to provide the load expected by the frequency changer. Too much signal here might cause nonlinearity.

There is more to IFT specification than -3dB frequency. A typical off-the-shelf IFT will probably be slightly undercoupled, to get a good shape combined with ease of adjustment - but of course this assumes a particular load and source impedance. Factory-built radios did not need to use off-the-shelf IFTs so variations are possible.

The infinite impedance detector is nearer to being an ideal envelope detector than a diode detector, but it still suffers from the limitations of an ideal envelope detector. It still can't do negative peaks at 100% modulation.
G8HQP Dave is offline  
Old 31st Oct 2017, 2:35 pm   #11
Bazz4CQJ
Dekatron
 
Bazz4CQJ's Avatar
 
Join Date: Apr 2009
Location: Oxfordshire, UK.
Posts: 4,923
Default Re: Infinite Impedance detectors.

Am I right that the HBR 16 doesn't have a filter in the IF chain, nor any provision to change the selectivity. Is it the case that the design is intended to achieve a fairly tight "one size fits all" bandwidth for a comms reciever, and the II is part of that strategy?

Don't suppose anyone has an "Anglocised" version of the circuit diagram?

B
__________________
Saturn V had 6 million pounds of fuel. It would take thirty thousand strong men to lift it an inch.
Bazz4CQJ is offline  
Old 31st Oct 2017, 3:41 pm   #12
ukcol
Rest in Peace
 
ukcol's Avatar
 
Join Date: May 2003
Location: Harlaxton, Lincolnshire, UK.
Posts: 3,944
Default Re: Infinite Impedance detectors.

Quote:
Originally Posted by Skywave View Post
In a phrase, cumulative frequency-sensitive attenuation.
Does this mean the -3dB point becomes -6dB overall.
ukcol is offline  
Old 31st Oct 2017, 6:25 pm   #13
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Arrow Re: Infinite Impedance detectors.

I wouldn't go so far as to be specific, as in "-6dB overall", but that seems to be a possible result. The point is this: each stage will have a BW of -3dB, but the overall BW will depend on the manner in which the two stages are coupled. For example, if we have an amplifier whose stated (& measured) BW is -3dB, that figure will obviously been obtained when that amp. is connected to a known (and presumably designed) load. If the O/P of that amp. is then fed to a similar amp., that second amp will need to present the same load value. Any discrepancy will produce a mis-match affecting the overall BW. (By "load", I am referring to a load of the form R ± jX).

Well that's my understanding, anyway.

Al.
Skywave is offline  
Old 31st Oct 2017, 7:01 pm   #14
ukcol
Rest in Peace
 
ukcol's Avatar
 
Join Date: May 2003
Location: Harlaxton, Lincolnshire, UK.
Posts: 3,944
Default Re: Infinite Impedance detectors.

OK thanks Al. I think I'm starting to get it now but I'm not quite there yet.
ukcol is offline  
Old 6th Nov 2017, 10:42 pm   #15
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: Infinite Impedance detectors.

Another AM tuner with an infinite impedance demodulator was the GEC model for home construction, circuit details attached:

Click image for larger version

Name:	Molloy p.190 GEC AM Tuner.jpg
Views:	365
Size:	58.1 KB
ID:	151905Click image for larger version

Name:	Molloy p.191 GEC AM Tuner.jpg
Views:	186
Size:	44.5 KB
ID:	151906

It appears to have used standard Weyrad IFTs. But the means by which the 1st IFT was provided with variable bandwidth was unusual. At a guess I’d say that in the “wide” position, the primary and secondary coils had their resonant frequencies displaced somewhat in opposite directions (along with additional damping for the secondary) in order to create the required double-humped response. More usually the coupling was varied, typically with a switched tertiary coil, but sometimes by a switched bottom capacitance network. But as with conventional practice, likely the 2nd IFT would have infilled the gap-between-the-humps of the 1st IFT, for which it required, by the conventional wisdom, half the loaded Q. So, its Q was likely determined more by circuit needs rather than what the infinite impedance demodulator would allow.


Cheers,
Synchrodyne is offline  
Old 7th Nov 2017, 12:26 pm   #16
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,059
Default Re: Infinite Impedance detectors.

Quote:
Originally Posted by ukcol View Post
Quote:
Originally Posted by Skywave View Post
In a phrase, cumulative frequency-sensitive attenuation.
Does this mean the -3dB point becomes -6dB overall.
Basically, yes (that's assuming that the two are connected in cascade, there's no parasitic feedback around the whole thing, etc etc).

So the 3db points would be somewhat narrower. How much narrower depends on the shape of the response curves of the original IFT's. If these were ideal, square-topped, brick-wall responses, there wouldn't be any narrowing at all!

Quote:
Originally Posted by G8HQP Dave View Post
The infinite impedance detector is nearer to being an ideal envelope detector than a diode detector, but it still suffers from the limitations of an ideal envelope detector. It still can't do negative peaks at 100% modulation.
True, but it could get there more nearly. And it doesn't have the limitations of AC/DC load ratio of less than 1, causing a limit to amplitude modulation depth without distortion.

It's also worth considering, a HF push-pull input, maybe a double-triode with cathodes strapped together. That would allow a halving of the cathode capacitor value for the same degree of HF ripple, so the audio frequency at which non-linear distortion started to kick in would be doubled.

Last edited by kalee20; 7th Nov 2017 at 12:27 pm. Reason: Typos
kalee20 is offline  
Old 7th Nov 2017, 2:00 pm   #17
ukcol
Rest in Peace
 
ukcol's Avatar
 
Join Date: May 2003
Location: Harlaxton, Lincolnshire, UK.
Posts: 3,944
Default Re: Infinite Impedance detectors.

Quote:
Originally Posted by kalee20 View Post
It's also worth considering, a HF push-pull input, maybe a double-triode with cathodes strapped together. That would allow a halving of the cathode capacitor value for the same degree of HF ripple, so the audio frequency at which non-linear distortion started to kick in would be doubled.
I asked a question about the full wave version of the diode detector used in the Murphy A8 back in 2008.

http://www.vintage-radio.net/forum/s...ad.php?t=28053

Perhaps this is why Murphy chose to use a full wave detector particularly in view of the low IF frequency used. The cathode capacitor in your statement above equating to the capacitor across the diode load in the diode detector.
ukcol is offline  
Old 8th Nov 2017, 10:37 pm   #18
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: Infinite Impedance detectors.

Heathkit UK used a full-wave AM demodulator in its AFM-1 hi-fi tuner of 1962. In respect of that it claimed:

“A push-pull detector circuit maintains the symmetrical loading of the driver stage; it uses two germanium diodes connected in such a way that one conducts for each half of the R.F. waveform. This reduces both R.F. and audio distortion and, since the R.F. ripple frequency is doubled from 470 to 940 Kc/s, filtering of the I.F. signal is more effective.”

Although the AFM-1 was a UK-only model, this form of AM demodulator appears to have been carried over from established Heathkit US practice, having been used for example in the PT-1. Heathkit US also used two-diode voltage-doubling AM demodulators in some of its models.

On the other hand, Sturley had this to say about full-wave (double-wave) demodulation:

“It has been shown that only under special circumstances can double-wave detection prove superior to half-wave detection, and the latter is almost universally employed in receivers.”

No elaboration was provided, but there was a reference to an article by W.T. Cocking in “Wireless Engineer” for 1935 November.

Regarding the infinite impedance demodulator and loading effects, Langford-Smith observed:

“There is seldom any difficulty with a.c. / d.c. ratios as the grid resistance of the following a-f stage is generally about 0.5 megohm.”

This suggests that although it may have been less sensitive to the AC/DC load ratio than typical diode demodulators, it was not immune.

When all is said and done, I suspect that the devil was in the detail, and that circuit implementation had as much, perhaps more weighting in terms of performance outcomes than choice of basic circuit. Certainly, many of the hi-fi makers who put some effort into obtaining good AM performance chose conventional diode demodulators, amongst them Chapman, Dynatron, Jason and Quad. And wth the exception of the AM44, Armstrong used diode demodulators. As well as demodulator distortion, some attention was usually given to minimizing differential and modulation rise distortions in these designs.

The setmakers in general may have been less concerned about AM performance, particularly after the advent of FM-AM receivers. But the Murphy A188C was an example of an AM-only receiver in which better-than-typical performance was self-evidently an objective. It included variable selectivity with a wide audio bandwidth that I’d guess was around 10 kHz. It had a conventional diode demodulator. But the export TA160 model of the same era had a double-diode, full-wave demodulator. That was counter-intuitive. The TA160 had single selectivity, probably tailored for shortwave reception, and I doubt that the IF bandwidth allowed an audio response much beyond 5 kHz. So post-demodulator filtering would have been an easier proposition in the TA160 than in the A188C, with the latter perhaps more in need of the potential benefits of a full-wave demodulator. Different designers with different ideas, perhaps? Or maybe since the TA160 used two 6LD20 double-diode triodes, in order to provide two AF triodes, someone thought that they may as well make some constructive use of the surplus of diodes that were available as a result.


Cheers,
Synchrodyne is offline  
Old 9th Nov 2017, 9:08 am   #19
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,059
Default Re: Infinite Impedance detectors.

The 'special circumstances,' though, are not mentioned by Sturley! He's an author of pedigree, but it's still appropriate, as good engineers, to challenge and try to understand WHY.

R fervency AC/DC load ratio, with a normal diode detector, it's quite possible to compensate for this by appropriately biasing the diode. The down side is that the bias needed depends on the mean carrier strength, so unless it's derived from a (non-delayed) AVC voltage, it is difficult to get it right. And the load on the final IFT could no longer be represented by an equivalent resistance, meaning that bandwidth would be dependent on signal level, giving hard-to-calculate performance.

With an infinite impedance detector, AC/DC load is not an issue as long as there is a sufficient cathode current drain to allow the cathode capacitor voltage to decay between HF cycles so that it can always follow downward modulation. Upward is not an issue because the valve charges the capacitor with little gulps of current. And load on the preceding IF transformer is infinite anyway so who cares if it's a signal-dependent infinity?
kalee20 is offline  
Old 9th Nov 2017, 9:31 am   #20
ukcol
Rest in Peace
 
ukcol's Avatar
 
Join Date: May 2003
Location: Harlaxton, Lincolnshire, UK.
Posts: 3,944
Default Re: Infinite Impedance detectors.

Quote:
Originally Posted by kalee20 View Post
..........signal-dependent infinity?
I like it.
ukcol is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 2:21 pm.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.