Re: Another infinite impedance detector scheme

[Tom WA3KLR]

Those Radio-Electronics issues have an amazing amount of information in them.

Infinite impedance circuit - Detector and phase inverter all in one.

Also of note in this issue are several articles by Rufus Turner who we talked about in another topic a year or so ago. On page 38 is his article on making a PNP transistor out of a 1N34 diode; 1949, amazing!

[K9MB]

I love all the old Radio Craft issues available on the site and when it became Radio-Electronics, it continued for several years. It is like reading he history of the development of the Radio Art.
I saw my first PNP germanium transistors in 1962. It was a Raytheon CK722.
It came in a kit that used a 1n34 detector and the CK722 was an audio amplifier.
 Y brother got the kit for Christmas, but he was not interested in it. I bought a Knight Star Roamer regen receiver and two years later, got my Novice and a Knight T60 and Lafayette HE-30.
I took Radio Electronics and then QST and a life of solder smoke…
That crystal set with a germanium transistor audio amp led me to the “hard stuff” 😉😂😂

[Pete, WA2CWA]

HERE IS THE ORIGINAL START TO THE THREAD:

See page 76.

https://worldradiohistory.com/Archive-Radio-Electronics/40s/Radio-Electronics-1949-05.pdf

[KD6VXI]

Radio Electronics was my favorite magazine growing up.


Every year my grandmother gave me a subscription for my birthday.


LOTS of great memories reading that magazine!


--Shane
WP2ASS / ex KD6VXI

[Opcom]

I have a question about the article/circuit. Won't the huge 100K cathode bias/load resistor make it necessary to raise the grid voltage substantially above ground to keep the tube from being cut off so hard that small signals might not tickle it?

The circuit looks like the 'accordion' phase inverter used in audio amps, that is usually direct coupled to the preceeding stage's plate with about +50V on it.

Would adding a regular cathode bias resistor as shown in the Resistance Coupled Amplifiers sections of tube manuals make any sense, or am I missing some principle of operation? Sorry to be stupid if so.

[K9MB]

I have a question about the article/circuit. Won't the huge 100K cathode bias/load resistor make it necessary to raise the grid voltage substantially above ground to keep the tube from being cut off so hard that small signals might not tickle it?

The circuit looks like the 'accordion' phase inverter used in audio amps, that is usually direct coupled to the preceeding stage's plate with about +50V on it.

Would adding a regular cathode bias resistor as shown in the Resistance Coupled Amplifiers sections of tube manuals make any sense, or am I missing some principle of operation? Sorry to be stupid if so.

Hi Patrick,
It is very audacious of you to claim all the stupidity here. I deserve a large helping myself, so you must share…😉😂

Seriously, I also can see, after you pointed it out that the 6J5 is turned of hard at -140volts below the cathode.
I have added the curve for a 6J5 below:

If the output is balanced at 140 volts each side, then there must be a 140volt drop across each 100k resistor, though the plate resistance will raise the R,,depending on the grid to cathode bias.
140v/100k = 1.4mA
On the curve, this looks like a grid to cathode bias of about -7 to -8 volts.
So your Rk needs to be the value that puts the cathode 5 volts above the  grid
5-7/0014= 4300-5000 ohms??
There- you see, my idea surpasses yours in pure, half baked stupidity-right?? 😉😁 73, Mike

Ps: Kidding aside, I have very often wondered if anyone ever actually built up a lot of published circuits. It does not mean that adding parts as you suggested vould not fix it, however, and I believe that you make a good point. MB

[Opcom]

I thought so, but ya never know about little-known odd circuits and unusual ways of using components. 

The audio people might say that the Rk should be part of the 100K for balance sake, if Rk is bypassed, but not on Thursdays.

As published, maybe the current through the tube is 50 microamperes (5V/100K), so the conflict is, why use the extreme cutoff part of the curve instead of regular current for a 6J5, and will it actually work, and is that better than a conventional tube current level regarding noise and everything else? And the part of the curve down there at 50uA reminds me of a manufacturer's reply "we don't guarantee operation under those conditions".

I seldom build circuits where too much oddness isn't explained. Having been stung a few times after many hours invested. LOL then in the next issue an obscure correction is made to the scamatic. What others do? Common sense should weigh in, but it's full circle to an oddball application.

But then there are things like the Starved Circuit amplifer. Not the same but very weird.
https://www.electronicdesign.com/technologies/analog/article/21768551/the-starvedcircuit-amplifier-is-revived-in-a-transistor-version

(apparently no schematic LOL)
But curiosity is piqued (peaked) as to whether a simulation could shed any light.

[K9MB]

I thought so, but ya never know about little-known odd circuits and unusual ways of using components. 

The audio people might say that the Rk should be part of the 100K for balance sake, if Rk is bypassed, but not on Thursdays.

As published, maybe the current through the tube is 50 microamperes (5V/100K), so the conflict is, why use the extreme cutoff part of the curve instead of regular current for a 6J5, and will it actually work, and is that better than a conventional tube current level regarding noise and everything else? And the part of the curve down there at 50uA reminds me of a manufacturer's reply "we don't guarantee operation under those conditions".

I seldom build circuits where too much oddness isn't explained. Having been stung a few times after many hours invested. LOL then in the next issue an obscure correction is made to the scamatic. What others do? Common sense should weigh in, but it's full circle to an oddball application.

But then there are things like the Starved Circuit amplifer. Not the same but very weird.
https://www.electronicdesign.com/technologies/analog/article/21768551/the-starvedcircuit-amplifier-is-revived-in-a-transistor-version

(apparently no schematic LOL)
But curiosity is piqued (peaked) as to whether a simulation could shed any light.

Hi Patrick,
I have never heard of the extreme cutoff curve, though the curves seem to imply an asymptote, which may mot ever reach perfect cutoff. I have always that must leakage current.
That starved amplifier seems to be associated with generating distortion.
Unless you are a heavy metal guitarist, seems useless.
I thought that the grid bias should be adjustable to maintain balance.
The grid is negative of the cathode, so impedance very high, so no loading-right?
Interesting and weird circuit…..

[DMOD]

I thought so, but ya never know about little-known odd circuits and unusual ways of using components. 

The audio people might say that the Rk should be part of the 100K for balance sake, if Rk is bypassed, but not on Thursdays.

As published, maybe the current through the tube is 50 microamperes (5V/100K), so the conflict is, why use the extreme cutoff part of the curve instead of regular current for a 6J5, and will it actually work, and is that better than a conventional tube current level regarding noise and everything else? And the part of the curve down there at 50uA reminds me of a manufacturer's reply "we don't guarantee operation under those conditions".

I seldom build circuits where too much oddness isn't explained. Having been stung a few times after many hours invested. LOL then in the next issue an obscure correction is made to the scamatic. What others do? Common sense should weigh in, but it's full circle to an oddball application.

But then there are things like the Starved Circuit amplifer. Not the same but very weird.
https://www.electronicdesign.com/technologies/analog/article/21768551/the-starvedcircuit-amplifier-is-revived-in-a-transistor-version

(apparently no schematic LOL)
But curiosity is piqued (peaked) as to whether a simulation could shed any light.

We have to keep in mind this a combination of a Class A amplifier AND a cathode follower. The AC cathode voltage will be about 0.9 X the grid AC voltage.

If one takes into account the 20k resistor in series with the PS, the numbers show the plate node would be about 125V (Vp) and 150V at the cathode (Vk) for 1.5mA IP and IK.

That is, 1.5mA X 20k = 30V drop. 280V - 30V = 250V; 250V/2 = 125V.

The major problem I see, if this is the case, is that if there is a Cathode Voltage of 125V to 150V, this would be about 35 to 60 volts over the Cathode-to-Filament spec of 90 volts.

I question how long the tube was actually able to operate.

A solution would have been to add another resistor in the cathode circuit and a 10 Mohm resistor to the grid to bring the bias into the class A operating region, a "TOO-High-Bias" circuit solution.

[K9MB]

I found this article on Infinite Impedance detectors from 1948 in an Aussie publication.
It seems obvious to me that this type of detector functions best with solid signals of higher amplitude to keep distortion low…
No panacea here…  

Edit: I just remembered that Jim Tonne wrote an excellent article on demodulators and he ran curves on a lot of them.
The infinite impedance detector indeed has lower distortion at higher amplitudes.
The durprise to me was the excellent performance of the Military ARC-5 detector circuit on all models above 1.5mHz.
They are low distortion and handle a lot of signal at low distortion numbers that appears to beat everything that Jim looked at.


http://tonnesoftware.com/appnotes/demodulator/EnvelopeDemodulators.pdf