Parallel Wired JFET Amplifier with Transformer

Thread Starter

Sir Kit

Joined Feb 29, 2012
188
If anyone is still checking this thread ... I was reviewing bordodynov's plot and wondering if the output should not appear inverted from the input instead of roughly in phase. Or am I wrong? Thanks for any comments.

Screenshot_20221208_203215.png
 

DickCappels

Joined Aug 21, 2008
10,153
This would have been the first time I've seen Bordodynov make a mistake.

The waveforms look fine. Out is inverted at the drains but the transformer corrects the polarity. Out2 is a source follower and has the same polarity as the gate as source follower are normally expected to behave.
 

Bordodynov

Joined May 20, 2015
3,177
Sir Kit only asked two questions:

Can anyone please tell me why the JFET's are wired in parallel?
(Answer: This increases the drain amps per gate volts, Gm)

What might be the purpose of the air core transformer?
Still a mystery.

Please stick to the questions
I didn't make any mistake. Look at the original diagram. It has the numbering of the transformer pins. But I can change the pins of the secondary winding. I will do that a little later. I have to get home.
 

Thread Starter

Sir Kit

Joined Feb 29, 2012
188
Thank you for the additional illustration. If I understand the first example correctly, phase is inverted at the drain (primary side), but restored to in-phase by the inverted transformer secondary.

But I would have expected more phase shift in the second example where the transformer windings are in-phase. What might be the reason for this?
 

Thread Starter

Sir Kit

Joined Feb 29, 2012
188
I received the plot below from another source who proposed that the negative feedback via the transformer could be intended to make the circuit oscillate.

Any comments?

1668217729496.png
 
I received the plot below from another source who proposed that the negative feedback via the transformer could be intended to make the circuit oscillate.

Any comments?
Hi,

Well, the above schematic shows actually positive feedback (as per the dots at the coils are put) to form an oscillator indeed. The input pulse seems to stop self oscillation though, simulation could be done when the input pulse is not present and the gate is connected to the negative supply rail, to see if self oscillation can start. Maybe the 50 Ohm source resistance could be increased to have a better DC bias for the jfets.

Regarding an earlier question : "What might be the purpose of the air core transformer?" I think this depends only on the frequency range used, so preferable in the range from several ten MHz to several hundred MHz where air cored coils in mutual coupling can be practical. However, in most such amplifiers the coils are wound on toroidal ferrite cores, especially when wide band frequency coverage is required (air cored coils are limited in this respect).

Basically such circuit was conceived in the early 1970s as loss less negative feedback amplifier, by its possible inventor David E. Norton.
You can get more pieces of information in the following papers and a link:
https://www.okdxf.eu/lankford/Noiseless Feedback Norton Amplifiers.pdf
https://www.researchgate.net/public...sless_Feedback_Amplifier_for_LWA_Applications
https://qsl.net/yo4hfu/Files/IMD/High dynamic range Input RX.pdf
http://www.ke5fx.com/norton.htm

Greetings

Winni
 

Thread Starter

Sir Kit

Joined Feb 29, 2012
188
Thank you for your reply. The propensity for this circuit to oscillate was not mentioned previously in this thread.

Would the circuit self-oscillate if the "input" was a ground referenced, tunable tank circuit (in the MHz range) that produced little or no signal?

In your opinion, is it likely oscillation was intended by the designer given that the transformer is air core and not ferrite?
 

Thread Starter

Sir Kit

Joined Feb 29, 2012
188
As I read post #2, it referred to the input signal being AM, not AM due to self-oscillation of the amp.

If I got that wrong, please explain your reasoning. Thanks.
 
Thank you for your reply. The propensity for this circuit to oscillate was not mentioned previously in this thread.
Well, okay and I did not mean that for the original circuit drawing but for the one in reply #27 where the phasing dot indicators are on the top of both coils. Member Bordodynov showed simulations for both cases of the phasing dot positions in his reply #25.
The positive feedback case was when the dots were on the top ends of the coils but we do not see oscillator behaviour at the output (only a reduced amplitude as blue Vout2 in the bottom right plot). Probably the simulator software did not find correct condition for self oscillations I suppose.

Would the circuit self-oscillate if the "input" was a ground referenced, tunable tank circuit (in the MHz range) that produced little or no signal?
No it would not, IMHO. If the parallel LC tank would connect the common gates to the ground and no input signal to the tank is given (if you meant this, please correct if not), then such tank would isolate the gates from the ground at and near its resonant frequency with a relatively high impedance. So at this single frequency the amplifier would have a floating gate RF wise, making self oscillation hard to come by at least at that frequency. Of course at many other frequencies the tank may connect the gates to the ground (having low impedance) and the circuit may self oscillate at a certain frequency established by the air cored coupled coils. 12T on 3/4" dia. air core (from your reply #10) may represent around 2.8 uH inductance for each coil as per this calculator https://coil32.net/online-calculators/one-layer-coil-calculator.html
So at what frequency these 2 coils would couple and constitute an oscillator would depend on the some pF self and stray capacitances and the direct circuit enviroment they are embedded. Besides, there is the question for self oscillation conditions for the circuit like the amplifier's amplification factor > 1 etc.

In your opinion, is it likely oscillation was intended by the designer given that the transformer is air core and not ferrite?
Well, it does not depend on that. The circuit can operate as an oscillator either with air or ferrite cored coils, this depends first of all on the coils phasings. You can read on this here in this paper
https://www.okdxf.eu/lankford/Noiseless Feedback Norton Amplifiers.pdf page 1 at the bottom:
"The phasing dots of the transformer in Fig. 1 should be observed. A BJT CBTF amp uses negative feedback, so
that reversal of the phasing of the feedback link n1 would provide positive feedback, which would likely cause
the amp to oscillate, and in any case would change amp gain, and degrade IMD performance and two way
impedance match to Z0."


Winni
 

Thread Starter

Sir Kit

Joined Feb 29, 2012
188
Thank you for providing that additional information.

Going by what few notes I have, the tank circuit was an 8 turn 2" diameter air core coil in parallel with a half plate variable capacitor, The old fashioned radio style with a typical tuning range of 10-100pF. I assume the tank circuit, in parallel with gates and ground, was intended to function as a detector of ambient signals over a certain frequency range.

Please advise if this changes your assessment of the circuit behavior.
 
Well, as you describe the additional notes, I can agree with your assumption, this circuit could serve as a kind of selective tunable RF 'sniffer' in the some MHz range. The input signal to the gates was the resonant voltage developed across the LC tank circuit from the ambient and at the output there was some means to detect a change in amplitude. Like in an Absorption wave or frequency meter. And the negative feedback amplifier may have been needed to improve linearity.
What is not 'straightforward' in this assumption is the text 'Test Signal (not RF)' written to the gates input in your original schematic. The LC tank with the 10-100 pF and 8 turn 2" diameter air coil must have been tunable certainly in the RF range.
 

Thread Starter

Sir Kit

Joined Feb 29, 2012
188
Yes, I see you are right about the "not RF' being wrong. As I mentioned, this was an old circuit I found. That notation may have had some other meaning.

In any case, it appears we now agree that the primary purpose of the circuit was not to oscillate, and it would not do so functioning as a tunable amplifier with unity gain.

Thanks again for your helpful input.
 
Top