Needed: single transistor audio amplifier with positive feedback

hello,
I am trying to amplify as much as possible the very few microwatts out of a radio detector.
I would like to use just one transistor if possible so I believe the best to do is to apply some positive feedback from the output to the input to further amplify the signal.

Is there any such schematic out there using a single transistor and positive feedback loop??
 
The OP was asking about positive feedback for amplifying the output from a radio detector. The regen uses positive feedback for amplifying the input to the detector. The superregen does this even more, and includes the detector too. There is also the reflex, which uses one device to first amplify the RF then amplify the audio on a second pass.

Could the OP give us a bit more detail on what he is trying to do? I would have thought an op-amp would give more gain than a transistor, and less risk of instability.
 
The OP was asking about positive feedback for amplifying the output from a radio detector. The regen uses positive feedback for amplifying the input to the detector. The superregen does this even more, and includes the detector too. There is also the reflex, which uses one device to first amplify the RF then amplify the audio on a second pass.

Could the OP give us a bit more detail on what he is trying to do? I would have thought an op-amp would give more gain than a transistor, and less risk of instability.

That is exactly df96! The regen/superregen/reflex uses feedback to amplify the radio or the radio and audio together. I do not want radio feedback. I leave the radio part to another circuit. I just need an AUDIO amplifier that uses positive feedback for ultra high gain, which I will put after the detector. For simplicity reasons (since my design is minimal) I would like it to use just a single transistor.

I have found many regens but this is just not what I am trying to do. It must not be that hard to feed some positive output signal back to the input of a single transistor, but I just do not know how to do it.
I think this would be of interest to many radioamateurs building minimal designs.
 

Elvee

Member
2006-09-08 2:04 pm
There are a number of fundamental impossibilities there: you cannot use a reflex stage to amplify twice the same frequency range, and you cannot use positive feedback on a single active stage without a "prop", such as a transformer for phase inversion.
But even if you allow a transformer, it would be silly, because it would be difficult to ensure the stability of such a stage, and a supplementary transistor is much less expensive than any transformer.

60 years ago, it might have been worth exploring the possibility, but not anymore.
 

jcx

Member
2003-02-17 7:38 pm
..
another view is that positive feedback just "unmasks" gain that is inherent in the device - in triodes there is internal negative feedback since the electrons "see" and respond to the plate V - it is possible (and was necessary in early tube circuits) to use positive feedback to partially cancel the triode's internal negative feedback over a restricted frequency range to obtain higher useful gain from the tubes

bjt transistors have much lower internal feedback and don't benefit much from positive feedback - although multistage amps may use positive feedback to make loop gain available from a stage otherwise using 100% local feedback
 
Last edited:
You can get some positive feedback with a single transistor and without a transformer by having some feedback from collector to emitter. Put a small resistor in the emitter circuit - by itself this will give negative feedback. Then add a resistor plus capacitor from the collector to the emitter, to give positive feedback. Personally, I would use an opamp.

Positive feedback does not "unmask gain", it actually adds gain. You can see this from any oscillator, where the power gain approaches infinity - it amplifies thermal noise right up to useful levels. The device did not have hidden gain this big!
 
There are a number of fundamental impossibilities there: you cannot use a reflex stage to amplify twice the same frequency range, and you cannot use positive feedback on a single active stage without a "prop", such as a transformer for phase inversion.
But even if you allow a transformer, it would be silly, because it would be difficult to ensure the stability of such a stage, and a supplementary transistor is much less expensive than any transformer.

60 years ago, it might have been worth exploring the possibility, but not anymore.

why do you need another stage or transformer? A transistor amplifier outputs two signals. One from it's collector, which is 180degrees out of phase from the input signal and one from it's emitter, which is in phase with the input signal. Couldn't I get a portion of the signal using a capacitor/resistor combination from the emitter and fed it back to the input, thus positive feedback?
 
You can get some positive feedback with a single transistor and without a transformer by having some feedback from collector to emitter. Put a small resistor in the emitter circuit - by itself this will give negative feedback. Then add a resistor plus capacitor from the collector to the emitter, to give positive feedback. Personally, I would use an opamp.

Positive feedback does not "unmask gain", it actually adds gain. You can see this from any oscillator, where the power gain approaches infinity - it amplifies thermal noise right up to useful levels. The device did not have hidden gain this big!

I have seen this microphone amplifier. It uses two stages but It feeds some signal back to the input using a resistor. Is that similar or is it negative feedback?
 

Attachments

  • ecm.gif
    ecm.gif
    21.1 KB · Views: 205

wakibaki

Banned
2008-01-08 11:51 pm
The OP was asking about positive feedback for amplifying the output from a radio detector.

Yeah, do you know of such a device?

For simplicity reasons (since my design is minimal) I would like it to use just a single transistor.

Exactly. The super-regen is minimal.

It must not be that hard to feed some positive output signal back to the input of a single transistor

No, it's not. It's called an oscillator.

But even if you allow a transformer, it would be silly, because it would be difficult to ensure the stability of such a stage, and a supplementary transistor is much less expensive than any transformer.

How are you going to prevent a single frequency growing in amplitude and occupying all the transistor gain?

You might say, make it supersonic and quench it. How would it then be fundamentally different from a super-regen?

w
 
Last edited:
You can have positive feedback to increase gain without necessarily getting oscillation. That is how a correctly adjusted regen works. In principle you could do the same at audio, but there are better ways of achieving high gain. I'm not sure you could do an audio superregen as it might have quite a narrow bandwidth - it is a quenched oscillator so not really a broadband device.
 

Elvee

Member
2006-09-08 2:04 pm
You can get some positive feedback with a single transistor and without a transformer by having some feedback from collector to emitter. Put a small resistor in the emitter circuit - by itself this will give negative feedback. Then add a resistor plus capacitor from the collector to the emitter, to give positive feedback. Personally, I would use an opamp.
This cannot work with a 3 terminal amplifying device: the current you add to a node is taken elsewhere, and at best, you get a status quo, or like here, a loss: see example 2.

How are you going to prevent a single frequency growing in amplitude and occupying all the transistor gain?
You just have to remain under the critical gain: if G is the amplifier's gain, you have to reinject less than a 1/G fraction of the output to the input.
The closer you are, the higher the gain, with Gtot=infinity if exactly 1/G is reinjected.
In the example, 1 (green) is a plain vanilla amp for reference, 2 (magenta) is DF's method and 3 (red) is the transformer method, implemented in a shunt-series combination.
In this case, the improvement is modest: the original gain is multiplied by ~3, but it is already sufficient to reveal the non-linearities.
 

Attachments

  • Comparison.jpg
    Comparison.jpg
    140.2 KB · Views: 191

poynton

Member
2005-03-10 11:57 pm
UK
The question you still have not answered "why a single transistor?"

Do you have voltage / space concerns ?

Way back when... pre-opamp days.... if I wanted more gain, I just added another single transistor gain stage, like (1) above. That usually gave enough gain, if not I added another.
Of course, the noise went up.!

Now, as suggested, I would use an opamp.



.
 

wakibaki

Banned
2008-01-08 11:51 pm
You can have positive feedback to increase gain without necessarily getting oscillation. That is how a correctly adjusted regen works. In principle you could do the same at audio, but there are better ways of achieving high gain. I'm not sure you could do an audio superregen as it might have quite a narrow bandwidth - it is a quenched oscillator so not really a broadband device.

And a regen isn't a narrowband device? I very much doubt that you can operate a broadband (audio) regen and get reliable gain improvements without it sliding into oscillation. I could be wrong about this, I look forward to somebody producing a working example.

I didn't say you could build an audio superregen, I said that if you could, there wouldn't be a useful distinction to be drawn between it and an RF superregen. Anyway the argument that it is a quenched oscillator, so not really a broadband device, doesn't hold water, what is an RF superregen if not a quenched oscillator? This doesn't prevent it producing audio in the form of PWM.

w
 

Ian Finch

Member
Paid Member
2010-04-11 4:22 am
Coffs Harbour, NSW
There have been lots of TRF radios based on the ancient ZN414 or later MK484 chips and these were often boosted with a single BC549 to drive an earpiece. This implies a very low output still. even though the devices will have higher output and lower impedance than a simple reflex detector.

I'm not trying to talk you into another project, neazoi, unless you like the very simple TO92 case hiding something sophisticated, but the limits are real for sure. Increased gain above a standard CE stage is sure possible but the only way up for gain is down for quality. Perhaps you could tell us what you wish to drive and if you can measure the output of what you have with any certainty.
 

wakibaki

Banned
2008-01-08 11:51 pm
Thanks Elvee, I didn't see your post before I posted.

A gain of 3 is indeed a gain, but whether it is a worthwhile improvement in gain is another matter, and demonstrating it in a simulation is not the same as making it work as a practical matter. At what point do you consider the increase in gain would become impractical either as a result of increased nonlinearity or instability? What is the usable bandwith?

w