Author Topic: Noise figure of RF stage with reaction?  (Read 1260 times)

0 Members and 1 Guest are viewing this topic.

Offline CirclotronTopic starter

  • Super Contributor
  • ***
  • Posts: 3168
  • Country: au
Noise figure of RF stage with reaction?
« on: September 26, 2022, 12:44:04 pm »
From what I understand, the noise figure of an AM receiver is mostly determined by the first RF stage (if it has one). If this RF stage is tuned then image rejection is improved as well. What though if we apply a little bit of positive feedback aka reaction to this stage? Certainly the selectivity and sensitivity will be improved, but will the noise figure of this stage also be raised by the same amount as the sensitivity, so no real improvement overall?
 

Offline TimFox

  • Super Contributor
  • ***
  • Posts: 7934
  • Country: us
  • Retired, now restoring antique test equipment
Re: Noise figure of RF stage with reaction?
« Reply #1 on: September 26, 2022, 02:28:00 pm »
General answer:  feedback itself (as opposed to the extra components used to effect the feedback) does not affect the noise figure.
However, the optimal source impedance for best noise figure remains the same, even while the feedback changes the input impedance.
Positive feedback was commonly used over fifty years ago in "regenerative" receivers.
https://www.electronics-notes.com/articles/radio/radio-receivers/regenerative-receiver.php
« Last Edit: September 26, 2022, 02:43:04 pm by TimFox »
 

Offline mag_therm

  • Frequent Contributor
  • **
  • Posts: 633
  • Country: us
Re: Noise figure of RF stage with reaction?
« Reply #2 on: September 26, 2022, 03:45:51 pm »
Armstrong's first regenerative receiver.
As the positive feedback (at 180 degrees at Wr) is increased, the amplifier will start to be an incipient oscillator.
I expect that the phase noise characteristics of ordinary oscillators (hartley etc) will then progressively apply along with the thermal noise at input of the amplifier.
However atmospheric noise from antenna on conventional AM frequencies will dominate.

Here is a tutorial about phase noise in oscillators and reciprocal mixing.
https://www.eecis.udel.edu/~vsaxena/courses/ece518/Handouts/Oscillator%20Phase%20Noise.pdf

The Q multiplier used on tube HF receivers was a similar application, however they were applied at the intermediate frequency.

The regen first stage and Q multiplier act to reduce the R component in the RLC tuned circuits to increase selectivity.
On Lafayette HE30 I used the Q multiplier setting of barely oscillating to resolve ssb.
That gave good fidelity however I no long use it on the digital modes because the Q mult is not stable enough. Replaced by xtal IF oscillators , Gilbert cell and ceramic filters.

An interesting experiment in analog domain  would be a regen receiver, or Q multiplier locked to DDS.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16547
  • Country: us
  • DavidH
Re: Noise figure of RF stage with reaction?
« Reply #3 on: September 26, 2022, 11:08:41 pm »
As a practical matter, the AM broadcast band, and shortwave bands, have such high atmospheric noise that input noise figure is irrelevant.  The only reason an RF amplifier is included is to block leakage from the local oscillator.
 

Offline CirclotronTopic starter

  • Super Contributor
  • ***
  • Posts: 3168
  • Country: au
Re: Noise figure of RF stage with reaction?
« Reply #4 on: September 27, 2022, 12:16:42 am »
I just asked the question only as a matter of interest. Did quite a bit of playing around with home made vacuum tube regenerative sets in 72-73. Yikes! That's fifty years ago.  :o The air was so much quieter back then too. It's really surprising how well you can get a simple regen set to work. There is a special kind of satisfaction in squeezing the last ounce out of a bare bones radio.  :)
 

Online vk6zgo

  • Super Contributor
  • ***
  • Posts: 7563
  • Country: au
Re: Noise figure of RF stage with reaction?
« Reply #5 on: September 27, 2022, 01:26:14 am »
As a practical matter, the AM broadcast band, and shortwave bands, have such high atmospheric noise that input noise figure is irrelevant.  The only reason an RF amplifier is included is to block leakage from the local oscillator.
If it is a tuned RF stage, a more important function is to increase the pre-mixer selectivity to minimise image interference.
 

Offline radiolistener

  • Super Contributor
  • ***
  • Posts: 3282
  • Country: ua
Re: Noise figure of RF stage with reaction?
« Reply #6 on: September 28, 2022, 09:24:38 am »
As a practical matter, the AM broadcast band, and shortwave bands, have such high atmospheric noise that input noise figure is irrelevant.  The only reason an RF amplifier is included is to block leakage from the local oscillator.

It depends on the frequency. AM broadcast stations working up to 26.2 MHz. Below 10-14 MHz atmospheric noise is really high enough to just ignore receiver noise, but above 10-14 MHz usual receiver noise can be high enough to beat atmospheric noise.
 

Offline RoV

  • Regular Contributor
  • *
  • Posts: 175
  • Country: it
Re: Noise figure of RF stage with reaction?
« Reply #7 on: September 28, 2022, 07:53:26 pm »
... but will the noise figure of this stage also be raised by the same amount as the sensitivity, so no real improvement overall?

Ignoring external noise contributions (e.g. just measuring the NF with an analyzer), I think that the NF will improve (means lower NF) with increasing reaction gain, but will never go below the best NF that the transistor can give. The effect can be significant if subsequent stages are noisy, not otherwise. You are increasing the gain of the 1st stage without changing its NF (first approximation...), so the NF of the chain will improve overall, unless the following stages already have a good NF.
In practice the positive feedback loop will include some noisy components (internal and external to the transistor), so the transistor NF will degrade somewhat.

Offline A.Z.

  • Frequent Contributor
  • **
  • Posts: 851
  • Country: it
Re: Noise figure of RF stage with reaction?
« Reply #8 on: September 28, 2022, 09:07:00 pm »
... but will the noise figure of this stage also be raised by the same amount as the sensitivity, so no real improvement overall?

Ignoring external noise contributions (e.g. just measuring the NF with an analyzer), I think that the NF will improve (means lower NF) with increasing reaction gain, but will never go below the best NF that the transistor can give. The effect can be significant if subsequent stages are noisy, not otherwise. You are increasing the gain of the 1st stage without changing its NF (first approximation...), so the NF of the chain will improve overall, unless the following stages already have a good NF.
In practice the positive feedback loop will include some noisy components (internal and external to the transistor), so the transistor NF will degrade somewhat.

right, but then imagine the above being fed by a relatively narrow preselection network and/or a narrow tuned antenna like (e.g.) a tuned loop one
 

Offline mag_therm

  • Frequent Contributor
  • **
  • Posts: 633
  • Country: us
Re: Noise figure of RF stage with reaction?
« Reply #9 on: September 28, 2022, 10:16:32 pm »
For  tunable LC preselector or 1st RF stage :
At 1 MHz   Q = 12  BW = 83 kHz
At 10 MHz Q = 38  BW = 263 kHz
At 50 MHz Q = 75  BW = 666 kHz

Q values are approximate, from the Coilcraft datasheets for the 1812CS series chip inductors.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf