Amplitude noise in LC oscillators

[Teleno]

the OP is trying to detect the presence of metallic objects by having the object act like a shorted turn and changing the voltage and frequency of the oscillator.

one problem with this question, is the OP has not mentioned what the detection distance is that he's trying to achieve, or the dimensions of L1, so it's impossible to make any solid recommendations, even for what possible noise sources to eliminate.

L1 is 20cm in diameter. The detection distance with this flicker noise and the filtering algorithm I mentioned is 10 cm. for a 0.3 g. gold nugget, which is not at all bad. A 1 EUR coin is detectable at 20 cm. and a can of RedBull at 100 cm. However, due to this flicker noise it's not possible to plot the amplitude vs. frequency changes in a meaningful way in order to achieve discrimination. A solid baseline would be required to achieve that.

 

[BobW]

All this noise coming off of an LC resonant circuit suggests that the Q of your circuit is far too low. If the Q is high, it should be fairly immune to noise. The reason for low Q could be the type of coil construction, or overly tight coupling of the LC tank into the feedback path. Most likely a combination of both.

The loading effect of nearby metal is in essence a Q-reducing effect. If you have low Q to start with, then the effect of nearby metal won't be as great as if the Q was high to start with.

 

[Teleno]

All this noise coming off of an LC resonant circuit suggests that the Q of your circuit is far too low. If the Q is high, it should be fairly immune to noise.

There are two kinds of noise in oscillators: phase noise and amplitude noise. A high Q implies a high voltage gain at the center frequency (more amplification of voltage noise) and a a quick gain drop off the center frequency (less phase noise). In fact, I believe lowering the Q would lower the amplitude noise and increase the phase noise, but since this also implies a lower gain the overall effect on sensitivity would be the same. There is no trade-off.

From everything I've read on this subject the answer seems to be a pulse-bias oscillator such as the Colpitts where the active (noisy) component is on only for a fraction of the signal's period. Thereby integration of the noise by the LC tank takes place for a shorter time, giving a lower noise figure.

But the Colpitts is not efficient. I'm not aware of any pulse-bias oscillator configuration involving two active elements that would allow me to put enough current int he tank (about 250 mA) with good efficiency (for battery operation).

 

[Beau Schwabe]

one problem with this question, is the OP has not mentioned what the detection distance is that he's trying to achieve, or the dimensions of L1, so it's impossible to make any solid recommendations, even for what possible noise sources to eliminate.

Not necessarily .... He knows the inductance of the coil and the parasitic capacitance (assuming without metal present) ... As more metal is introduced the resonant frequency between the parasitic capacitance and the coil will go up, so the +577kHz can be filtered with a low-pass filter. The 81kHz (resonant frequency between the 380uH and the 10nF Capacitor) will be allowed to pass, and will also increase with the additional presence of metal Typically for an air core coil, the inductance will range from 100% (no metal) to roughly 150% (metal saturation) ... So roughly 380uH to 250uH ... or 81kHz to 100kHz

 

[BobW]

Typically for an air core coil, the inductance will range from 100% (no metal) to roughly 150% (metal saturation) ...

Non-ferrous metal will cause the inductance to decrease. Ferrous metal will cause the inductance to increase.

 

[BobW]

From everything I've read on this subject the answer seems to be a pulse-bias oscillator such as the Colpitts where the active (noisy) component is on only for a fraction of the signal's period. Thereby integration of the noise by the LC tank takes place for a shorter time, giving a lower noise figure.

Could you provide some references? I've never heard of a Colpitts oscillator referred to as "pulse bias."

Edit: The reason why I asked for references is that your terminology seems strange, and I'm thinking that you're looking at something that wasn't originally written in English, and that the translation of technical terms may be inaccurate or at least not conventional. I'm thinking that "pulse bias" may mean Class C.

 

[Teleno]

Could you provide some references? I've never heard of a Colpitts oscillator referred to as "pulse bias."

Edit: The reason why I asked for references is that your terminology seems strange, and I'm thinking that you're looking at something that wasn't originally written in English, and that the translation of technical terms may be inaccurate or at least not conventional. I'm thinking that "pulse bias" may mean Class C.

These are some examples: https://ieeexplore.ieee.org/document/7571191/?denied https://www.researchgate.net/figure...LC-tank-pulsed-bias-oscillator_fig1_216498544

 

[BobW]

These are some examples: https://ieeexplore.ieee.org/document/7571191/?denied https://www.researchgate.net/figure...LC-tank-pulsed-bias-oscillator_fig1_216498544

Just as I thought.

 

[Teleno]

Just as I thought.

The on-time of the active components in pulsed bias is much shorter than class C.