working of opampp in L-C meter

[ymmen]

can anyone please explain me the operation of comparator LM311 in the ciruit below?

 

[FvM]

It's working as an oscillator (schmitt-trigger with positive feedback). It oscillates a the resonance frequency of the connected LC circuit.

 

[ymmen]

It's working as an oscillator (schmitt-trigger with positive feedback). It oscillates a the resonance frequency of the connected LC circuit.

then can you plz explain how the comparator is setting up the two threshold levels and how to calculate those values? In circuits what i have learnt i never used a resonant circuit in non inverting terminal of opamp.

 

[FvM]

In circuits what i have learnt i never used a resonant circuit in non inverting terminal of opamp.

Yes, the circuits is somehow unusual.

I presume you can calculate the DC thresholds yourself by applying literature formulas. See e.g. https://en.wikipedia.org/wiki/Schmitt_trigger#Non-inverting_Schmitt_trigger

The actual hysteresis is set by the LC circuit impedance. The oscillation magnitude at higher frequencies also depends on the comparator output delay and won't be easily to calculate.

 

[ymmen]

Yes, the circuits is somehow unusual.

I presume you can calculate the DC thresholds yourself by applying literature formulas. See e.g. https://en.wikipedia.org/wiki/Schmitt_trigger#Non-inverting_Schmitt_trigger

The actual hysteresis is set by the LC circuit impedance. The oscillation magnitude at higher frequencies also depends on the comparator output delay and won't be easily to calculate.

can you explain what do you mean by "The hysteresis is set by LC circuit". LC circuit is responsible for sinusoidal waves, how can it set threshold levels?

 

[FvM]

The circuit has positive feedback, the feedback factor which defines the hysteresis is set by a voltage divider, part of it is the LC circuit. So the feedback factor depends on the LC impedance.

 

[ymmen]

The circuit has positive feedback, the feedback factor which defines the hysteresis is set by a voltage divider, part of it is the LC circuit. So the feedback factor depends on the LC impedance.

can you guide me from where should i start in order to comprehend this circuit, i just know how the relaxation oscillator works.

 

[D.A.(Tony)Stewart]

The positive feedback makes it oscillate, but the negative feedback keeps it self biased in the middle to make relaxation oscillation possible.

I see that the RC time constant of the negative path to pin 3 will dominate the frequency with ripple equal to the hysteresis defined by the positive feedback defined by thresholds set by the 5V out and R ratios.

I do not see how it works. The "precision" L and C' parts installed do not seem to influence the relaxation frequency controlled by the 47k+10uF.

 

[FvM]

It's not actually a relaxation oscillator, rather a LC oscillator with excessive gain. Whatever you call ist, you can see that it's oscillating at the LC resonance frequency as long as the LC circuit has sufficient Q.

It's no classical text book oscillator circuit and it's operation is depending on unspecified device parameters like the frequency characteristic of the LM311 comparator.

 

[D.A.(Tony)Stewart]

It's not actually a relaxation oscillator, rather a LC oscillator with excessive gain. Whatever you call ist, you can see that it's oscillating at the LC resonance frequency as long as the LC circuit has sufficient Q.

It's no classical text book oscillator circuit and it's operation is depending on unspecified device parameters like the frequency characteristic of the LM311 comparator.

The only explanation can be is that there is an error in the schematic with pins 2&3 reversed. Then the high gain negative feedback supports your argument and built in LC parts should resonate at 555kHz with 1000pF , 82uH.

Otherwise as shown when the output saturates, there is no gain except the slope of the relaxation effect on RC on -ve feedback, then max gain when the input levels cross and linear mode resumes for the output transition, then the gain is zero again.

With the schematic corrected the Low freq slew rate is much slower on the 10uF*47k that 1nF*50k so negative feedback prevails in regulating the bias point as a linear oscillator, self biased for 50% symmetry

However I still wonder how sensitivity is achieved well below 82uH when in series, L does not change and C=<1pF //1000pF ... Same issue.

Even if the PIC counter measured for 10 seconds the stability is'nt good enough to detect and resolve 0.1pF in parallel with 1000pF with good accuracy.

note : unlike an Op Amp, with +,-,out on pins 3,2,6 this chip is 2,3,7, which may explain the error on this design.

 

[FvM]

It's intentionally positive feedback. And it works as an oscillator at the LC resonance, as you can check in a circuit simulation.

 

[ymmen]

It's intentionally positive feedback. And it works as an oscillator at the LC resonance, as you can check in a circuit simulation.

yeah you are right, in simulation the output frequency is according to frequency of LC tank circuit. But how will i analyse this circuit, its not making any sense.

 

[FvM]

It's a strongly nonlinear circuit and can't be described with usual linear feedback circuit theory, although I presume that the Nyquist criterion manifests instable behaviour. But stating instability doesn't tell the actual oscillation frequency.

It has been already mentioned that the circuit can be analyzed as a relaxation oscillator if the LC circuit is disconnected.

SunnySkyguy stated correctly that the circuit has no feedback gain when the comparator output is in saturation. What however happens is that the output transition which is caused by the relaxation configuration after some time stimulates an oscillation of the LC circuit. With sufficient Q, the comparator immediately switches back after a half period and a continuous wave starts.

So I think it can be intuitively understood why the circuit is working. Stability and C sensitivity is a different thing. I didn't build this circuit and can't tell anything about the actual performance. The 0.1 pF resolution doubted by SunnySkyguy may be achieved in frequency measurement (corresponding to demanding 0.01 ppm frequency resolution!) with multi-period measurement, but I don't know if it has any significance.

 

[D.A.(Tony)Stewart]

It oscillates, at 555kHz ok , i get it.


Even assuming non-linear correction is applied it must have a routine to calibrate before use with open/short cct. to compensate for ppm drift in Xtal and 2% error onboard LC parts.