RL circuit practice

Tony Stewart · Dec 31, 2015

Series C values are as I stated correctly, in other words 1/C = 1/C1 + 1/C2

Also the OP & I were the ones who pointed out using high R reduces error, while you insisted on using an R lower than DCR

Meanwhile as a Test Engineer for 35 yrs, I know of no instrument using the RL method and LC method with resonance or oscillation is preferred.
I also have 15yrs of R&D experience in analog systems alone

Ratchit · Dec 31, 2015

Tony Stewart said:
Series C values are as I stated correctly, in other words 1/C = 1/C1 + 1/C2

Yes, the formula for series capacitance was never in doubt. I just wanted to confirm whether you meant series or parallel in the circuit. The capacitor placement in the circuit is ambiguous as to whether they are in series or parallel.

Also the OP & I were the ones who pointed out using high R reduces error, while you insisted on using an R lower than DCR

No you all didn't. I talked about using a 10K resistor to swamp out the coil resistor as early as post #7. Unfortunately, the OP's generator was not able to generate the frequency needed. I suggested a lower resistance and lower frequency to eliminate the high frequency effects.

Meanwhile as a Test Engineer for 35 yrs,

Good, then you will understand what I am talking about.

I know of no instrument using the RL method and LC method with resonance or oscillation is preferred.

I never said the TC method was the best way to measure inductance. I only suggested it as a simpler alternative.

I also have 15yrs of R&D experience in analog systems alone

And I respect that. Do you now agree that L = 1/(w^2)*C) for series resonance? You seemed to disagree about that in a previous post.

Ratch

Tony Stewart · Dec 31, 2015

Even if ground is between two caps in series, it is not ambiguous to the experienced. Why did you doubt me?

Ratchit · Dec 31, 2015

Tony Stewart said:
Even if ground is between two caps in series, it is not ambiguous to the experienced. Why did you doubt me?

Because the op-amp output goes to the right-side capacitor and simultaneously to the left side capacitor through a resistor and coil. That is the parallel perspective. The series perspective is around the loop. I wanted to know which way you were calculating the capacitors.

You still have not answered the question of what is the correct formula for series resonance.

Ratch

Tony Stewart · Dec 31, 2015

I see I need to correct your false assumptions.

1) it is not an Op Amp, rather a logic gate
2) it is not a series resonant circuit.
It is a parallel resonant circuit with series caps.

Thanks for pointing out my typo
ω²=1 / (LC)
L = 1 / (ω²C)
L = 1 / {(2πf)²C}

Ratchit · Dec 31, 2015

Tony Stewart said:
I see I need to correct your false assumptions.

1) it is not an Op Amp, rather a logic gate

A overdriven inverting Op Amp could be made to do the same thing functionally.

2) it is not a series resonant circuit.
It is a parallel resonant circuit with series caps.

Now that I examined it better, I don't think it is either a series or parallel resonant circuit. It looks more like an electronic multi-vibrator to me, which is not a linear oscillator.

In some equipment, like power supplies, there are many capacitors connected to ground. You would not say those capacitors are in series, would you?

Thanks for pointing out my typo
ω²=1 / (LC)
L = 1 / (ω²C)
L = 1 / {(2πf)²C}

Glad to do it.

Ratch

Tony Stewart · Dec 31, 2015

Ratchit said:
A overdriven inverting Op Amp could be made to do the same thing functionally.

Now that I examined it better, I don't think it is either a series or parallel resonant circuit. It looks more like an electronic multi-vibrator to me, which is not a linear oscillator.

In some equipment, like power supplies, there are many capacitors connected to ground. You would not say those capacitors are in series, would you?

Glad to do it.

Ratch

1. OP Amp ? Not , never
as Op Amp as I had shown at that f.
Op Amps are integrators (1/f response) and not broadband linear CMOS amplifiers with high gain (10 per stage of 3) at 10Mhz
- thus useless for high f oscillators for measuring LC resonance

2. electronic multi-vibrator ? Not , never
It is not a relaxation oscillator or "multi-vibrator" or "astable".
It is a linear resonant circuit amplifier up the point until output saturates into a square wave since the gain at resonance is >>1 with Barkhousen criteria satisfied for oscillation. So there is a sine wave output feedback after the bandpass parallel filter with 180 phase shift and the logic inverter is 180 phase shift resulting in 360 deg loop phase at resonance. The inductor self biases input DC for 50% duty cycle. this is a "101" oscillator circuit design

3. parallel caps on supply? series? no not never
this is an obvious error in your implied critical comments. They are shorted at each end,
This design has an inductor between each cap. ( which has equal magnitude of impedance at resonance)
Allow me to redraw for you to understand.

https://www.falstad.com/circuit/circuitjs.html?cct=$+16+5.000000000000001e-11+10.20027730826997+50+5+43 r+160+512+288+512+0+1000 w+288+512+288+544+0 c+352+544+288+544+0+1.5e-9+-0.4342750419355297 l+288+512+400+512+0+0.000001+-0.07080797846745183 c+400+544+352+544+0+1.5e-9+4.571412541865394 w+400+512+400+544+0 I+400+592+160+592+0+0.5 g+352+544+352+560+0 w+400+544+400+592+0 w+160+512+160+592+0 o+9+64+0+298+5+0.0015625+0+-1 o+8+64+0+290+5+0.00009765625+1+-1

Tony Stewart · Dec 31, 2015

moreover if you analyze the above link and press RESET, you can see the startup performance for DC self-bias, Step Input broad-band noise, resonant output which is amplified in a linear fashion when Vcc/2 or 2.5V is reached and then operates in pure linear mode until saturation of output.

This is identical with all Pierce Crystal Oscillators
https://en.wikipedia.org/wiki/Pierce_oscillator

capiche?

The linear aspects of a logic gate are easily misunderstood. Gain is fixed below to unity gain by 1:1 R ratio added to above oscillator to make a "sine wave buffer" ,

In LC resonator the feedback gain it is affected by Q ratio. = R/X(f)

https://goo.gl/8V1qJd

Ratchit · Dec 31, 2015

Tony Stewart said:
1. OP Amp ? Not , never
as Op Amp as I had shown at that f.
Op Amps are integrators (1/f response) and not broadband linear CMOS amplifiers with high gain (10 per stage of 3) at 10Mhz
- thus useless for high f oscillators for measuring LC resonance

Functionally speaking, it is a overdriven inverting amplifier, even if an Op-Amp is not capable of reaching those frequencies.

2. electronic multi-vibrator ? Not , never
It is not a relaxation oscillator or "multi-vibrator" or "astable".
It is a linear resonant circuit amplifier up the point until output saturates into a square wave since the gain at resonance is >>1 with Barkhousen criteria satisfied for oscillation. So there is a sine wave output feedback after the bandpass parallel filter with 180 phase shift and the logic inverter is 180 phase shift resulting in 360 deg loop phase at resonance. The inductor self biases input DC for 50% duty cycle. this is a "101" oscillator circuit design

By any other name, it looks like a pulsed Colpitts oscillator. The output, when using a gate, looks like a rounded triangular wave, not a nice sine wave.

3. parallel caps on supply? series? no not never
this is an obvious error in your implied critical comments. They are shorted at each end,
This design has an inductor between each cap. ( which has equal magnitude of impedance at resonance)
Allow me to redraw for you to understand.

https://www.falstad.com/circuit/circuitjs.html?cct=$+16+5.000000000000001e-11+10.20027730826997+50+5+43 r+160+512+288+512+0+1000 w+288+512+288+544+0 c+352+544+288+544+0+1.5e-9+-0.4342750419355297 l+288+512+400+512+0+0.000001+-0.07080797846745183 c+400+544+352+544+0+1.5e-9+4.571412541865394 w+400+512+400+544+0 I+400+592+160+592+0+0.5 g+352+544+352+560+0 w+400+544+400+592+0 w+160+512+160+592+0 o+9+64+0+298+5+0.0015625+0+-1 o+8+64+0+290+5+0.00009765625+1+-1

Some power supplies have inductors between each filter cap also.

Ratch

Tony Stewart · Dec 31, 2015

Ratch , your comments are incorrect or irrelevant.

1. Functionally speaking, it is a overdriven inverting amplifier,
This would imply external input.
This is a resonant oscillator not a multi-vibrator

I stand by my comments, No Not never.

Your arguments deflect your errors and conflict with general wisdom of experienced designers.
Fallacious at best.

Tony Stewart · Dec 31, 2015

Again you make fallacious irrelevant comparisons to deflect admitting your errors.

The Colpits Osc. is not what I demonstrated . That design uses non-inverting amplifier.
i.e. 180 phase inverted amplifier is used for Pierce Osc.
When saturated output with high gain, it is also the common uC PIC square wave clock with external reactive element such as LC, crystal or tuning fork.

a Colpitts or a Multi-vibrator
No, not! , never

Ratchit · Dec 31, 2015

Tony Stewart said:
Ratch , your comments are incorrect or irrelevant.

In what way? Just saying so does not make it so.

1. Functionally speaking, it is a overdriven inverting amplifier

Click to expand...

This would imply external input.
This is a resonant oscillator not a multi-vibrator

External input? No more so than the gate does.

Yes, there is resonance involved, like there is for many nonlinear oscillators. You even said that there is saturation involved. So if you don't want to call it a multi-vibrator, then perhaps a switching oscillator would be a better descriptive.

I stand by my comments, No Not never.

I think you mean to say, "No, not ever".

Your arguments deflect your errors and conflict with general wisdom of experienced designers.
Fallacious at best.

You can say anything, but you should describe and prove what you say.

Ratch

Ratchit · Dec 31, 2015

Tony Stewart said:
Again you make fallacious irrelevant comparisons to deflect admitting your errors.

Again you say things but don't back them up.

The Colpits Osc. is not what I demonstrated . That design uses non-inverting amplifier.
i.e. 180 phase inverted amplifier is used for Pierce Osc.
When saturated output with high gain, it is also the common uC PIC square wave clock with external reactive element such as LC, crystal or tuning fork.

Your submission still uses a inductor with a cap at each end. Instead of connecting the caps together and then to ground, connect them to ground directly. Now it looks like the pi filter of cap--inductor--cap, like so many power supplies used to do. You drive your circuit at a different point than the Colpitts does, but it still uses the same basic circuit.

a Colpitts or a Multi-vibrator
No, not! , never

It is matter of perspective.

Ratch

Ratchit · Dec 31, 2015

Tony Stewart said:
moreover if you analyze the above link and press RESET, you can see the startup performance for DC self-bias, Step Input broad-band noise, resonant output which is amplified in a linear fashion when Vcc/2 or 2.5V is reached and then operates in pure linear mode until saturation of output.

This is identical with all Pierce Crystal Oscillators
https://en.wikipedia.org/wiki/Pierce_oscillator

capiche?

How does all this relate to the capacitors being in series when the topology shows them to be connected in parallel.

The linear aspects of a logic gate are easily misunderstood. Gain is fixed below to unity gain by 1:1 R ratio added to above oscillator to make a "sine wave buffer" ,

In LC resonator the feedback gain it is affected by Q ratio. = R/X(f)

How come the output does not look like a nice sine wave. Instead it looks like a rounded triangle wave.

https://goo.gl/8V1qJd

What does 5 volts on the 1.5 nf capacitor do for anything?

Ratch

Tony Stewart · Dec 31, 2015

Dear Hopeless;

-granted there are similarities, but the Colpitts is in-phase and the Pierce is anti-phase.
-big difference

Neither above is called https://en.wikipedia.org/wiki/Multivibrator

There is no rounded triangle in my simulation.
Only a slew rated limited square wave , which can be edited at will.

Ratchit · Dec 31, 2015

Tony Stewart said:
granted there are similarities, but the Colpitts is in-phase and the Pierce is anti-phase.

I will go with that.

Neither is a called https://en.wikipedia.org/wiki/Multivibrator

I never called a Colpitts a multi-vibrator. I believe I changed the name of your circuit to a switching oscillator in a previous post. A multi-vibrator is a type of switching oscillator.

There is no rounded triangle in my simulation.
Only a slew rated limited square wave , which can be edited at will.

It just does not look like a good sine wave. Hopefully, the real circuit does.

Ratch

Tony Stewart · Dec 31, 2015

Ratchit said:
A overdriven inverting Op Amp could be made to do the same thing functionally.

Now that I examined it better, I don't think it is either a series or parallel resonant circuit. It looks more like an electronic multi-vibrator to me, which is not a linear oscillator.

In some equipment, like power supplies, there are many capacitors connected to ground. You would not say those capacitors are in series, would you?

Ratch

You incorrectly identified the classic design as an electronic multi-vibrator

The sine wave is very clean but the Simulation time sampling gives quantization effects which can be reduced by reducing the sample time. in Options > other options but then unlike digital scope this has less memory. For better slew rate , I edited the gate to be faster below.

I changed one cap to pullup and one to ground to speed up the startup DC bias.. Same f results as Vcc to ground ~0 Ohms impedance

Tony Stewart · Dec 31, 2015

https://goo.gl/7mgDQf

jjw · Dec 31, 2015

Ratchit said:
.........

It just does not look like a good sine wave. Hopefully, the real circuit does.

Ratch

I simulated this in LTSpice, L=10ouH C1=C2=10nF using 74HCU04 as inverter, frequency ~ 225kHz and the second harmonic measured from input pin was about 0.3% and from output pin about 3%

Edit: corrected the value of L

Tony Stewart · Dec 31, 2015

Not bad, thanks for the confirmation jjw.

I think Q affects harmonic rejection and THD with the values selected.
As stated the design was a starting place for measure L.
For improved THD, Q can be increased and linear gain adjusted.

There are subtle differences with U04 types and Buffered types with spurious potential with overtones, but 3 stages of gain can extend range of operation for low Q inductor tests. U type is recommended for crystals but often Buffered (B) types work well too.

RL circuit practice

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Well-Known Member

Active Member

Well-Known Member

Similar threads

Privacy & Transparency

Privacy & Transparency