Calculate damping frequency

[Jan Audio]

Hi, with a opamp i make a inverted-amplifier,
then i take a cap (350nf) together with the feedback-resistor (18K) from the opamp output to the inverting-input.
How do i calculate the frequency it cuts off ?

I also make 2 pole bessel/butterworth type filter schematic,
the problem here is i only find websites that generates part values.
I need a website that lets you add part values, and give you the frequency.
Dont that exists ?

thanks

[penfold]

You can simulate your circuit using something like LTSpice if you can't/won't/don't want to do the hand analysis?
If you want some pointers on how to approach the analysis, posting a circuit diagram would help clarify your question

[Benta]

There is no such thing as a "Damping frequency".

There is Damping Ratio and Quality Factor, both related to each other.

That apart, your circuit description is so vague that nothing can be said: "i take a cap (350nf) together with the feedback-resistor (18K) from the opamp output to the inverting-input."
What does this mean? Series? Parallel? How are the opamp inputs really connected?

Here's a helpful link:
https://www.10stripe.com/articles/kb/how-to-ask-a-question.php

[Jan Audio]

That apart, your circuit description is so vague that nothing can be said: "i take a cap (350nf) together with the feedback-resistor (18K) from the opamp output to the inverting-input."
What does this mean? Series? Parallel? How are the opamp inputs really connected?

You know a inverting-amplifier, it has a feedback resistor, it is 18K.
I also place the capacitor of 350nf in parallel with the resistor to get lowpass filtering.

You can simulate your circuit using something like LTSpice if you can't/won't/don't want to do the hand analysis?
If you want some pointers on how to approach the analysis, posting a circuit diagram would help clarify your question

Yes, then i need to fight with it, i also have found some sites that allow dragging modules, to complicated, it dont work for me, thanks

[djerickson]

What you have is a 1 pole low-pass filter. I don't even have to look that one up:-)
Cutoff is when R = Xc, and Xc = 1/(2 * pi * F * C)
The cutoff frequency of a single pole RC filter is  F = 1/(2 * pi * R * C)
R=18K , C=350nF
F = 25.3Hz
The math is the same for a 1 pole high-pass filter.
For single pole LC filters, I leave it as an exercise for the student.
2 pole filters are trickier, I use Spice for them.
Dave

[Jan Audio]

What you have is a 1 pole low-pass filter. I don't even have to look that one up:-)
Cutoff is when R = Xc, and Xc = 1/(2 * pi * F * C)
The cutoff frequency of a single pole RC filter is  F = 1/(2 * pi * R * C)

Ok let me try this now

R=22K , C=350nF

22.000 * 350.000 * 3.1416 * 2 = 48.380.640.000

1 / 48.380.640.000 = 2,066942479471127e-11

F = ehh

I dont have a scope with FFT.

[TimFox]

You have paid no attention to the powers of 10 in your calculation.  350 nF = 350 x 10^-9 F.
If you are using the European convention, “350.000” would mean 350 x 10⁺³.  An error of x10⁺¹² is huge.  If you use the American convention, “350.000” means 350, which is 1000 times less huge.

[Jan Audio]

Oh sorry i posted the Pi value wrong : should be : 3,1416 ( i,m EU )

Google gave me the wrong thing.

[Jan Audio]

I still getting the same 48.380.640.000

[Jan Audio]

By the way : F = 25.3Hz(USA) dont matches reality. ( 25,3Hz for EU )
I would have no sound at that frequency.

25KHz is more what i was thinking.

Oops i have 350pf.

[TimFox]

1/(2 x 3.1416 x 22 x 10⁺³ x 350 x 10^-9)  = 20.7 Hz (American convention on decimal points.)

[Zero999]

When R = 18k and C = 350pF, F_C = 25kHz.

[Benta]

@Zero999:
How did your simulation manage to get phase shift below -90 degrees? I'm curious.

And why does the simulator display "kelvinhertz"?

[Zero999]

@Zero999:
How did your simulation manage to get phase shift below -90 degrees? I'm curious.

It's an inverting amplifier, so the phase shift is really a little over 90 degrees. The op-amp model also has a -11 degree phase shift at 1MHz. Here's a plot without the capacitor.



Running the simulation again, with C = 350nF, gives a 91.4° phase shift at 1kHz, which is really -88.6°.


It's confusing because it's an inverting amplifier. 180° phase shift is equivalent to zero phase shift, for a non-inverting circuit.

And why does the simulator display "kelvinhertz"?

Since Kelvinherz doesn't make sense, it should be obvious what it means. I don't know. Ask Linear Technology, why they chose to use the capital, rather than lowercase K.

[Zero999]

If the voltage source is flipped and the op-amp model upgraded to 100MHz, we can see that the simulator is giving the expected phase shift, going from 0° to 90°.

[Benta]

@Zero999:

"EDIT: What are you talking about? It has a phase shift of 86°, not over 90°!"

Sorry, but your first simulation has a phase starting at 180 degrees (I presume, as it does not start at 0 kelvinhertz) and ends at ~86 degrees.

More than 90 degrees phase shift in a first order filter is impressive.

Always question a simulator. Maths are precise.

[Zero999]

@Zero999:

"EDIT: What are you talking about? It has a phase shit of 86°, not over 90°!"

Sorry, but your first simulation has a phase starting at 180 degrees (I presume, as it does not start at 0 kelvinhertz) and ends at ~86 degrees.

More than 90 degrees phase shift in a first order filter is impressive.

Always question a simulator. Maths are precise.

Sorry, I got a bit confused. I re-edited it out. Everything has phase shift, including the op-amp, so it's not really a first order filter.

[Benta]

Well, if a simulator builder is incapable of using correct SI prefixes and units, and then presents unrealistic results due to hidden parameters, I'd be wary of using it.
I always do a sanity check by hand, but in this case that wasn't even necessary...

Fail! Big time!

[Zero999]

Well, if a simulator builder is incapable of using correct SI prefixes and units, and then presents unrealistic results due to hidden parameters, I'd be wary of using it.
I always do a sanity check by hand, but in this case that wasn't even necessary...

Fail! Big time!

No fail. The simulator is working perfectly. The SI prefix is a minor issue. It's obvious what it means and only bothers extreme pedants, who have nothing better to whinge about. LTSpice is used by Linear Technology to develop their own ICs. It isn't an unrealistic result. One would get a similar result if they actually built the circuit using a real op-amp. There are no hidden parameters. The GBWP of that op-amp model is easily adjusted. The more sophisticated models have more options such as being able to set the phase margin, slew rate, offset voltage etc. I just chose the most basic one.

Yes, it's important to do a sanity check, but one needs to know why the results, either from SPICE, or a real circuit, differ from ideal models assumed by simple calculations.

[penfold]

Well, if a simulator builder is incapable of using correct SI prefixes and units, and then presents unrealistic results due to hidden parameters, I'd be wary of using it.
I always do a sanity check by hand, but in this case that wasn't even necessary...

Fail! Big time!

Its an inverting amplifier... the phase should tend to 180 degrees at LF surely?
As with the phase shift of the opamp, its not hidden, its just the default for that particular spice model,
As for the Kelvin-Hertz, I seem to remember Mike Englehardt saying it was a limitation to one of the software libraries, he wanted to change it along with the default interpretation of "M" being interpreted as milli and having to use Meg to mean mega, but Linear wouldn't let him. Its a shame Linear never really invested in it as a software package, but as a spice engine, it easily beats any commercial offering... at least it tells you when you results are offset by a factor of 1 Kelvin 

[Benta]

The SI prefix is a minor issue. It's obvious what it means and only bothers extreme pedants, who have nothing better to whinge about.

You may be right. I value accuracy in communication highly, especially mathematical and scientific. That may make me seem pedantic.

But your postings of failed (super-simple, first-order) simulations raises questions.

I think I'll stay pedantic. It provides better results.

Over and out from here.

[TimFox]

One legacy of Fortran Spice (perhaps relevant to k vs. K) is that it started with Hollerith cards, all majuscules.  Therefore, “mega” was denoted by “MEG” and “milli” by “M”.  “Kilo” was denoted by “K”.

[Zero999]

Well, if a simulator builder is incapable of using correct SI prefixes and units, and then presents unrealistic results due to hidden parameters, I'd be wary of using it.
I always do a sanity check by hand, but in this case that wasn't even necessary...

Fail! Big time!

Its an inverting amplifier... the phase should tend to 180 degrees at LF surely?

Yes, that confused me too. The phase angles all off by 180°, so if it has a phase shift of 10°, it'll show up as 170°.

But your postings of failed (super-simple, first-order) simulations raises questions.

I've answered the questions. Do you have any more, or are you just trolling now? For the benifit of the orignal poster, I'll go through it once more.

Once an op-amp is added to the equation, it's no longer a super-simple, first-order system. The most blatent red flag in my first post was the plot goes to 1MHz, which is beyond what most common jellybean op-amps will do. I made it closer to an unrealistic, ideal op-amp, by increasing the GBWP to 100MHz. If a total beginner used a 100MHz op-amp, it wouldn't behave, because they'd use a probably use poor layout, such as a breadboard, which would create additonal phase shifts and oscillation.

Go and build it with a real op-amp such as the NE5532 and you'll find the results of the first simulation are fairly accurate. Your basic, rough calculations are a failure, because they don't take the delay of the op-amp into account.

[TimFox]

Yes, the op-amp itself has substantial "excess phase" over its useful frequency range.  It starts at zero at DC, then increases past the first break point (approximately the gain-bandwidth divided by the DC gain) to approximately 90 deg, and increases further at very high frequencies.  The negative feedback in the usual circuit reduces this phase, but does not eliminate it.  Your demonstration with different GBW frequencies shows this.  At a frequency of, say, GBW/10, the op amp gain’s magnitude is a measly 10, and the loop gain has a weaker effect on the excess phase.

[Jan Audio]

1/(2 x 3.1416 x 22 x 10⁺³ x 350 x 10^-9)  = 20.7 Hz (American convention on decimal points.)

Thanks, what does 22 x 10⁺³ x 350 x 10^-9 in normal numbers ?