Low Pass Filter (Resonance Proble)

[ali2002]

Hello I am facing a problem regarding design of passive LPF.I have already found the order of the filter that comes out to be n=4.For this order the C AND L values were calculated and praticaly filter was designed and tested.
During the testing stage it was found that after the implementation of passive filters on the circuit the voltage onthe circuit on which it was implemented boosted up above 25% of previosly observed value.
Ithink so resonace is happening and there is a continous unstable swing of wavefrom at the output from 1 top 25% above the refrence volatge value.Now i want to know why this resonance/ringing is happening and what is the solution for this problem in order to bring stability in design .
Secondly i want to know what should be the range of resonace frequeny whether it should be set within the fps and fsb frequency or if there is any relationship between fps &fsb frequency and resonance frequency.

Waiting for your kind reply

 

[Microwave123]

Did you observe this effect across the spectrum?

Microwave123

 

[LvW]

Hi Ali,

During the testing stage it was found that after the implementation of passive filters on the circuit the voltage onthe circuit on which it was implemented boosted up above 25% of previosly observed value.

I don't understand the real meaning of the above sentence.
What is the "voltage on the circuit" ? Output voltage?
What means "previously observed"? Have you two circuits? Or do you compare with simulation results?
The best thing would be to show us the circuit diagram in connection with the observed signals.

 

[barry]

Without seeing anything, I'm speculating there's an impedance mismatch.

 

[BradtheRad]

Filters with a high order number often have crests and troughs near the corner frequency.

The output can be a few decibels greater than the input. This is possible when inductors are used. And with combined inductors and capacitors it can generate resonant action.

These may or may not be acceptable as a tradeoff for getting the advantage of a steeper rolloff curve.

These characteristics can be generated by mathematical formula, and plotted across the frequency spectrum.

Link to one example showing a typical graph:

https://www.ti.com/lit/an/slyt113/slyt113.pdf

This article goes into more depth. Lots of graphs. (5 MB to load.):

https://www.analog.com/library/analogDialogue/archives/43-09/EDCh 8 filter.pdf

 

[LvW]

Filters with a high order number often have crests and troughs near the corner frequency.
The output can be a few decibels greater than the input. This is possible when inductors are used. And with combined inductors and capacitors it can generate resonant action.
These may or may not be acceptable as a tradeoff for getting the advantage of a steeper rolloff curve.

Just to avoid misunderstandings:
I like to point out that the basic operation principle of passive LC filters (creation of conjugate-complex poles) is based on such "resonant actions" between L and C.
More than that, the steepness of the roll-off is determined by the filter order. The ripple factor (with some amplitude peaking due to higher pole Q values) sharpens the transfer characteristic of the filter at the end of the passband only.

 

[zorro]

Hi,

It is not completely clear to me whether the effect is:
a) there is a peak in frequency response: with sinusoidal input, at some frequency the ouput voltage (not the power, of course) is greater than without the filter inserted
b) there is a peak in time response: with (for example) square wave at the input, at some instants the ouput voltage has peaks (ringing) before stabilize its level
Ali, please can you clarify?
Regards

Z

 

[ali2002]

Yes i observed the effect aross the spectrum .It was quite regular.Previously without filter pk-pk voltage seen on the scope was 25V smooth.After the implementation of filter voltage jumped upto 40 Vpk-pk with filter continously decaying down towards 25V, (1V step by ste) and then jumping upto 40 V after reaching 25V.

Did you observe this effect across the spectrum?

Microwave123

---------- Post added at 08:38 ---------- Previous post was at 08:37 ----------

i observed the effect aross the spectrum .It was quite regular.Previously without filter pk-pk voltage seen on the scope was 25V smooth.After the implementation of filter voltage jumped upto 40 Vpk-pk with filter continously decaying down towards 25V, (1V step by ste) and then jumping upto 40 V after reaching 25V.

---------- Post added at 08:40 ---------- Previous post was at 08:38 ----------

HOW can i avoid the resonace frequency.It is quite damaging .At now my resonance frequency is falling between pass band stop band frequency.

---------- Post added at 08:42 ---------- Previous post was at 08:40 ----------

i observed the effect aross the spectrum .It was quite regular.Previously without filter pk-pk voltage seen on the scope was 25V smooth.After the implementation of filter voltage jumped upto 40 Vpk-pk with filter continously decaying down towards 25V, (1V step by ste) and then jumping upto 40 V after reaching 25V.

---------- Post added at 08:43 ---------- Previous post was at 08:42 ----------

i observed the effect aross the spectrum .It was quite regular.Previously without filter pk-pk voltage seen on the scope was 25V smooth.After the implementation of filter voltage jumped upto 40 Vpk-pk with filter continously decaying down towards 25V, (1V step by ste) and then jumping upto 40 V after reaching 25V.

---------- Post added at 08:46 ---------- Previous post was at 08:43 ----------

Sharpness can only be obtained after the stabilty is attained.There circuit is not stable .Help me out so that i can attain stability

---------- Post added at 08:47 ---------- Previous post was at 08:46 ----------

i observed the effect aross the spectrum .It was quite regular.Previously without filter pk-pk voltage seen on the scope setting it to DC OFFSET .It was 25V smooth.After the implementation of filter voltage jumped upto 40 Vpk-pk with filter continously decaying down towards 25V, (1V step by ste) and then jumping upto 40 V after reaching 25V.

---------- Post added at 08:49 ---------- Previous post was at 08:47 ----------

How this mismatch can be avoided or solved.

---------- Post added at 09:06 ---------- Previous post was at 08:49 ----------

I didnt use op amp only passive elements L & C are being used order n=4

 

[BradtheRad]

Don't know if this applies to your situation...

With order numbers higher than 1, and with no load attached, there are filter topologies where it is possible to draw extreme current from the source at a resonant frequency. Example, if you do NOT attach a speaker when using some crossover designs, you can blow the amplifier at a resonant frequency. Even though it's a passive filter.

The result might be large buildup of emf on the coils with each cycle. This would not necessarily be easy to measure.

And perhaps it is an increasing buildup that creates greater swing with each cycle, until it is as much as 50 percent greater than the applied AC voltage? The question being whether your load is high resistance?

 

[ali2002]

At full load condition i have observed that the stability comes in the circuit i.e jumping stops but the attenuation for which filter is designed has not been attained by this passive filter.This LP filter was designed for 80db Attenuation choosing order N=4 but hardly it has given 20db attenuation . Higher frequency components are also coming in the result seen on the scope.
How this problem can be solved.

 

[BradtheRad]

Glad you got some positive development.

Higher frequency components are also coming in the result seen on the scope.
How this problem can be solved.

The webpage below (allaboutcircuits.com) tells about various types of filters. Two-thirds of the way down it describes "Capacitive Inductive low-pass filter." (Re 2nd-order filter)

Resonant filters : FILTERS

It brings up a problem similar to yours.

"What was supposed to be a low-pass filter turns out to be a band-pass filter with a peak somewhere around..."

A few sentences later:

"The problem is that an L-C filter has a input impedance and an output impedance which must be matched. The voltage source impedance must match the input impedance of the filter, and the filter output impedance must be matched by “rload” for a flat response."

It tries new values and then states:

"Figure below shows the “flat” response of the L-C low pass filter when the source and load impedance match the filter input and output impedances."

Your 4th-order filter has more going on inside it than this 2nd-order filter, but they work on similar concepts.

 

[ali2002]

Hello sir i have checked the Source resistance Rs and Load resistance R L.In my design they both are equal but still the problem is right their where it was i.e. continuous resonance effect is coming and required attenuation which the design was supposed to provide have also not been achieved.

 

[BradtheRad]

the problem is right their where it was i.e. continuous resonance effect is coming

Are there any vertical (sudden) transitions in your signal? If so these can trigger 'ringing'.

The ringing will be at one or more resonant frequencies created by LC interaction in your filter (per post #6).

These resonant frequencies can also get emphasized in response to your incoming signal. But I believe your signal is supposed to damp them. Or your load is supposed to have sufficiently low impedance to damp the resonant frequencies.

It may pay to reduce your filter down to a 3rd order. Etc. See if anything changes. It could indicate which components are contributing to what undesired effect, and where component values need to be changed, etc.

 

[FvM]

Strange discussion about detail problems without clarifying the basic circuit setup. You're likely jumping into conclusions.

- Involved load and source, anything known about their impedance?
- which side is expected to excitate the "oscillations"?
- some information about the kind of application. From what's said up to now, it can be anything like power supply, RF PA, whatsoever.

Are we in a guessing game, where it's neither allowed to tell clearly about the subject nor ask direct questions?

 

[ali2002]

Hello sir hope u r fine .Sir after acting upon ur advise i have come to a result may be the inductor core that i am using is not properly selected one.My frequency range of ripples is 1200hz .For 1200hz frequency range ripples which inductor core should i use?Can you suggest me one sir that will be helpful for me sir?waiting for your reply sir.
REGARDS

 

[FvM]

The questions, I asked two months ago are still pending.

 

[ali2002]

Hello sir i am still waiting about your guggestion about the core type(i.e. type of material of core) that i should use in making inductor .My frequency range is 1khz to 1.2Khz and load is 100A.Ithink in my previous filter inductoir is not working properly as i am using ferrite core.

 

[FvM]

I am still waiting about your guggestion about the core type.

My frequency range is 1khz to 1.2Khz and load is 100A.

It's the first time that you mention about a power filter (although it could be guessed somehow).

All power inductors in a frequency range up to a few kHz that I'm aware of are using regular silicon steel transformer cores.