what is differant between LPF circuit and Intigrator and so with HPF circuit and diff

[MrAl]

Hi again,

What it sounds like you are saying is that you accept some circuits as being an application specific integrator but you dont accept other circuits as being an application specific integrator.

All i have left to say then is that if you dont like a given circuit for a given application then dont use it. If you would rather use a more conventional circuit then use that instead. It's up to the individual to use what circuit they think is best and investigate the consequences.

Another aspect of this discussion is where we say that an RC network adds an integration operation to the circuit. That's because we have to use an integration to get the RC response:
Vo=Vi/s-Vo/s

where we have the integration of the output subtracted from the integration of the input which forms the new output. The integration of the output is the feedback.

 

[Winterstone]

Hi again,

What it sounds like you are saying is that you accept some circuits as being an application specific integrator but you dont accept other circuits as being an application specific integrator.
All i have left to say then is that if you dont like a given circuit for a given application then dont use it. If you would rather use a more conventional circuit then use that instead. It's up to the individual to use what circuit they think is best and investigate the consequences.
.

MrAl, I am afraid that - because of my bad knowledge of the english language - you did not get my point.
I never have used the phrase "aplication specific integrator" and all my arguments were not application specific.
My only intention was twofold:
(1) to explain to FLINTY the differences as well as common properties of a passive RC lowpass in comparison to an active circuit we call "integrator"
(2) to point to some errors (you call it "misquotes") I have detected in your replies - in particular your suggestion to use a passive second-order network (two RC sections) for integrating purposes.

This discussion has absolutely nothing to do with "application specific" integration actions. I think the term "integration" is unequivocal - only the realization can be different (with more or less approximation errors).
But I think, a newcomer/beginner like Flinty should know that it is not sufficient for integrating purposes to ensure a 90 deg phase shift for one single frequency. For example, this would be provided by each 2nd order active filter at the pole frequency - and I hope nobody would claim that the filter works as an integrator at this point. The relation to the time domain (Laplace) clearly requires a first order denominator in the transfer function .
Don't you agree?
Regards
W.

 

[MrAl]

Hi,


Well i agree that there are better integrators than others, but that's all i can agree with. An op amp integrator makes a pretty good integrator over all time, but a single resistor single capacitor also makes an integrator over shorter time periods.

"Integration" can mean we have an integrating circuit, we take a mathematical integration, or even that simply an integration in the system function. For example, we can say that when we add an RC network into a system we add an integration.

It sounds like you only want to call some things integrators and other things not integrators. If it doesnt integrate the way you want it to, you dont want to call it an integrator. Some things are usually called integrators (such as op amp and RC) and other things like we are talking about may be called RC networks, but that doesnt mean that we can not use an RC network (of any order) to obtain an integration as long as we know the limits of it's functionality. That's why i keep mentioning "application specific". In this case we would still call the RC network an integrator or perhaps an integrating network or "the integrating network".

And it only requires a first order denominator in the transfer function if it has to integrate for every possible application you can find. In other words, it would be most universal that way but that doesnt mean there are other circuits too.
For a good example, take an op amp integrator with a single polarity input output. It works just fine with a single polarity input. But try to use the opposite polarity and it doesnt work anymore. It's still an integrator, but i has limits just like any practical integrator.

 

[Winterstone]

Hello again MrAl,

It sounds like you only want to call some things integrators and other things not integrators. If it doesnt integrate the way you want it to, you dont want to call it an integrator.

I do not want to argue again against this. Instead, I think, I have given the answer with the simulation results and their evaluation/interpretation.

... and other things like we are talking about may be called RC networks, but that doesnt mean that we can not use an RC network (of any order) to obtain an integration as long as we know the limits of it's functionality. That's why i keep mentioning "application specific". In this case we would still call the RC network an integrator or perhaps an integrating network or "the integrating network".

I do not understand the meaning of these sentences (...other things...?). Please, can you give an example for one RC network - other than those we have discussed discussed - which we could use as an "integrating network" ?

For a good example, take an op amp integrator with a single polarity input output. It works just fine with a single polarity input. But try to use the opposite polarity and it doesnt work anymore. It's still an integrator, but i has limits just like any practical integrator.

I think, this example has nothing to do with the subject of our discussion. I suppose nobody will expect a negative opamp output when the supply voltage does not allow it. We are discussing circuits and not restrictions caused by single supply operation.
______________________

Finally, I propose to end this special discussion since the question from Flinty has been answered, I think.
It would be fine if Flinty could confirm that the simulatiuon results I have shown clearly reveal the differences and common properties of the passive and active "integrator versions".
______________________
Our discussion, perhaps, has demonstrated that integrating circuits are interesting devices.
I have an example for this claim - and perhaps it deserves an extra thread:
Two integrators in a closed loop (inverting/non-inverting) form a harmonic oscillator.
But, surprisingly, the oscillation frequency does NOT exactly fulfill the oscillation criterion as far as the loop phase is concerned. The loop gain - as expected - is LG=0dB.
And another question: What about those frequencies with loop phase=0 deg but loop gain LG>0 dB ?
All other positive feedback oscillator circuits wiil oscillate (with clipping effects) at such a frequency.
____________________
All for now
W.

 

[ericgibbs]

hi Guys,

I suspect the OP, 'flinty' has almost given up on getting a simple understandable answer for his original question.

Is there any chance we could give a simple summary in order to help him.?


E.

 

[Winterstone]

hi Guys,
I suspect the OP, 'flinty' has almost given up on getting a simple understandable answer for his original question.
Is there any chance we could give a simple summary in order to help him.?
E.

Hi Eric,

good suggestion. What do you think about the simulation results as given as an attachement to my posting #18 ?
For my opinion, these results clearly show for the two candidates (RC passive and RC active) which frequency regions are suitable for integration purposes.
And the realtionship between 1st-order low pass filtering and integrating can be derived from these graphics.
I really don't know if another kind of "summary" is possible resp. necessary. Perhaps the "OP" (Flinty) should give an answer.

Thank you.
W.

 

[MrAl]

Hi Eric,

Yeah i have a feeling there is a language barrier here. I've quoted circuits where a limited form of the integrator works even though it's not strictly speaking an integrator, then note that this is an application specific issue, the other party says that it can never be used, then later agrees that it can be used in another application, then says it's not application specific. So im stumped as to what this discussion is about now

I guess i would sum up as saying that there are "Classic" integrators that work with a wide range of input types and not so classic integrators that only work with a small range of input types. An example of the classic integrator is the resistor, capacitor and op amp circuit, and an example of the not so classic integrator would be an RC network with resistor and capacitor. The latter circuit being more limited in it's application.

And with that i'll end my replies now so that this thread doesnt get any more confusing. Many thanks to all that participated.

 

[ericgibbs]

hi,
I would have thought with the explanations given so far and the plots in post #18, the OP could now refine his original question and perhaps say what is the actual application.

hi flinty,
Would you please post more details of the application and exactly what you are trying to do.?

E.

 

[Winterstone]

Yeah i have a feeling there is a language barrier here. I've quoted circuits where a limited form of the integrator works even though it's not strictly speaking an integrator, then note that this is an application specific issue, the other party says that it can never be used, then later agrees that it can be used in another application, then says it's not application specific. So im stumped as to what this discussion is about now

One last word from the "other party":
Thank you, MrAl, for this nice and fair summary. Nevertheless, I suppose everybody who has read the last 24 postings will be able to weight your wordings.
Regards.
W.

 

[Winterstone]

hi,
I would have thought with the explanations given so far and the plots in post #18, the OP could now refine his original question and perhaps say what is the actual application.
hi flinty,
Would you please post more details of the application and exactly what you are trying to do.?
E.

Hello Eric,
because it has happend within this thread: May ask a more general question?
Is it possible and allowed that an author modifies/correct his own contribution - even in case it was already quoted in a corresponding reply? For my opinion, this would destroy the logical sequence of question and answer.
Thank you.
Winterstone

 

[ericgibbs]

hi Winterstone,

Many posters go back and edit earlier posts, as you say it can make the later replies appear out of context.

One way that it could be done, would be to leave the 'original' text unedited and simply add new text to that post as required, giving a reason for the edit.

This, hopefully should keep everyone happy.

example:
EDIT:
Correction/Amendment to the above text......

 

[Winterstone]

Eric, thanks for the answer.
Yes, of course this could be one way - but, more than that, I would appreciate if the moderators would urgently require this procedure as proposed by you.
Otherwise, a fair exchange of arguments (pro/con) and a reasonable sequence of questions and answers (in particular: corrections of errors) is not possible.
This is true, in particular, for forum members not directly participating on the discussion. They must feel lost if they cannot follow the arguments.
I think, especially for newcomers/beginners such an exchange of arguments is a very good way to learn (a) something about electronics and (b) how to discuss in fair manner.

Thank you and regards.
W.

 

[garimella]

I think LPF and Integrator cannot be equated. An LPF acts as an attenuator and not as an integrator at high frequencies (revealed through bode plot).

 

[garimella]

I agree with winter that LPF acts as an integrator only at high frequencies. LPF computes the average at high frequencies which reflects as an attenuated output

 

[MrAl]

Hi,

This is an old old thread, but here is a graph of the two. The main differences are the amplitude is very different at DC and similar at higher frequency, and the phase is constant for the integrator and varies for the LPF near DC, and they are both constant for higher frequencies.
A striking difference is the integrator has infinite response to DC, while the LPF has finite response.

blue: integrator amplitude
red: LPF amplitude
green: integrator phase
violet: LPF phase