How to select a low-pass filter

[DragonPetter]

Try do the math on the integrator I designed in #13.

I can treat it as an integrator and do the math on the integrator. The switching complicates it with non-linear cases, where I would have to break the problem down into time slices of operation depending on the state of the switch. There are not exact solutions for non-linear circuits and the whole problem with your challenge is that the switching, non-linear part has nothing to do with the circuit integrator or with any filtering. It is only a practical complication you introduced to give you certain results that apply to the application.

If I take a textbook, classic, low pass filter, and I put a diode on the input of it so that it breaks down into a non-linear circuit analysis problem with no analytical solution, has my textbook, classic, low pass filter block suddenly become a different circuit and I can no longer call it a low pass filter?

 

[yungman]

Both parts are math and both parts are analog. The switching part is non-linear and it has nothing to do with low pass filtering and is not inherent to an integrator. Do some applications use a discharge switch to reset an integrator? Yes. Is the reset switch part of the basic fundamental idea of an integrator? No.

Reset switch is a real part of a real integrator. Again, read #13 where I explained why a theoretical integration like a LPF don't work well in real life where you have to wait for it to discharge.

 

[yungman]

I can treat it as an integrator and do the math on the integrator. The switching complicates it with non-linear cases, where I would have to break the problem down into time slices of operation depending on the state of the switch. There are not exact solutions for non-linear circuits and the whole problem with your challenge is that the switching, non-linear part has nothing to do with the circuit integrator or with any filtering. It is only a practical complication you introduced to give you certain results that apply to the application.

If I take a textbook, classic, low pass filter, and I put a diode on the input of it so that it breaks down into a non-linear circuit analysis problem with no analytical solution, has my textbook, classic, low pass filter block suddenly become a different circuit and I can no longer call it a low pass filter?

And this is a real life integrator and that's where the textbook ends.

 

[DragonPetter]

Reset switch is a real part of a real integrator. Again, read #13 where I explained why a theoretical integration like a LPF don't work well in real life where you have to wait for it to discharge.

If I add a switch in a circuit with a light bulb, is it a special real life light bulb now since it can now turn on and off? By your argument, it is no longer a light bulb (just as you say the LPF integrator circuit is no longer a LPF when a switch is added). I would argue its still a light bulb, its just working with a switch function to enhance/accomplish a more specific task. Likewise, a LPF integrator circuit is still a LPF, but you add the switch to give it a more desired behavior . . that doesn't mean the LPF circuit goes away. And believe it or not, integrator circuits are used without switches in some applications.

 

[DragonPetter]

Also, this might resolve the disagreement. Can you tell me why the switch is there? The opamp you use for the integrator will saturate, and so the integrator can no longer increment the voltage, right?

 

[rbj]

No, only by definition. As I repeat over and over, read #13 about the real life circuit application and the defect of the simple integrator.

and you're still wrong. an integrator is a low-pass filter.

I never argue about the math, I said that is the easy part.

sounds like an easy way out of a losing argument. there are all sorts of crackpots that promulgate their crackpottery who say, in defense, "don't bother me with the math."

Implementation is the key and that's where the rubber hit the road.

it's an issue of definition of terms. implementation is a different thing altogether.

This is EE forum, not the Calculus or physics forum. BTW, I know the math, but that's not the point.

it's an EE forum, and when you say that "an integrator is not a low-pass filter", this electrical engineer (me) is here to say that is decidedly incorrect. the integrator depicted in this circuit:
http://www.electro-tech-online.com/...ineair-when-using-integrator-opamp-setup.html
is decidedly an LPF. and it is so by the definition of LPF.

r b-j

 

[yungman]

If I add a switch in a circuit with a light bulb, is it a special real life light bulb now since it can now turn on and off? By your argument, it is no longer a light bulb (just as you say the LPF integrator circuit is no longer a LPF when a switch is added). I would argue its still a light bulb, its just working with a switch function to enhance/accomplish a more specific task. Likewise, a LPF integrator circuit is still a LPF, but you add the switch to give it a more desired behavior . . that doesn't mean the LPF circuit goes away. And believe it or not, integrator circuits are used without switches in some applications.

I never disagree with the math, but it's only a small part. The integrator defined in the math is not very useful as stated in #13.

 

[yungman]

and you're still wrong. an integrator is a low-pass filter.
sounds like an easy way out of a losing argument. there are all sorts of crackpots that promulgate their crackpottery who say, in defense, "don't bother me with the math."

Come up with the math of the integrator in #13

it's an issue of definition of terms. implementation is a different thing altogether.
it's an EE forum, and when you say that "an integrator is not a low-pass filter", this electrical engineer (me) is here to say that is decidedly incorrect. the integrator depicted in this circuit:
http://www.electro-tech-online.com/...ineair-when-using-integrator-opamp-setup.html
is decidedly an LPF. and it is so by the definition of LPF.

r b-j

You design a real integrator circuit? or you just do it in DSP and programming?

The circuit diagram you pull out in the link really shows how much you really know about circuit. That example you show does not work in real life. Go bread board this and see how good an integrator you get. Output will rail out before you have time to put a probe on. And you come here to insult?

That's the problem people study theory and talk. OR doing programming and think they know.

This is a technical discussion forum, not a place to trade insult.

As I said, the one you show or the LPF is the most basic form of integrator. If you don't understand this, don't even talk integrator. I expect people here talking know a WHOLE LOT MORE than this. Read the #13 on part that I described the short coming of the basic integrator.

Case in point, the paper I publish is base on the self reset integrator that over come the last generation system using a LPF as integrator. The original system need 300nS for the LPF to settle before it is ready to accept the next event. My design cut the dead time to 25nS. I got $500K of government SBIR grant and successfully developed the new product.

http://rsi.aip.org/resource/1/rsinak/v71/i11/p4144_s1?isAuthorized=no

And don't say I don't know math, I do plenty of math in EM and RF, those are more the steady state where all the vector calculus, PDE, fourier, laplace transform become useful. do a search of my pass posts on those subjects, I had more math and formulas than people care to see.

 

[yungman]

Also, this might resolve the disagreement. Can you tell me why the switch is there? The opamp you use for the integrator will saturate, and so the integrator can no longer increment the voltage, right?

Yes, but more than that. It's the ability to reset after the event and ready for the second event in a short time. Like I explained in #13, without the reset, you have to wait for the output to recover back to ground state before you can do another integration. Or you will be piling up and getting error.

 

[rbj]

Yungman, i have never insulted you. i have compared your "math ... is the easy part." argument with those used by crackpots. no where did i call you a crackpot.

nonetheless, i will make it clear to other people reading that you are decidedly mistaken when you say that an integrator is not an LPF.

i am not decoding all you "repeat[ed] over and over" in "#13". to make it clear, i pointed to a circuit schematic of what we mean by "an integrator". if your integrator has a shorting or reset switch, it's not a time-invariant system anyway unless you never press it. by "integrator", i and the other readers here, mean an LTI system.

in my previous life (i was a grad student about 35 years ago), i designed and built a sawtooth VCO out of such an integrator (and it had a transistor shorting switch), a state-variable filter out of integrators, and a PID controller out of integrators. we even had, in our EE department, an analog computer left to us by the Air Force (i'm sure it's gone into the trash by now) that had integrators. and all the integrators took the form of the circuit shown in the graphic: http://www.electro-tech-online.com/...ineair-when-using-integrator-opamp-setup.html . of course it wasn't a perfect integrator because the op-amp was not an ideal op-amp. of course there is a DC issue because the op-amp would saturate if there was any DC getting integrated, including that non-zero DC bias going into the inverting and non-inverting terminals of the op-amp (so this integrator had to be surrounded by other parts and feedback to avoid that).

so, it would be accurate to say "no perfectly ideal integrator really exists in practice", but it is not accurate to say that "an integrator (ideal or not) is not an LPF." even if you repeat it over and over again, you are simply wrong over and over again.

no insults, no BS, just the facts.

r b-j

 

[DragonPetter]

Yes, but more than that. It's the ability to reset after the event and ready for the second event in a short time. Like I explained in #13, without the reset, you have to wait for the output to recover back to ground state before you can do another integration. Or you will be piling up and getting error.

Yes, so can you tell me what either a or b has to do with integration fundamentally?

a. saturating your opamp
b. reset after an "event" and ready for a second "event"

I mean, the fundamental act of integrating, not what your application wants. Events are arbitrary and what you define as an event means little other than a detail for your application in the realm of what integrator implies.

A. This is a practical limitation of the technology and means nothing to an integrator more than it means to any other function you might encounter taking you to an amplifier rail.
B. You are taking samples here, that simple. A sample is not an integration. You are choosing your limits of integration basically. A pure integrator in the sense that we call it a filter does not have these limits of integration; it is a moving integration of all past signal.

Also, piling up error is a fundamental factor in integration (look at trying to track displacement with an accelerometer for example). Its not just something specific to a circuit called the integrator. Error is still part of the integration process. You may be correcting for it in your application though, but your application doesn't define what everyone else calls an integrator. This really just seems like an argument of semantics again on what a term, the integrator in this case, applies to. It seems like yungman knows what a LPF is of course, and he knows that a LPF can act like an integrator, its just the philosophical standpoint that he draws a line at the integrator circuit that the rest of us do not.

 

[yungman]

The LPF integrator is just a fundamental building block for a real practical integrator. This fundamental building block is as simple as it get. It is the implementation into a real integrator that is the majority part of it. It is the difficulty of designing the integrator to do the job per the requirement that make it hard. Inside the real integrator, there is that little block of ideal integrator like your circuit, but the support circuit is what make the real integrator.

I assume at our level of discussion, everyone know the basics of a simple integrator function and build onto it. The integrators that I designed can never be used as a LPF in any stretch unless you eliminate all the support circuit. I never even consider that simple basic integrator building block is a stand alone integrator, it is only part of the whole picture of a real integrator.

Back to the op's question, if you say a basic integrator building block like the diagram you show is a LPF, I don't have any problem. But I cannot agree on a blanket statement of an integrator is a LPF. To me as a circuit designer, integrator is a composition of circuits including the small integrator building block. And there are different kind of integrator designs depend on the need.

The integrator in #13, the center of it is just a simple integrator like the circuit you posted. BUT I had to design the circuits to pulse the input to charge up the integrator to a steady level representing the charge of the pulse. Then the same pulse travel through a 15nS delay line and inverted and buffered before driving into the summing junction of the same integrator. The second pulse being same shape by inverted will discharge the capacitor and everything return to 0. So in the middle of the event, I have a 15nS integrated flat pulse for sampling. Then after the event, the circuit only need about 3 to 4 nS for it to totally recovered and ready for the next pulse.

This whole integrator design is so far remote from the basic function block you can't really put it in the same statement. We just stipulate that there is that functional block in the circuit everyone else called an integrator. But my circuit comprises of pulse amplifiers, delay lines, compensate network etc. The basic "integrator" block is only about 10% of the circuit.

 

[DragonPetter]

The integrators that I designed can never be used as a LPF in any stretch

Then it isn't an integrator. The very act of integration, which is what an integrator should do, is low pass filtering the signal. You cannot escape that, and so you can never escape the argument you make.'

Let me ask you this. During operation, which I assume excludes dead time where you shunt the signal, or perhaps you corrected for dead time, what does the output signal look as a function of the input signal? If you can tell me its an integration (which you have to say it is if you want to call it an integrator) then by default you have admitted that it acts as a low pass filter during this operating period. My first post linked to a thread that shows how the operation of integration in the time domain results in a low pass filter behavior in the frequency domain.

The integrator in #13, the center of it is just a simple integrator like the circuit you posted. BUT I had to design the circuits to pulse the input to charge up the integrator to a steady level representing the charge of the pulse. Then the same pulse travel through a 15nS delay line and inverted and buffered before driving into the summing junction of the same integrator. The second pulse being same shape by inverted will discharge the capacitor and everything return to 0. So in the middle of the event, I have a 15nS integrated flat pulse for sampling. Then after the event, the circuit only need about 3 to 4 nS for it to totally recovered and ready for the next pulse.

This whole integrator design is so far remote from the basic function block you can't really put it in the same statement. We just stipulate that there is that functional block in the circuit everyone else called an integrator. But my circuit comprises of pulse amplifiers, delay lines, compensate network etc. The basic "integrator" block is only about 10% of the circuit.

All of the extra stuff does not define an integrator. I get the impression that any textbook that wants to teach what an integrator is and what it does should have a figure of your schematic in it. If I ask a colleague about integrators, should I say "the yungman integrator" when I want to talk about a real one as opposed to a textbook one?

The basic "integrator" block is only about 10% of the circuit.

There you say it yourself. The very circuit you're arguing over isn't even the integrator. You argue that all of the ancillary functions and processing that you claim in your circuit cannot be a low pass filter, but that 90% stuff you are arguing over isn't even the definition of an integrator. An integrator is not a complex circuit like you describe, and its a fundamental function that also happens to low pass filter a signal. You have admitted that the integrator part of your circuit behaves as a low pass filter, isn't that the end of the discussion? Everything else is just bells and whistles that no one here has assumed works as a low pass filter. It sounds like your design was some kind of charge accumulator or low current/low signal measurement, and in that case you aren't using it directly as a low pass filter, but you are still using the behavior of a low pass filter to make your measurement, whether you like it or not.

 

[yungman]

I consider that integrator as a whole circuit, not just that little small block. All the surrounding circuits are part of the integrator.

I design plenty of low pass, band pass filters. I designed active filter, state variable filter. But my main experience is passive multi-pole LC filters. Bessel, elliptical, Butterworth either in lump components or microstrip or stripline designs.

Any of you try using a elliptical or Butterworth multi pole LPF as an integrator. It really do not work very well. In pulse application, it rings, it under shoot and ring for a long time.

The simple op-amp circuit and Bessel type of filter, or linear phase type of filter can still work as a simple integrator building block, but elliptical or butterworth has very little usefulness to perform any integration. But they are all LPF. That's exactly what the prior design that I replace in post #13. They use a 4 pole LC filter and it ring and ring.

 

[yungman]

Then it isn't an integrator. The very act of integration, which is what an integrator should do, is low pass filtering the signal. You cannot escape that, and so you can never escape the argument you make.'

Let me ask you this. During operation, which I assume excludes dead time where you shunt the signal, or perhaps you corrected for dead time, what does the output signal look as a function of the input signal? If you can tell me its an integration (which you have to say it is if you want to call it an integrator) then by default you have admitted that it acts as a low pass filter during this operating period. My first post linked to a thread that shows how the operation of integration in the time domain results in a low pass filter behavior in the frequency domain.




All of the extra stuff does not define an integrator. I get the impression that any textbook that wants to teach what an integrator is and what it does should have a figure of your schematic in it. If I ask a colleague about integrators, should I say "the yungman integrator" when I want to talk about a real one as opposed to a textbook one?



There you say it yourself. The very circuit you're arguing over isn't even the integrator. You argue that all of the ancillary functions and processing that you claim in your circuit cannot be a low pass filter, but that 90% stuff you are arguing over isn't even the definition of an integrator. An integrator is not a complex circuit like you describe, and its a fundamental function that also happens to low pass filter a signal. You have admitted that the integrator part of your circuit behaves as a low pass filter, isn't that the end of the discussion? Everything else is just bells and whistles that no one here has assumed works as a low pass filter. It sounds like your design was some kind of charge accumulator or low current/low signal measurement, and in that case you aren't using it directly as a low pass filter, but you are still using the behavior of a low pass filter to make your measurement, whether you like it or not.

As I repeat, a real integrator has so much more than just that little building block. We all know the fundamental of that. The real integrator need a whole lot more that just that little block.

 

[DragonPetter]

I consider that integrator as a whole circuit, not just that little small block. All the surrounding circuits are part of the integrator.

How does 100% = 10%? Pulse amplifiers, delay lines, and compensator networks are part of an integrator, or they're part of your integrator? Again, I get the impression that if someone wants to talk about a real integrator, they need to talk about your design. Does your integrator define all integrators such that you can make the statement that an integrator is not a low pass filter?

Again, the disagreement is resolved if you answer my question about what your circuit does. If a voltage signal is supplied on its input during its integration period of operation, and it gives a time domain voltage output that is proportional to the integral of the input voltage, then it has low pass filtered the signal in some way, no matter what the circuit actually looks like. If you disagree with this statement, then you disagree with mathematics.

 

[yungman]

How does 100% = 10%? Pulse amplifiers, delay lines, and compensator networks are part of an integrator, or they're part of your integrator? Again, I get the impression that if someone wants to talk about a real integrator, they need to talk about your design. Does your integrator define all integrators such that you can make the statement that an integrator is not a low pass filter?
My design is one case of the integrator only.
Again, the disagreement is resolved if you answer my question about what your circuit does. If a voltage signal is supplied on its input during its integration period of operation, and it gives a time domain voltage output that is proportional to the integral of the input voltage, then it has low pass filtered the signal in some way, no matter what the circuit actually looks like. If you disagree with this statement, then you disagree with mathematics.

I don't think you hear what I said. I don't disagree with the basic integrator or the math. I said there is so much more in implementing a real integrator that just that little circuit. I take it that is only part of the complex design. Every integrator design is specific according to the requirements.

I used my circuit as an example, not every integrator are like this. There are other designs of integrators with FET switches and all.

 

[DragonPetter]

I don't think you hear what I said. I don't disagree with the basic integrator or the math. I said there is so much more in implementing a real integrator that just that little circuit. I take it that is only part of the complex design. Every integrator design is specific according to the requirements.

I used my circuit as an example, not every integrator are like this. There are other designs of integrators with FET switches and all.

Sorry, I have gotten lost in all of the stuff we've talked about, so I hear what you said now. But, now that we are done with this (I'm really tired of typing integrat---). Do you think that integrators are low pass filters, although they may not be used as a low pass filter (like in your design)? I think that was the original and probably only disagreement people have had in this thread.

 

[yungman]

Sorry, I have gotten lost in all of the stuff we've talked about, so I hear what you said now. But, now that we are done with this (I'm really tired of typing integrat---). Do you think that integrators are low pass filters, although they may not be used as a low pass filter (like in your design)? I think that was the original and probably only disagreement people have had in this thread.

The basic integrator building block is a LPF, I never disagree on this. It's the additional requirements that required on the real circuit that I am referring to. I look at it as a system, not just that particular circuit. Integrator design can be challenging.

 

[rbj]

This whole integrator design is so far remote from the basic function block you can't really put it in the same statement. We just stipulate that there is that functional block in the circuit **everyone else** called an integrator. But my circuit comprises of pulse amplifiers, delay lines, compensate network etc. The basic "integrator" block is only about 10% of the circuit.

I consider that integrator as a whole circuit, not just that little small block.

yungman, it's not about what you consider to be the case. it's about a common lexicon of terms and concepts. you are not the czar of this lexicon, nor am i nor anyone else (but "everyone else" might have authority to define a term such as this). the term "integrator" in the context of electrical engineering has meaning outside of your world of activity or mine or anyone else's. it's in textbooks. it's in published papers. it's on the internet. it's in Wikipedia.

i think we should ditch the thread now. there is no meeting of the minds, and there need not be. but i won't let your personal lexicon sit uncorrected on this forum where it may mislead other people. and i haven't and nothing else is needed from me about it.

 

[yungman]

I maintain that talking about integrator in that basic configuration is too simplistic a point of view. You need to be more specific on the application.

Talking about integrator as just a LPF inverted op-amp with a cap as feedback is like talking about a RF power amp as just a transistor in common emitter or common source configuration. You miss the moon talking about it like this. People can write a whole book on JUST this common emitter stage when you get into the RF world. It is all about the matching, filtering, efficiency and other calculation and design. Yes the simple little few formulas of basic transistor is still the main equation, but there is just so much more to this little common emitter stage than just the mathematical equation. It is not just a common emitter stage.

Lastly, this is my opinion and it's not up to you to say whether it is correct or not. Many people here are studying to be an engineer. It is important to know not to just look at everything just from the textbook. A lot of people can get good grades but never transition into a good engineer because they cannot transition from the textbook theory into real world. That's the reason it is more important to talk about application, looking at things in a more macro point of view. I spent 30 years designing these kind of circuits and looking at an integrator as just the circuit you show is woefully inadequate.