(?) That's a long time to be working on the problem.

 

And I created many different filters including Butterworth 3rd order filters.

 

And I created many different filters including Butterworth 3rd order filters.

So why did you ask the question about an analytical solution, since you apparently already know that?

 

So why did you ask the question about an analytical solution, since you apparently already know that?

 

I would like to calculate it, then optimize the capacitor values, and then select the resistor values. Now I am using the web site which you mentioned before and optimizing the filter parameters to get the Butterworth parameters (or other) as close as possible. The web site tool is not accurate as direct calculation. You can compare their website tool for 2nd order filter and calculate the values directly and you can see it. It is easy.

 

"... a virtual Ground is not needed."
What are you trying to say? What you recommended was just using a virtual ground. When using a voltage, which IS NOT GND of Earth as Ground for a circuit then you call that virtual ground. Because voltage is relative, the Real Ground then gets a VIRTUAL negative voltage ( negative voltage at virtual ground ), but NOT a REAL negative one. Then as i said it many many times before, i would have to apply a offset to the SIGNAL. I need a REAL negative Voltage with low noise. A transformer can't achieve such a purpose. That's why i ordered that chinese dual voltage board. If i could smooth out the voltage of that chinese board, i would be able to use it, because it outputs REAL negative voltage. You can see that it is real negative voltage, by applying the negative probe of your multimeter to REAL ground and the positive probe to the voltage, you want to measure. If the multimeter then show's you a negative value, the voltage really is negative and not virtually negative.
Why do you still recommend that virtual voltage thing @Audioguru again ?
Disadvantages: You have to apply a DC offset to the signal, which later has to get removed, which is (as far as i know) very hard.

Please recommend a method to get REAL negative voltage relative to ground, with low noise in it.
Thank you, cheers

 

I showed you a simple unregulated +14/-14V circuit using a center-tapped 22VAC transformer. If has output filter capacitors to kill any noise. It is simple so it produces a little amount of 100Hz or 120Hz hum which is not fed to the opamp inputs so it is reduced to almost nothing. Look at the Power Supply Rejection of an audio opamp. An OPA2134 dual audio opamp has 100Hz power supply noise reduced 100dB which is one/100000, higher frequencies are reduced even more. You can use a higher voltage transformer and use 7815 and 7915 voltage regulators if you want.

A "virtual ground" circuit has everything grounded to the voltage divider. In a simple positive supply for opamps then the input and output have coupling capacitors (that remove any offset voltage) and the opamps are grounded to the real ground, not to the voltage divider. The voltage divider is filtered and is used only to bias the input of the opamps which require an extremely low current. There will be a POP while the coupling capacitors charge unless their RC's are balanced so that they charge equally.

Here are the two ways an opamp can be powered:

 

Thank you so much, that was just what i wanted to know If you could tell me now, which values to choose for the Resistors and for the Capacitors, i would be even more thankful
Greetings Lukas

 

You need to learn the basics of electronics and look at the opamp spec's in its datasheet to calculate resistor and capacitor values.
Some online calculators can do it if your know the details of what you want the circuit to do.
You might find a project online that does what you want but it might have errors.

 

Okay, i just watched a presantation of how that circuit you recommended me works. I got 2 Questions about it:
Why is there A capacitor at the 6v to GND rail/ The non inverting input to the Ground. ( We are talking about an inverting configuration and a Single supply, so the picture at the bottom right corner ) The next question is just for helping me understanding:
How does the Output capacitor remove the DC offset of 6v? I mean, it gets charged up to 6v and the electric Field, which is building up, attracts the opposing charges ( In that case electrons ) on the other side, right? So i will get -6v there?
Or is this just what you meant by having to calculate capacitor values? ( The resistor values should be as high as possible to keep the current low. The Formula Loks like that:
1 /R1 = 1/ R2 + 1/ R [input impedance opamp]
The resistors for the Gain already have been calculated.)
What capacitors do i need if i want to use a TL072 with a +24v ( so +-12v ) single power supply? Those are called smoothing and coupling capacitors, right?
Thanks again a lot

 

That's the Circuit, i know came up with. I just have to calculate the input and output Impedances, to determine a Coupling capacitor for the usage of one Power supply.
I only want to ask you, whether this circuit seems okay or not, or if you expect it to work or not.
Thank you

 

An output capacitor passes AC but does not pass DC. Also it does not charge and discharge by the signal, instead it charges when the power supply is turned on and it discharges when the power supply is turned off.

I posted this in your other thread about converting this circuit to use a single positive supply:

 

Here is another possible circuit which does not require an extra op amp as long as the input source impedance is low enough. If not, reduce the value of the first 9.1k so the total series resistance is around 9.1k.
Also note the input source must be DC coupled not AC, if it is AC we have to add another resistor.

Note the additions:
One 9.1k resistor going to +Vcc
Two 22k resistors (although they can be almost any reasonable value as long as they are the same) one goes to 1/2 Vcc.
Change of two capacitor values.
Reposition the 160k resistor going to 1/2 Vcc.
Small changes in the 25nf capacitor should cause either a smoother response or a more peaked response going up or down on the value of the cap (one way makes it smoother, the other way makes it more peaked but dont change it too much).

Do a simulation to make sure it works right this was all done theoretically.
Also note the circuit was reduced to gray scale for easier reading (ha ha that green junk is nuts)

 

Here is another possible circuit

Where?

 

MrAl,
This circuit has its +24V fed from a noisy switching power supply. In your new circuit you are feeding that noise directly into the inputs of the opamps.
My circuit had the Vcc/2 biasing filtered to remove the switching noise.

Oh, are you thinking that this lowpass filter circuit will not pass high frequency switching noise?

 

Audioguru again
If i understood your circuit correct, it should look like this:
Firstly i want to Thank you, for your help etc, but i got some questions...
(actually without the blue selected wire )

If this circuit is what you meant, there will be a voltage of the unfiltered input signal fed into the non inverting input of the last opamp, right?
Is this "noise" enough to affect the complete filter circuit and mess it up?
I mean, there is a capacitor (22uF) to firstly Filter the input voltage (12v) but also Filter the noisy signal which is being added there, but is this enough?

What frequency did you use for calculating fc (-3db) for the input capacitor?

This is my simulation of the input circuitry and how it should affect the Signal. ( The one resistor which usually is connected to +24v is connected to gnd, since i don't have the possiblity to simulate this in that program. But it is the same, right? A DC voltage source does not change the filtering aspect here, right? )

If this simulation is true, my signal will be completely weak, that is also what came up with my upper simulation.
I have one idea of what to do:
use all opamps as inverting filters, so i can just connect half the supply voltage to the non inverting input, right?
Then with 2 opamps the signal will be 360° phase shifted. ( This only affects my Line outs, since they are playing a bit earlier then. If this is the case, i will add 2 inverting voltage followers/buffers to my line out, to again get 360° phase shift. )
Thank you again and
Greetings from Germany

 

MrAl,
This circuit has its +24V fed from a noisy switching power supply. In your new circuit you are feeding that noise directly into the inputs of the opamps.
My circuit had the Vcc/2 biasing filtered to remove the switching noise.

Oh, are you thinking that this lowpass filter circuit will not pass high frequency switching noise?

Well actually the Vcc/2 lines all have to come from a divider anyway if this is to work off of a 24v supply only.
The straight Vcc line could be filtered too if needed.
In fact now that you mentioned that maybe the entire Vcc line should be filtered to all places in the circuit.

So some filtering could be used there too, unless the filter itself rejects the noise that is.

 

Where?

At the end of post #93. You dont see it?

 

Here is another solution which doesnt required using unusual capacitor values...

[LATER]
I thought i should mention that if we change the 97k and 172k resistors slightly we can get a flatter response in the pass band, but we loose some sharpness in the cutoff region. We can also allow a little more overshoot (like maybe 5 percent) and get a sharper response.
I was able to get a flat response to within 0.1 percent but the response becomes somewhat less sharp. Allowing 3 percent overshoot the response sharpness is the same as the original circuit.

 

Here is another solution which doesnt required using unusual capacitor values...

[LATER]
I thought i should mention that if we change the 97k and 172k resistors slightly we can get a flatter response in the pass band, but we loose some sharpness in the cutoff region. We can also allow a little more overshoot (like maybe 5 percent) and get a sharper response.
I was able to get a flat response to within 0.1 percent but the response becomes somewhat less sharp. Allowing 3 percent overshoot the response sharpness is the same as the original circuit.

Here are the three main types of responses possible by varying those two resistors a little bit.
One is slow and smooth, the middle one (97k,172k) is only slightly peaked but sharper, the last one is very peaked but sharper yet.