Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

guidance with op-amp bandpass filter, help!

Status
Not open for further replies.
Therion said:
those dotted lines are supposed to be ‘nets’ that indicate the actual wiring of the circuit, so yeah, your right, op amp 2 has been shorted!
Lots of things are still shorted.

How does changing the ratio between the resistors for gain
That is how opamps work.

effect the resistance you use for the Fc derision?
That is how filters work.

I admit I am unsure how you got 2k resistance value for the HPF circuit?
You had the 1nF capacitor in series with two 1k resistors. The resistors have a total resistance of 2k.

[/quote]Any advice on how could I combat the opposing phase shifts in each filter circuit?[/quote]
Filters have phase shift. You can't change the phase shift of a filter.

Your new lowpass filter does not have the correct frequency response. Its cutoff frequency looks to be much lower than 15kHz.
The graph shows a huge signal loss, instead of the gain of 500 from the 3rd opamp.
 
Yup, unfortunately, the nets are not the only indication of connection, the majority of those nets are running beneath the wires that are present.

I thought op-amps derived their gain by the ratio between the gain resistor and the feedback resistor, for op – amp 2, and although there are two 1k resistors in series, there is still a 500k ohm feedback resistor present.

Im happy with the frequency response as it is, I just wish to see it working. And yes, I see that there is large signal loss occurring, this is what im trying to remedy. .

Im sorry, have I done something to warrant you being quick with me? If I’ve done something to wind up or offend, please let me know ..
 
Hi Therion, I think you are doing great in your *paper* experiments. I wish the simulators had been available when I started. Keep up the digging and asking questions and you will end up understanding and will be able to avoid common mistakes when building the actual hardware.

Your most recent diagram still shows Op Amp 3 shorted out, its not there. Therefore, you don't get any gain from it at all. I also noticed you changed Res 8 from 1K to 500K. If the Op Amp 3 was not shorted that would give you a gain slightly less that 1, i.e., 500k / (500K + 1K). The gain is Rfeedback / (Rinput) where Rinput is the total series resistance in the input leg. Cap 4 is also in the series input leg but is not a resistance - it causes phase shift and affects the LPF cutoff but not midband gain. Looking from the input of Op Amp 3 toward the output of Op Amp 2 you see Res 13 + Res 8 (originally 1K + 1k to give the 2K mentioned above). They make up the Rinput, the Total resistance. The output impedance of Op Amp 2 is very low so its output impedance (resistance) compared to Res 13 and Res 8 is negligible. So that's why the Rinput is calculated as 2K previously and now is501K in the new circuit. Cap 4 affects the LPF frequency calculations. So if you change Rinput, to change gain, then you must adjust Cap 4 to restore the ratio to give you the desired frequency response.

In the formula: Fc = 1 / (2 x Pi x R x C), note I added the ( ) to clarify that 1 gets divided by all the denominator terms, consider the RC product as controlling the frequency response - everything else is a constant. Hence if either R, (the total resistance as discussed above) or C, (Cap 4) change, then Fc will also change. But you probably understood all this anyway.

I really haven't looked at calculating your Hi and Lo cutoff frequences but even if the shape of the Amplitude curve is what you want, the -133dB tells you its 133dB of attenuation, not gain. When you are looking for Gain, a Negative dB, rather than a hoped for Positive dB, means it is signal attenuation. [The negative dB can also mean a 180 degree phase shift between input and output signals. Since the sign of the dB can be interpreted two ways it can get a bit confusing but if the sign (+ or -) of the expected Gain is opposite what you are expecting, then you have attenuation.] Clear as mud??? Note that for an Op Amp, signal input on the negative, (-) terminal or inverting input will give a 180 degree inversion of the signal (at MIDBAND - the frequency range where the LPF and HPF neither give any attenuation), where as a signal input on the non-inverting, (+) input will give a positive gain, slightly different in magnitude than the negative input gain.

Just as Gain is controlled by the ratio of Rfeedback and Rinput, the phase shift at higher and lower frequencies (outside *mid band*) will change as the Capacitive reactance (Xc) of the capacitor changes with frequency.

Xc = -j1/(2 x Pi x F x C) where j represents a 90 degree phase shift and the calculations represent the numerical value of Xc. So, as far as phase shift is concerned, you get what your particular circuit gives you at different frequencies. And it changes with frequency because the *capacitive effect* of the capacitor changes with frequency. The actual phase shift which occurs depends on the ratio of R to C or C to R. Again, sorry for the mud. May I suggest you read up on RC circuits, particularly the mathematical effects which lead to the phase shift (due to the present of Capacitance or in other circuits the presence of Inductance). The math will get you into the Tangent function. If you are familiar with trig functions studying some RC circuits (attenuation and phase shift) from a mathematical perspective will clear up a lot of things for you.

As for the very low *gain* (in the mid-band) in your circuit there are two possible causes.

1. Op Amp 3 (is shorted out and therefore) does not exist, electrically. Res 8 = 500K is a very high series resistance. Op Amp 2 has a mid band gain of ONE, Rf = 10K and Rin = 10K. Op Amp 6 has a mid band gain of Res 15/ (Res 8 + Res 13 + Res14) or just under ONE.

2. The other is the frequency overlap (Midband range) of the Hi pass and Low pass sections. The pass frequencies of each must overlap where neither LPF or HPF have any appreciable attenuation - this is the MIDBAND. If the pass frequencies of the LPF and HPF don't overlap then there is no Midband region. The HP filter section should come first and the LP second and the cutoff frequencies of each adjusted so that signals in the desired bandpass are not attenuated by either filter.

Edit Note: The LPF cutoff frequency, Fc(LP), should determine the highest frequency desired and the HPF cutoff frequency, Fc(HP), should determine the lowest frequency desired.

Its fun to learn but you have many different things happening simultaneously that you have to sort out several different effects which are happening together.

Hope this helps and not confuses. If you have specific questions about my comments I'll try to further explain.

Edits to correct spelling and add a couple of sentences, hopefully to clarify.
Jess
 
Last edited:
Therion, I just *cursored* through the on-line manual for Filter Lab 2 by Microchip which gramo recommended (thanks gramo).

Its apparently available for free download. I highly recommend you at least read through the manual (video reference) as it graphically gives a good explanation of the design concept and appears to be an easy to use tool. You could quickly design your amplifier with it and then transfer the RC values into Edwin to calculate the simulation. It does not get into the mathematics appreciably but gives graphical tools which give an equivalent understanding. As you use it to design your amp, it checks things and may give you an error pop-up indicating a potential problem which you can quickly fix.

Really seems to be a neat tool. I'd like to play for it for a while, but Taxes must get done first. Think I'll download it anyway.

Thanks again gramo.

Jess
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top