I apologise for being a little off topic but yall are the only board I post to that may understand this kind of thing.
I'm putting together a project for school that deals with harmonics of certin types of waves. (square, triangle, sawtooth ect.) I need to design variable gain bandpass filters for 2kHz, 4kHz, 6kHz, 8kHz and 10kHz.
There are a few ways to do this and I've found one way to do this with only one Op-Amp for each harmonic instead of two. ( 
) The only problem is they dont give any formulas for setting the center frequency or Q.
I apologise for being a little off topic but yall are the only board I post to that may understand this kind of thing.
I'm putting together a project for school that deals with harmonics of certin types of waves. (square, triangle, sawtooth ect.) I need to design variable gain bandpass filters for 2kHz, 4kHz, 6kHz, 8kHz and 10kHz.
There are a few ways to do this and I've found one way to do this with only one Op-Amp for each harmonic instead of two. ( ) The only problem is they dont give any formulas for setting the center frequency or Q.
Anyone farmiliar with this design and know the formulas? Or might anyone have a diffrent suggestion? Also might anyone know a diffrent board that this question might be a little more appropriate for?
AM1650WITT.fortunecity.com
[email protected]
Hi,
I don't recognize that circuit in the link so I can"t be of much use. I will mention that it is difficult to get a stable filter with so few parts. They tend to ring and oscillate.
I think a much better approach will be to use the Sallen and Key circuit. This requires two stages cascaded, but the parts count is reasonable.
The bandpass is adjusted by setting the Fc's, one for the high cutoff and the other for the low cutoff. This is done by choosing capacitors and calculating the resistances, or if using pots, the cutoff frequencies can be variable.
Floyd's "Electronic Devices" textbook covers this topic quite well. It is very often used in Tech Schools and you will probably find a copy in the Library or on a Professor's bookshelf.
Though I have not tried it, a web search on "Sallen and Key" , "Butterworth", "Chebyshev", or "active filters" will probably provide something you can use.
Neil
The most important part here is that the gain has to be variable. With a sallen key filter I would need a 3 rd op amp in order to do that. (I think)
I got alot of good ideas. Problem is, most of em suck.
George Carlin (WINO)
One op amp will have to be dedicated to setting the gain, one for the lowpass stage and one for the highpass stage. You are right, three will be needed.
Start by deciding what order (determines the roll off past cutoff) and what passband response you want. Chebyshev would be a good choice for your situation since it gives a pretty high roll off. Butterworth is more common and it might be easier to find the gain tables for this. Then select the gain for each stage from a reference table which you will find in most texts on the subject.
Generally, active filter gain is what is known as "realizable to a constant". This means that once the order and response (Butterworth, etc.) are set, the gain is fixed to a particular value and cannot be changed because it is the only gain which will give the desired response. The designer has to accept this. The common solution is to follow the filter with a buffer amplifier and use that to set the gain.
For example, a filter may give a gain of 1.7 and you want to adjust this to a gain of 4. The buffer gain would be 4/1.7 = 2.35 and the overall gain is now 1.7x2.35 = 4 in the passband. You could use a buffer amplifier with variable gain to set the overall gain to whatever you want.
This sounds complicated because it is. That's why people spend a long time learning how to design filters. I regret that I cannot just point you to a circuit that will work for you, but if this is a class assignment it was designed to provide a learning experience and you will need to study the topic. I taught this subject and it required 20 hours of lecture and lab time for my students to understand the topic so they could do what you want to do. Don't give up on this, just be willing to spend some time on it.
I hope my suggestons are helpful.
Neil
Good stuff there neil, one way I was thinking of going however was with a multiple feedback bandpass filter (pictured here
http://www.ecircuitcenter.com/Circuits/MFB_bandpass/MFB_bandpass.htm) then another op amp for variable gain in order to only use two op amps. O well...anyone know a diffrent place I should go for this type of discussion.
I got alot of good ideas. Problem is, most of em suck.
George Carlin (WINO)
am1650witt.fortunecity.com
[email protected]
Hi,
Hope what I posted helped. You are on the right track. The link you gave did not open so I couldn't see the circuit, but I know of multiple feedback circuits. It will probably work for you, just be aware that they can be unstable. There is a thing called S Q of K which measures the stability of Q for filters as the gain changes. When transistors were used as the amplifier, the gain would change due to current and temperature effects and this was a concern. The multifeedback filters did not perform well according to this measure. With op amps it is not as much of a concern since the gain is determined by resistors and is very stable.
This filter will have a fixed gain so use the second op amp to set the overall gain.
I would like to hear how this project works.
Neil
here is a good site for general electronics info....
http://www.twysted-pair.com/cgi-bin/eboard40/index.cgi