Passive notch filter needed

[yesway]

Hi!

Would be helped by finding a simple (RC based) filter to blank out 600Hz (and later possibly add notches for 1kHz and 1k2Hz) in a design. At this point I have my eyes on tiwn t's or bridged t's, but would like some expertise of what would suit the situation best, I am sure there are other alternatives.
Think some 10-15dB attenuation at center f, and -3dB at +-100Hz would be a good start, if possible?

best regards

 

[Lewish]

Please indicate the number of watts you would like to pass thru the filter. Maybe I can help more.

 

[yesway]

Hi Lewish!

The filter is intended to go in the signal path between a cell phone and a small speaker amp, so the signal is about 50-100mVrms, the source impedance is low (10 ohms?), and the load impedance is high (a volume pot of about 10kohms).

 

[Lewish]

Hi yesway,

With that much impedance matching, i.e. 1000:1, it will be very difficult to design a passive filter that will be stable. It will have a strong tendency to ring (echo effect).
If you would like a reference, please consult "Electronic Filter Design Handbook" by A.B. Williams, chapter 6 - LC Band-Reject Filters.
I am curious as to how you have a problem signal at 600Hz in your cellphone. Do you know the source? Maybe you can do something to control the source and solve the problem.

 

[buzz41]

Pretty staightforward app. You are OK with the impedances. Ringing should not be a prob, particularly with a BW of 200 HZ. Of course, make sure there is no inadvertent feedback loops involved in the circuit you are inserting it into.

Try a bridged t and if the performance is adequate, you're done, else do a bridged t. Good luck. BTW, these passive RC filters work best with a stiff source and a light load.

 

[buzz41]

Where's the edit button? %-) It should be:
Try a bridged t and if the performance is adequate, you're done, else do a TWIN t.

Should also say that if you apply more that one in series, you will get a high insertion loss, so an emitter follower, etc. between filters might be needed. Relatively low source impedance and high load impedance is needed to keep insertion loss down. Classical impedance matching is not applicable.