Elliptical LPF question
It is infact two elliptical LPFs connected in series, for greater rejection.
Question 1. I wonder if I can wind the two inductors onto the same toroid, to save cores?
Question 2. If I wind them using air coils, what is the best orientation for the two coils (coaxial, side by side, etc)
A good elliptical filter won't use two identical sections, instead implement different real zeros.
The design equations expect uncoupled inductors. Any coupling will ruin the frequency response.
I have simulated it using rfsim99. It attenuates the first harmonic by more than 100db and the second by about 40db. The frequency in the passband is the desired one. Can it becave much different than the simulation in real life?
Also I would like to know the answers to the two questions.
Thank you
The orthogonal orientation is suited to minimize magnetic coupling among them.
Anyway, I would not expect a reasonable performance of an air-based inductor designed to have a calculated value at an 3rd order filter, notably because further interactions of its electromagnetic field with other stuffs on the board will happen.
I have just re-read the coupling information you give and this answers the question. Thanks.
So it has to be made using two separate toroids for best performance.
What do you mean by "implement different real zeros"?
To design the second filter so that it will have a notch at the second harmonic instead of being identical to the first one that has a notch at the first harmonic?
An elliptic filter is characterized by equiripple pass- and stop-bands (parameter passband ripple, minimal stopband attenuation and transition width). The zeros are generated according to this specification and not placed at arbitrary frequencies. But it can be meaningful in some cases to adjust the parameters so that a zero hits a specific interfering frequency. For elliptic filter properties, see https://en.wikipedia.org/wiki/Elliptic_filter
My point in post #2 was that cascading two equal filters doesn't follow the equiripple design method. Of course the resulting filter can be useful for your purposes, but it isn't an elliptic filter, strictly spoken.
I agree that low μr or air core ("material 0") torroids should give stable inductors with effectively no mutual coupling.
So you say I can use air core solenoids with success instead of toroids?
The emphasis is on low μr for stable inductance. 850 nH may be a bit too high for small air core, but still possible. Ferrite cores with air gap should be considered too.
you could do that, but the coils will be cross coupled, and you would have to add that to your analysis model. A LOT of recent filter papers actually USE such and effect to make dual band filters.
BTW, if your program predicted 100 dB of rejection.....you are not modelling your component Q's realistically