Passive Analog filter inductors

[Bernhard]

I have reverse engineered and simulated a brickwall passive analog filter that has those coils with adjustable ferrite.

The coils have two windings with one common pin,
for example one measures 63,5mH & the other 0,5mH but in series the reading is 75mH


Which values can I trust ?
Simulation looks good.

Any problems to wind such coils without the ferrite ?

[hagtech]

You trust them both. Nothing wrong with your measurements. Inductance is proportional to turns-squared.

Try this, take two of these devices, wire up the 0.5m winding on one and the 63.5m winding on the other in series. Then they will add up to 64m.

jh

[Bernhard]

Thanks,

is it possible to wind a 65mH coil with reasonable size ?

[hagtech]

Yeah, it's a piece of cake. But at what level would you like it to saturate? It all depends on how much signal you want to pass. Only so much power can transfer through a given core size and material. Coil winding is an art, balancing many different requirements.

Reasonable size could mean it fits in the trunk of a car. If you want to wind your own, it'll probably take a lot of up front research first. That is, if you want decent performance for a particular application. I'd suggest looking at catalog parts from Mouser or DigiKey. Find something that operates well in the audio band. Or maybe look at crossover coils from Madisound.

jh

[Bernhard]

I modified a passive R-C-L-C-L-C-L-C brickwall filter in simulation to fit the values of the caps I have.

Before there were variations in the coil values like 50-58-64.
Is that necessary ?
Now I have something like 50-50-64

Also two caps have the same value.

But perfectly flat to 20kHz ( 0,05dB pp ripple ) and -60dB at 44,1 kHz.

[hagtech]

Yeah, but what is the time domain response of your filter?

jh

[Bernhard]

Even less ringing compared to the original values.
Simulated with square wave.

[hagtech]

How does the response compare to this filter?

http://www.hagtech.com/theory.html#hagerworth

jh

[Bernhard]

Your filter has a slight rolloff and is less steep, only about -40dB, but has a little less ringing.

My brickwall filter has a small coil in series to each of the 3 shunt caps in the middle.

Are the relations between the values of the shunt caps important ?

You have C1 = C4 & C2 = C3

[hagtech]

Are the relations between the values of the shunt caps important ?

It's a symmetrical filter, based on the Butterworth. Looks like with the extra coils you are running an elliptical or Cauer type. I would guess some transient issues. Don't do all your work in the frequency domain. Brick walls are notorious for time domain phase problems.

jh