About a discrete active crossover

[gaetan8888]

Hello

I want to made a discrete active crossover, here is low pass possible circuits and in the next message the high pass possible circuits.

Do it need a nfb ?

What is the formula to calculate the value of L1 L2 C1 for the low pass and C1 C2 L1 for the high pass, for a second order bessel 2khz low pass and high pass filters ?

Thank

Gaetan

[gaetan8888]

Hello

Here is the high pass circuit;

Gaetan

[Svante]

Well, apart from giving not-so-good distortion data, will the circuit work att all? I mean should it not be a DC path to the input of the first FET? (It was ages since I built something with FETs, so I'm not sure about this).

[phase_accurate]

Yes, there are missing DC paths. Apart from that I'd use RC filters with positive feedback and the FET stages would have to be source-followers.

Regards

Charles

[gaetan8888]

Hello

I was think to use op-amps like THS4032, LM6172 or AD625, but it may not sound as good as a discrete circuit.

With op-amp for the low pass and high pass, I would do a 2 nd order modified Bessel filter.

Thank

Gaetan

[analog_sa]

Originally posted by gaetan8888
Hello

I was think to use op-amps like THS4032, LM6172 or AD625, but it may not sound as good as a discrete circuit.

Maybe not but it will certainly sound better than your proposed circuit. Unless you really adore lots of second harmonics.

Unless you include some gate biasing network your circuit won't work anyway.

[Svante]

Originally posted by gaetan8888
I was think to use op-amps like THS4032, LM6172 or AD625, but it may not sound as good as a discrete circuit.

I'd say it is the other way around, unless you like distortion.

I remember a test in a Swedish magazine some 20 years ago, where they tested a ratsnest of 6 OPamps (TL074 and NE5534), "poor" capacitors, 12 metres of cheap signal cable and more. They compared this against a "straight wire" and were very careful with level matching throughout the audible range.

Several audiophiles got the opportunity to blind test this in their own equipment.

None of them managed to detect it.

Now, this was 20 years ago, still in the vinyl era, but ever since I read that test I am sceptical to statements that OPamps "sound bad". I think the test would give the same result today.

[MJL21193]

Originally posted by gaetan8888

I was think to use op-amps like THS4032, LM6172 or AD625, but it may not sound as good as a discrete circuit.

That's a completely false assumption. Op-amp filter implementation is not only easier, but usually much better.

Originally posted by Svante


I'd say it is the other way around, unless you like distortion.

The bias against op-amps is more audiophoolery.

[phase_accurate]

Now, this was 20 years ago, still in the vinyl era, but ever since I read that test I am sceptical to statements that OPamps "sound bad". I think the test would give the same result today.

Most probably.
And then there are OP-AMPs available today that are even much closer to the "wire with gain".

I think one should not overdo it and use masses of them.

Regards

Charles

[Keruskerfuerst]

You also could use BB OPA143PA.

Be aware, that the input impedance should not exceed 2kOhm, otherwise distortion will rise due to the non-constant input capacity of the FET input stage.