JFET Active Crossover

[merlin el mago]

Thanks Andrew, attached pic to be sure.

About 3 way: I have to make 2 HP + 1 LP right?

 

[AndrewT]

Remember to check your cascade of output offsets.
If it gets too high you may want to consider the Feucht added source resistor.

The Bass needs a 4pole Low Pass LR

The mid needs a 4 pole High Pass LR and a 4pole Low Pass LR. The Low Pass LR feeds the Mid amp. The output of the High Pass LR feeds the treble stage.

The treble needs a 4 pole High Pass LR.

This gives the basic LR4 3way. All the 2pole stages are Butterworth.

You may want to consider adding a High Pass for the Bass to protect against over excursion.
And consider a Low Pass for the treble to cut a bit more of the RF.
And attenuation+Buffer for the two most sensitive drivers.
And Baffle step for the Mid.
And some other EQs for any or all three drivers
And .....

 

[merlin el mago]

Oooops to much complicate, please Andrew could post schematics?

 

[merlin el mago]

Anybody can help me posting the schematic for the LP filter.

 

[AndrewT]

I've got some pics in the camera during today's daylight.

 

[merlin el mago]

TIA to post and help me.

 

[AndrewT]

Here is the HP sch and the 3way for an LR4
I have not implemented any EQ yet.
Will listen for a long while before I think about adding anything.

Note that with the HP, the input cap to each filter blocks the previous output offset.
Only in the LP & B do the offsets add up.
Thus the mid has 4 cascaded stages that should be checked for excessive output offset. The bass has 3 cascaded stages.

I took the power from the raw PSU +ve and 0V of the power amp. That's why I incorporated RCRC and used fairly high values of R (330||330) to drop the voltage.

 

[merlin el mago]

Thank you Andrew, (like we said in Spanish) "an image suits more than one thousand words". Could you post the LP filter? Attached your schematic with some doubts that I have? what are the bias voltage value? could I use other FETs? I guess isn't necessary to be matched, right?

 

[AndrewT]

The bias voltage is set with the 10k & 9k1. Smoothed with the 220uF.
I set it very slightly below 50%, to get a little further away from the 2nd harmonic that sets in with single ended followers at the extremes of input voltage.
You should be able to get a fairly clean 5Vac with 19V on the supply. This is well above (+6dB) the 2.2Vac to 2.4Vac that comes from highish output CDP. I don't know if any other Source is higher.

 

[merlin el mago]

Sorry my ignorance, 2R are connected to 10k 9k1 bias? Could you post schematic for low pass filter?

 

[AndrewT]

C, R and 2R (different from 2r2) are the Frequency and Q setting components.

You already have the Low Pass in post 121 !

 

[merlin el mago]

Thanks.

 

[merlin el mago]

How to check the cascade of output offsets?

 

[AndrewT]

Measure them. See how far from the 47.6% setting of the 9k1&10k they move away from.

 

[merlin el mago]

Of DC rail?

 

[AndrewT]

I usually measure from the Zero Volts line.
Sometimes I measure across a component.
In both cases it should be clear which is being used.
Your measurements should be just as unambiguous.

 

[merlin el mago]

Thanks for support.

 

[merlin el mago]

I have different amps for each way, how to balance to have the same gain in each way? Are the trimmers between the filters & buffers? If yes wich value and how measure to have the same gain in each way?

 

[merlin el mago]

Could you post the buffer schematic? Could I use others JFETs?

 

[AndrewT]

The buffer is the B in that first pic.
That buffer is included in both the schematics.
How much spoon feeding do you need?