Subwoofer filter with sensing LFE switch


2008-02-19 3:14 pm
For my little brother (he is literally quite small :)), I'm currently working on a subwoofer. Because he has some specific wishes, I'm making the filter myself. Before I start on the schematics, I would love your input on the block diagram. I'm for instance still contemplating if the relais is in the right spot; I can also but the input buffer for the L+R signal after the summing circuit and put the relais before the input buffers.

Filter Michiel Block Diagram.jpg

-Black arrows are signal, blue are control
-Speaker level inputs (L+R)
-Line level inputs (L+R)
-Summing circuit is an inverting opamp summing
-An automatic switch which senses if there is an LFE signal and with no signal present holds the relais for a certain amount of time (few minutes), relais can be released by pressing button

Cheers Maarten
Very similar to something I built for somebody a few years ago, but I chose to let the LFE signal run to all intents and purposes unfiltered as by it's nature it should be the correct bandwidth anyway. Then I just put the LFE/stereo select switch/relay before the EQ (using your diagram).

If that takes a bit of getting your head around, there is a block diagram for it on my website in the Bespoke section.
Last edited:


2008-02-19 3:14 pm
Thanks for the input Richie. That was also the plan to skip the LR filter for the LFE, but seeing my diagram again I see it's not correct (not having the most 'awake' day, as my mondays tend to be :rolleyes:).

The LR low pass should be directly after the summing block.

Thank you for pointing this out.


2008-02-19 3:14 pm
Finally finished the schematic. This is my first somewhat more elaborate putting together of building blocks. The different blocks I got from different desings / places.

Pleas ignore IC9B (not used right now) and the 2nd switch of the relais.

Still had some questions for you kind experts, if your willing to answer them:

-Are C204, C212, C222, necessary?
-Is the volume pot (VR3) in the correct place? (Before the output buffer)
-Is this the rorrect way to hook up VR3 or should I make a voltage divider (pot in series and a resistor to ground)?
-Is this the correct way to build a 0-180 phase control (are these the correct values for this frequency range)?
-Any other input and advice is welcome of course..

The schematic:

Filter Michiel Schematic.JPG

Thank you,

Cheers Maarten
1. No the caps are not needed.
2. VR3 is OK there but consider your max input level.
3. You can make a volume control that way but it's OK as you have it.
4. I think you have it almost correct, omit the feedback cap. As for values I suggest you simulate.

The jumpers across the tone controls will cause problems when you short the op-amps out.


2008-02-19 3:14 pm
Thanks again for taking the time too look at it Richie.

I modelled the phase control in LTSpice. This is the first time I used a modelling program, so I'm not sure I did it correct. But from my modelling it seems that in the frequency range of interest (20-80 Hz), a 220nF cap with a 100Kohm pot gives the best trade-off between amount of phase-shift and phase differences over the frequency range.

220nF with pot at 10kOhm:


220nF with pot at 100kOhm:


100nF with pot at 10kOhm:


100nF with pot at 100kOhm


  • AllPass100k100nF.JPG
    43.2 KB · Views: 168