Active XO with LM3886 as opamp

[xplod1236]

I'm building some 2 way speakers using Dayton drivers from Parts Express, which will be bi-amped. The idea here is to design a high/low pass filters with a gain of about 20, and use the LM3886 as the opamp and power amp in the filter. The drivers would be connected to the output of the filter (with built-in power amp). What are the pros/cons of doing it this way? One pro I can think of is less parts. Any comments and suggestions are welcome.

[MJL21193]

Hi,
I'm working with a similar concept here:
Active filter plus LM3886 - one board

Good idea for a two-way, as you can really simplify things by using the amp as the filter.
Good luck.

[theChris]

The LM3886 is not unity gain stable. you will need to chose a topology with high closed loop gain that tolerates component errors with the higher gain. This does add some design difficulty that isn't present for unity gain stable opamps.

[consort_ee_um]

The TDA2030A datasheet describes how to do active crossovers. You could substitute the LM3886 for the TDA devices on page 10 here:
http://www.st.com/stonline/books/pdf/docs/1459.pdf
Clearly the actual filter values would have to be tweaked to suit your particular drivers, but this satisfies Chris's criteria

[AndrewT]

Hi Thechris,
the 3886 is configured for normal power opamp gain.
The preceeding but integrated filters use the unity gain concept. I wish he hadn't chosen unity gain, due the limited flexibility and complex maths involved in changing Q values.
He is effectively emulating a plate sub-amp that takes wideband source material and filtering it to actively drive individual narrowbanded speakers.

[altor]

I've made "tri-phonic" system to my son some years ago, it use 1 LF channel and 2 MF+HF channels, but of course - it can be used with 2 LF speakers as "Bi-Amping".

Schematics: http://one.fsphost.com/altor/DIY_Hob...amplifier.html

text in russian, but schematics I hope are clear to understand.

I put 2 2nd order filters in serial, and the final stage - LM3886 based.

1st picture: volume control, LF-mixer, HighPass filters to sattelites.
2nd picture: phase invertor+phase shifter
3rd picture: Linkvitz transform and LowPass filter.
4th picture: 3886 output stage as the HighPass filter and High output impedance ("current source")
5th picture: 3886 output stage as the LowPass filter with negative output impedance.

Of course - you can use this 3886 filter schematics with the minor changes, if You don't want negative or high output impedance.

[MJL21193]

Originally posted by AndrewT
Hi Thechris,
the 3886 is configured for normal power opamp gain.
The preceeding but integrated filters use the unity gain concept. I wish he hadn't chosen unity gain, due the limited flexibility and complex maths involved in changing Q values.
He is effectively emulating a plate sub-amp that takes wideband source material and filtering it to actively drive individual narrowbanded speakers.

Andrew,
I think you are referring to my project, when you made the above remarks. TheChris was actually (I think) referring to xplod1236's plans to use the LM3886 as the filter.
My unity gain filter sections are 4th order Linkwitz-Riley, which have (almost) perfect phase-coherency, with no peaks or dips at the crossover frequency.
I have no reason to change the Q or the crossover frequencies, as I am well within the capabilities of my selected drivers.
The best part of my design is it's low cost - about $12.00 for each completed stand-alone filter/amp unit. With the cost this low, it becomes disposable, if a change is needed.