Active Crossover Activity thread

[Bricolo]

I send it with the forum's interface, with the email button under your posts

Is this the good adress?

[Fred Dieckmann]

I found a great set of links for active filter design. I have not checked them all out but will report further after looking at some of them more closely

http://www.newwaveinstruments.com/re...t_software.htm


I also recommend the Active Filter Design Techniques SLOA088.pdf at
http://www-s.ti.com/sc/psheets/sloa088/sloa088.pdf


As an introduction to Sallen Key filters:
http://www-k.ext.ti.com/SRVS/CGI-BIN...Sxi=9,Case=obj(26896)

http://www-s.ti.com/sc/psheets/sloa024b/sloa024b.pdf


Also I forgot the picture of Sallen and Key

[peranders]

Originally posted by till
Thank you Fred,

I don´t have spice software, but i see it would be good to learn using it.

You can download SwitcherCAD from Linear Tech for free.

http://www.linear.com/new/design_tools.html
http://www.linear.com/software/

[Fred Dieckmann]

As our fearless leader Mr. Pass said.....

"It certainly is true that the brighter of our readership can reverse engineer an XVR1 from the owner's manual, which is
online at www.passlabs.com

Go ahead, I dare you......."

The attached schematic is pretty much what I believe to be the basic topology of the resistor matrix resistor values and capacitor values for the basic 2nd order filter section. The Q switching appears to be a bit simpler on the passlabs schematic but this was my best shot for now. I cranked up my Spice resolution and think this is real close. The Q's were figured out from the gain specs on page 29 of the manual and the Spice plots compared against the curves on page 34. This was a real pain in the butt to figure out and I can't imagine doing this without a good Spice program and lot of spare time. It was very educational and a good exercise for the Grey Matter to figure out the resistor and capacitor values from the frequency steps and Rpack values in the drawing. The layout of the jumpers with respect to the Rpaks was also a useful clue since the order of the frequency steps appears to minimize trace length to the Rpacks. This is an extremely good PCB layout and I am curious as to the number of layers as I expect it is probably 4 layers.

Do I win anything besides the inevitable lecture about where I went wrong, Nelson?

[planet10]

An article on Sallen-Key at my site

http://www.t-linespeakers.org/tech/f...allen-Key.html

and a set of links to active Xo info

http://www.t-linespeakers.org/linx/xolinks.html

dave

[moamps]

Originally posted by till
High pass should be first or second order at ca. 700Hz.
I want the crossover for the high freqency speaker to provide a 6dB/okt ramp more signal from 10kHz upwards for compensation of the high frequency horns roll of.
I want it to provide about 4dB more between 700Hz or 800Hz and 5,5kHz than beween 5,5 and 10kHz.

Hi,

one idea (simulation only)

Regards

[Fred Dieckmann]

What you have posted is a bandpass filter.
It does not meet the requirements of either the high pass or low pass fillters, which are actually two filters. Back to the drawing board...........

[Fred Dieckmann]

Originally posted by planet10
An article on Sallen-Key at my site

http://www.t-linespeakers.org/tech/f...allen-Key.html

and a set of links to active Xo info

http://www.t-linespeakers.org/linx/xolinks.html

dave

Thanks Dave. Maybe we can work on some Mathcad models for some of the standard filter types. I have not used it in years and need to read up.

[moamps]

Originally posted by Fred Dieckmann
What you have posted is a bandpass filter.
It does not meet the requirements of either the high pass or low pass fillters, which are actually two filters. Back to the drawing board...........

Talk to me?????

Till says:

"High pass should be first or second order at ca. 700Hz.
I want the crossover for the high freqency speaker to provide a 6dB/okt ramp more signal from 10kHz upwards for compensation of the high frequency horns roll of.
I want it to provide about 4dB more between 700Hz or 800Hz and 5,5kHz than beween 5,5 and 10kHz"

and

"The circuit below sorts out the high end roll of of my horn. About the different response in 700<f<5500 and 5500<f<10000: is it possible to make something like a band pass for the lower intervall and parallel it with a padded down high pass 5,5kHz ? Behind this 5,5kHz highpass the correction circuit for the roll of?"

My circuit is good approximation of desired fr.response in TIll's first post, nothing else. IMO

[Fred Dieckmann]

"a band pass for the lower intervall (700<f<5500) "


"parallel it with a padded down high pass 5,5kHz ? Behind this 5.5kHz highpass the correction circuit for the roll of?"


For the low interval I think you want a high pass at 700Hz followed by a low pass at 5500Hz.

The upper range is a high pass at 5.5KHz followed by EQ for a boost in amplitude for frequencies over 10KHz.

I posted only the shelf EQ. The point off the thread was to investigate adjustable Q and frequency crossovers. I am not going to design specific crossovers for people as it would be immpossible without knowing the driver responses. I can not tell what your bandpass does with no units for the amplitude (y axis) unless I simulated it.
This is not a hand holding thread and reading some of the references is required to follow the thread.