diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Analog Line Level (http://www.diyaudio.com/forums/analog-line-level/)
-   -   Passive line level 3th order RC XO with no insertion losses? (http://www.diyaudio.com/forums/analog-line-level/220902-passive-line-level-3th-order-rc-xo-no-insertion-losses.html)

miksin 3rd October 2012 05:09 PM

Passive line level 3th order RC XO with no insertion losses?
 
1 Attachment(s)
how to implement one?
That's what I got and tried recently. It certainly works very well for me instead of active unit, between pre amp and power amps, but what is it I am missing as there are no insertion losses here. Shouldn't 3rd order passive RC cause 18 db attenuation?
4pcs 100n caps, two R's, that's it. ???

Davey 3rd October 2012 05:53 PM

The crossover in your photo looks like a first-order network. A third-order network would utilize many more components.

I think you're confusing two things. Insertion loss refers to the db signal loss across the series-connected components...in the pass band of the networks.
The "18db attenuation" refers to the roll-off rate (db/octave) outside of the pass band.

For a passive line-level crossover constructed with R's and C's it's possible to achieve nearly zero insertion loss in the high-pass portion.....because the capacitors are connected in series.
In the low-pass portion it's not possible because the resistors are connected in series and form a voltage divider with your power amp input resistance.....which yields the insertion loss.

It's still (sometimes) a viable approach because an attenuator could be used in the high-pass portion....or a power amp with lower voltage gain.....or relative speaker sensitivity difference.....or some combination of those.

Hope that helps.

Cheers,

Dave.

tinitus 3rd October 2012 06:45 PM

isnt the usual two components giving only 6db ?(instead of the expected 12db)
and why its almost impossible to do steep filter this way ?

I have thought about low cut 6db this way, and followed by a simple ordinary passive
wouldnt the low cut line level 6db be able to make up for a huge expencive inductor ?

but remember, you can still screw up speaker phase just like doing passive, if not done right

Davey 3rd October 2012 08:26 PM

Quote:

Originally Posted by tinitus (Post 3188947)
isnt the usual two components giving only 6db ?(instead of the expected 12db)
and why its almost impossible to do steep filter this way?

Yep, because a resistors action is not frequency dependent.

In a high-level crossover you'd have a series element (C or L) and a shunt element (C or L) to create a second-order filter. That won't work at line-level because the required inductor values go extremely high. That said, Marchand does offer the XM-46 passive line-level crossover which uses wound pot-core inductors of very high values and allows 4th-order slopes. The impedance levels get very low though. There's an existing thread here on diyaudio.com regarding the XM-46.

For passive line-level crossovers using R/C combinations, anything greater than second-order slopes aren't really practical.

Cheers,

Dave.

tinitus 3rd October 2012 08:31 PM

is there a lower limit for a 6db low pass filter (subwoofer) ?

miksin 4th October 2012 10:55 AM

Quote:

Originally Posted by Davey (Post 3188898)
The crossover in your photo looks like a first-order network. A third-order network would utilize many more components.
....

Just what I thought too and that's so confusing, because by the manufacturer states it is 3rd order.


Quote:

...
In the low-pass portion ... the resistors are connected in series....which yields the insertion loss.
...
I was referring just to assumed 3rd order low pass section.
I should find time to make some measurements. Unfortunately I have easy acces only to acoustical measurement equipment, which is of course a little stupid way to measure 80hz line level XO.:grumpy:
I could easily buy I could not make otherwise difference in low pass slope just by listening, that would be only expected, but I still expect at least 6db attenuation for first order at the low pass, and I dont find it. That is a little surprising.

DF96 4th October 2012 11:09 AM

You keep saying 'X dB attenuation' which either means you are confused, or you really mean 'X dB/octave attenuation in the stop band'. Which is it?

It is conceivable that a supplier might cheat and make use of the 2nd-order acoustic rolloff in the speaker itself. So then a 1st-order crossover with a 2nd-order speaker could be called a 3rd-order crossover. This would suggest that the supplier is incompetent or dishonest.

tinitus 4th October 2012 01:49 PM

Quote:

Originally Posted by miksin (Post 3189629)
.....but I still expect at least 6db attenuation for first order at the low pass, and I dont find it. That is a little surprising.

what is optimal amp load recommended by manufacturer ?
I suppose there would be one ?

but even if the 6db works, on its own it will probably only attenuate upper frequency
could be fun to ad a passive baxandall tone control :D


All times are GMT. The time now is 02:57 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2