Active system converting to passive system

[dahlenmarcus]

Hello!


I have an active 3 way plus sub system.

The drivers are
2x Acoustic elegance ib 18 sub
2x acoustic elegance td 15 midbass
2x AudioTechnology 4H52 'FlexUnits
2x beyma tpl 200

Drivers firing upwards
2xVisaton ti100
2x b&c de250

Electronics are
Streamer pc
RME adi 2 dac
Behringer dcx2496 le
Behringer crossover for subs
5x Behringer A800


I crossover at the following frequencies
40hz
340hz
3300hz
LR-48

Many people say passive crossover sounds best and I want to find out for myself.

My plan is as follows
*Make a passive crossover for tweeter and midrange
*bypass active crossover for midrange and tweeter
*get a more advanced amplifier to run midrange and tweeter


I like all my drivers very much
The beyma tpl200 amt is very nice and derailed but it gives a harch sound if not crossed over very steep.
It also sound hard if crossed too low.

The flexunits is very smooth and nice, just like I want it.
The sensitivity differs 10db so it could be hard to design a crossover.
Impedance is 8ohm both drivers

My options so far is
*Buy a new midrange with higher sensitivity
*buy new tweeter, perhaps beryllium with low sensitivity
*buy another pair of flexunits midrange and parallel them to make an MTM in order to increase their spl.

I like the last option if it's possible but also fun to try something new.

What do you think is the best choice?

Best regards Marcus

 

[newvirus2008]

If I were you, I would let the Behringer do its job, instead of converting to passive, for more or less the same reasons anyone would list. However, opinions vary.

 

[dahlenmarcus]

newvirus2008

Yes the opinions differs alot but I think it would be nice to passive.

If the current setup is better I will have use for it anyway so why not try😌

 

[AllenB]

Do you need help with this or do you have it worked out?

 

[adason]

If you have mic and laptop, measure each speaker separately, design passive, lots of people here can help you with it, measure again, test, listen, see if you are happy with the sound, if not, tweak it, mod it, have fun with it, why not...

 

[AllenB]

I wouldn't use a mic to convert this if I were already happy with the filters.

 

[mark100]

I wouldn't use a mic to convert this if I were already happy with the filters.

Are you saying to just replicate the existing active filters with passive filters?

That is what i would do for an apples to apples comparison.
And would make a great test imo.....

 

[hifijim]

Are you saying to just replicate the existing active filters with passive filters?

Sure why not. Use the existing electrical signal response of the active filters as the target electrical filter response of the passive filters... Makes complete sense to me.

 

[newvirus2008]

Well, my opinion originated from the 48dB/oct slopes!

 

[hifijim]

I see your point. 48 dB/oct will be "a challenge" with a passive crossover circuit.

I would think that 4th order LR should be interchangeable with 8th order LR... ?

 

[AllenB]

48dB/Oct may mean a lot of group delay, and it shouldn't be necessary to go that steep if the issues are properly identified. So it may be worthwhile dialling back without losing the character of the response. I'm not seeing where this will best go yet, and how OP feels about it..

I assume we know what happens when we exactly copy filters like this to other methods.

 

[kevinkr]

I suggest measuring the impedance of all of those drivers across the frequency range you intend each to cover, it's going to be difficult to replicate the behavior of the active XO with passive if you do not know how the drivers will interact with your networks.

A 48dB/octave LR is going to be hard to replicate with passive - that's an 8th order network. Even a 4th order LR in passive is going to be a challenge.

Good luck! (I went the other way from passive 2nd order to active 4th order and DSP about 3 years ago, as Allen can attest there have been struggles.)

 

[Face]

Whatever you do, please don't rely on on generic online crossover calculators.

 

[plasnu]

The speakers with 48dB passive filter really exist?

 

[kevinkr]

It is possible if not exactly practical - I've never seen one in the wild, but that does not mean someone hasn't done it.

 

[AllenB]

Typically a second order filter is used to get a fourth order target with a real driver. Making eighth order filters is mostly academic.

You can use resonance to get a higher order fairly easily with only few components. The rebound does have an effect, drivers are usually cutting out there anyway, the more components the closer you get.

So somewhere in the middle you'll have yourself a high order filter.

 

[dahlenmarcus]

Do you need help with this or do you have it worked out?

Hello!

Yes I like to discuss the 3 options that I have

*Get new tweeters with low sensitivity

* get new midrange with higher sensitivity

*get 2 more flexunits 4 midrange, parallel to increase sensitivity in a mtm.


I like the last option but not sure how to do it right.

I need to design Avery steep filter
I don't know how much better sensitivity I can get by parallel , 8 dB?
That would be great but 6db would be enough because I like a little elevated tweeter.

Is this a good idea or not would be the first question?


Best regards Marcus

 

[dahlenmarcus]

I have some good advice on measuring and crossover designs from you, thank you.

I will have to look more into this today.

Best regards Marcus

 

[newvirus2008]

I need to design Avery steep filter

You would need to use filter design tables for higher order passive filters. If I remember correctly, this way you would first get a filter for 1Hz (or 1rad/sec) and then scale the L,C values to reach the target cutoff frequency accordingly. I think it's also possible to adapt a passive network to a different load impedance by scaling, if required.

 

[Jonasz]

2 parallell mids will give you 6dB higher voltage sensitivity, check so the impedance don't gets too low though.