Dialling in active crossovers

What exactly is the question?

It's not really difficult Mark for as long as adjustments for turnover frequency, slope and gain are all on the front panel. If you will still be using the passive mid and treble crossover, you can just knee the response at its mechanical roll off frequency to save the watts and adjust the output for the bass panel the same way one would on a subwoofer plate amp. Now if you'll need to use internal jumpers it can get inconvenient but the principles are exactly the same. If you look at the bright side however, even with jumpers, once you get things right, you shouldn't have to use them ever again.

Here is a (mediocre) pic of the inside the midrange bandpass control module of my PassLabs XVR1.

The photo shows about 60% of the acreage inside the case. The bottom 3/4 of the pic is the left channel. The right channel above, mostly out of the photo is the same. The other control module has the low pass to the woofers and the high pass to the tweeters

The build quility is superb and as Jack said, once you have it set. you're done.

While inconvenient I do think jumpers are the way to go in a fixed installation. Every so often potentiometers get gummed up and need to be "wiped" leading you to start all over again. Dirty pots can be a source a lot of degradation. I don't see this happening with jumpers.

I converted my speakers to active in stages. I am currently implementing a digital active crossover - so I have plenty of experience with crossovers.

But I have to second what Jack said - what is your question?

spirit

Before jumping in, let me be frank and ask you some questions. I will apologize if these are deemed improper and will gladly remove this post

Do you have an idea of what a crossover does? What slopes are and how they relate to drivers performance?
Are you very clear on why you want to go active?
Of the trade-offs involved?
Of the increase complexity of the system?
Of the serious cost involved in cabling active systems?
That you may need to measure your speakers ( I would recommend that you do)?

FrantzM said:

(...) That you may need to measure your speakers ( I would recommend that you do)?

Click to expand...

Besides measuring the size and the weight, I have real doubts that any of us knows and has the conditions to properly measure panel speakers ... Not to speak about interpreting the results!

microstrip said:

Besides measuring the size and the weight, I have real doubts that any of us knows and has the conditions to properly measure panel speakers ... Not to speak about interpreting the results!

Click to expand...

As opposed to measuring cone speakers ...???

So you'll go active and not measure the relative levels of the drivers? Interesting method... To each its own I suppose.

FrantzM said:

As opposed to measuring cone speakers ...???

So you'll go active and not measure the relative levels of the drivers? Interesting method... To each its own I suppose.

Click to expand...

Surely. Planars are not point sources and can not be measured as typical cones. Do you know why Magneplanars, Soundlabs and Apogees sound great and measure miserably in typical reviews?

I have told before - I am not against properly carried measurements when properly documented and used. I am against the simplified pseudo "measurements" that spread in audio forums, that are misleading, although they can be excellent placebos ...

Can I ask you how you would measure "the relative levels of the drivers"? How would you then correct for the different phases of the passive and active crossover? IMHO everything is simple when we just theorize on a subject. BTW, I have tried to emulate the crossover of the Soundlab's with the Behringer DSP crossover - I spent a lot of time on it and the result was a disaster.

microstrip

You try to fight battles when there is none to be fought... I'll entertain you.

He goes on and get to his amp and crossovers.. Do you suggest him to measure the level of the drivers by ears? Or would you suggest a method?
Come on man . Did I post anywhere that it was going to be easy?
Are you actually asking another person to embark in active crossover without at least finding a way, however flawed to measure the level of his speakers' drivers... No measurements? Really!? Because taking measurements is so difficult it is best not to take any? and that said with a straight face?

Since you know how. Just propose him a process, a method. That is what he's asking for. Not for you to prove to him that I am wrong.
Now the fact you, I repeat you had a disaster with the Behringer, is a proof that it can't be done???? Really?

FrantzM said:

microstrip

You try to fight battles when there is none to be fought... I'll entertain you.

He goes on and get to his amp and crossovers.. Do you suggest him to measure the level of the drivers by ears? Or would you suggest a method?
Come on man . Did I post anywhere that it was going to be easy?
Are you actually asking another person to embark in active crossover without at least finding a way, however flawed to measure the level of his speakers' drivers... No measurements? Really!? Because taking measurements is so difficult it is best not to take any? and that said with a straight face?

Since you know how. Just propose him a process, a method. That is what he's asking for. Not for you to prove to him that I am wrong.
Now the fact you, I repeat you had a disaster with the Behringer, is a proof that it can't be done???? Really?

Click to expand...

No Frantz, IMHO the responsibility of suggesting a valid method is yours. People systematically use the word measurements in audio with the same casualness they would suggest to use a meat thermometer to roast a turkey. I just warned that, unless you know what to do the casual measurement will be misleading.

And yes, most of the time it is better to have no measurement than a wrong and misleading measurement. IMHO this is one the cases.

I (and others better than me) explained long ago why the simple DSP crossovers with the basic functions can not emulate in practice real passive speakers. We even had long threads on converting passive designs to active designs in WBF some years ago.

And yes, I was only reporting my humble experience. It is easy to show a mathematical proof that the basic filter functions can not reproduce such complex waveform due to the interaction of the passive crossover with the capacitive load of the panels, as I have measured with REW, but as I was not proving anything, no one would learn anything from it and I have better things to do, I will just smile at your usual "rolleyes". BTW, I am really expecting that one day you will provide WBF with your great measurements, and we will then post about them ...

Considering Spiritofmusic future, using the knowledge I have got about him from many hundreds of great posts I have read, I would suggest : DO NOT go on and DO NOT get to the amp and crossovers...

What you are trying to do is simply convert from a passive to Bi-amp set-up. It's not as hard as it sounds on the surface. You need an active crossover that matches the passive network. All you need to do is model the passive network in a good crossover program like LEAP. Once you do that you take the voltage drive from the passive network and use that as your target curve for the active network. This can all be done in LEAP. You may have to tweak the existing passive high pass crossover but this can also be modeled for the values.

Once the curves match you are essentially done. You can measure the speakers to verify driver levels and phase. I have done this with excellent results using this method

Rob

Can I ask you how you would measure "the relative levels of the drivers"? How would you then correct for the different phases of the passive and active crossover?

Click to expand...

Hello micro

You need to be doing this in a good simulator. Read my previous post. It's really easy to measure the driver levels and there won't be any phase differences if you model it correctly and match the slope and driver polarity.

Rob

Robh3606 said:

What you are trying to do is simply convert from a passive to Bi-amp set-up. It's not as hard as it sounds on the surface. You need an active crossover that matches the passive network. All you need to do is model the passive network in a good crossover program like LEAP. Once you do that you take the voltage drive from the passive network and use that as your target curve for the active network. This can all be done in LEAP. You may have to tweak the existing passive high pass crossover but this can also be modeled for the values.

Once the curves match you are essentially done. You can measure the speakers to verify driver levels and phase. I have done this with excellent results using this method

Rob

Click to expand...

Rob,

Thanks for helping making my point - proper adequate tools and deep knowledge are needed. LEAP is an advanced CAE tool for speaker designing. LEAP is not exactly freeware and a some training is needed - my congratulations on using it. I have contacted it long ago, but never used it, when trying to measure Thiele-Small parameters of subwoofers to design a dipole subwoofer for Quads.

Robh3606 said:

Hello micro

You need to be doing this in a good simulator. Read my previous post. It's really easy to measure the driver levels and there won't be any phase differences if you model it correctly and match the slope and driver polarity.

Rob

Click to expand...

Yes, I know about it - the question was not aimed at experts!

I think that some of these replies don't help the OP. In my experience, using software like LEAP is not required.

Implementing crossovers is both easy and complicated. Grasping the basics is quite easy. But a proper implementation, understanding the effects of crossover slopes on phase, and attempting some advanced functions - that is difficult.

This is what I did when I converted my speakers to active:

1. Take measurements of your system (using freeware like REW) and decide what crossover point and slopes you are going to use. I take Microstrip's point that panel speakers are not as easy to measure as cone speakers - because of the large radiating surface and the fact that it is difficult to remove the room influence when measuring drivers - but it is still better than nothing.

2. Remove the passive crossover from the speaker. In my case - unscrew an access panel, cut off the wire, re-solder the cable from the driver directly to the speaker terminals.

3. Study the passive crossover and confirm that the values used in the passive x-over are exactly the same as what you measured. If you are in doubt what the capacitor and choke values mean, ask someone who knows. Also take note if there are other corrections within the x-over, e.g. notch filters.

4. Replicate the settings exactly with your active crossover. In my case, I was using a digital crossover as a development tool. Much easier to make changes this way. As a side note, digital crossovers can be quite inexpensive. You can pick up a MiniDSP for a couple of hundred bucks.

5. Repeat the measurement of your system and listen to hear if you like the sound.

Of course, you may not be aiming to remove the passive crossover from your system. If you wanted to, you COULD simply high pass (i.e. allow higher frequencies to pass) your mid/treble and use your valve amp, and low pass bass into your SS amp - and rely on the internal passive crossovers of your speakers to sort things out. There are both pros and cons to doing this. The main con is that you now have two crossovers in the signal path. This will really mess up your phase coherence at the crossover point. The main pro is that you have now relieved your valve amp of having to drive the bass. Whether you ultimately come out on top with such an approach ... you will just have to try and see.

The BEST way to use an active crossover is the way I have mentioned. Remove the passive crossover, and convert your speaker to active. Give each driver its own channel of amplification.

I won't pretend that I did it all myself. I had plenty of help from speaker designers, and I had to be prepared to shell out money for consultancy fees.

The way I would do it is as follows.

1) Remove the passive crossover completely...make all channels active.
2) Get a digital crossover (a cheap Behringer DCX24/96 or MiniDSP will do) or software based active crossover and the amps you want to use and begin to experiment with xover frequency, xover type (1st, 2nd, 4th, butterworth, bessel etc.) and relative driver levels (might prefer something different than was dialed in passively).
3) Once you have got what you think is optimal (hopefully you have measured a bit as well to make sure you didn't get something wacky) then you can talk to Marchand and get a nice tube crossover with the exact settings you determined from your experiements OR you can get a different and significantly more expensive digital crossover that has digital outputs. In the first, you will have a somewhat expensive but not excessively so analog tube active xover. The other will be fully digital so you can keep playing around...also with time alignment but you will need to buy at least two or three DACs to handle the digital outputs. Of course you want really good DACs, not what is built-in the digital xover so this will cost.
4) Hook it all up and listen to music and hope that you did step 2 correctly...of course if you go with the more expensive option of the Digital xover with digital outs and DACs you can readjust everything as you wish...this could also simply be step two if you think you really want the digital route.

Now the downside is that for analog lovers, the idea of digitizing everything is not a pleasant thought and then the analog active xover is a must. I must caution anyone doing analog xover route that all the ones I have heard are unacceptably coloring the sound with a SS "haze" overlay on the sound. The closest I had to a good one was the Accuphase F25 that I imported from Japan (fortuneately you can switch the voltage but it was a pain)...but it had a slight dampening effect on dynamics that I ultimately didn't like. The Bryston BP10 is also no good for ultimate listening. The Audio Research EC-2B was pretty good as I heard it on a friend's big Apogees but it is a tube crossover. We were also doing the part active/passive thing there.

You can do the digital xover with digital outs for relatively cheap from Mini-DSP but I don't know how well it works...haven't tried it. I know DBX makes a big studio model that has digital outs and costs about $3K.