Active crossovers, worth the effort?

[everettT]

I'm a huge active crossover fan because of the many reasons listed above, but there are a lot more cables to deal with so keep that in mind.

Money might be his crux...

 

[Winkleswizard]

I have tried taking a passive system and making it active and I am very pleased with the results. Assuming you have the time and measurement equipment, the other benefit is being able to eq the system to the room. As mentioned, the wiring can be daunting as you add another set of line level and speaker wiring for every driver.

I have been contemplating on whether all the extra amps and wiring are worthwhile for a new design. There are situations where passive components are still useful. One is Zobel networks for impedance equalization and another high pass filters for tweeter protection. In active, it is easy to miswire and send the low signal to the tweeter and damage it. I also know that that my active setup is complex enough that nobody else in my house could fix it! As many of the benefits for active are in the lower octaves, this suggests that a active/passive hybrid may be a worthwhile compromise. This also would help offset the cost of additional amplifiers for a fully active system.

In our home theaters, many of us are already using active/ passive hybrids as our subwoofers are usually active and other speakers are passive. Due to the system complexity, I think this is about as “active” as we are likely to see for a while.

Has anyone seen any research that investigates the viability of active crossovers for midrange and higher frequency drivers? The arguments against large inductors and woofer voice coil heating issues seem pretty clear for the lower frequencies, but the main argument for higher frequency drivers seems to be mainly one of efficiency. With the cost of amplification dropping, am not sure that alone justifies using active crossovers...

Ww

 

[Verdinut]

I have tried taking a passive system and making it active and I am very pleased with the results. Assuming you have the time and measurement equipment, the other benefit is being able to eq the system to the room. As mentioned, the wiring can be daunting as you add another set of line level and speaker wiring for every driver.

I have been contemplating on whether all the extra amps and wiring are worthwhile for a new design. There are situations where passive components are still useful. One is Zobel networks for impedance equalization and another high pass filters for tweeter protection. In active, it is easy to miswire and send the low signal to the tweeter and damage it. I also know that that my active setup is complex enough that nobody else in my house could fix it! As many of the benefits for active are in the lower octaves, this suggests that a active/passive hybrid may be a worthwhile compromise. This also would help offset the cost of additional amplifiers for a fully active system.

In our home theaters, many of us are already using active/ passive hybrids as our subwoofers are usually active and other speakers are passive. Due to the system complexity, I think this is about as “active” as we are likely to see for a while.

Has anyone seen any research that investigates the viability of active crossovers for midrange and higher frequency drivers! The arguments against large inductors and woofer voice coil heating issues seem pretty clear for the lower frequencies, but the main argument for higher frequency drivers seems to be mainly one of efficiency. With the cost of amplification dropping, am not sure that alone justifies using active crossovers...

Ww

Although active bi-amplification has several advantages for splitting frequencies between a woofer or subwoofer and drivers which handle higher frequencies, I don't see much advantage in actively bi-amping to split the signal between the mid-range and tweeter frequencies.

I actively bi-amp my three front channel speakers using subs as woofers in same cabinets as the midrange drivers and tweeters. The bi-amping configuration is between subwoofers and midrange drivers. I designed the passive crossovers for use between midwoofers and tweeters. I am getting an excellent performance with that arrangement. See my Post #10 in this thread for additional comments.

Cheers,

 

[Winkleswizard]

Although active bi-amplification has several advantages for splitting frequencies between a woofer or subwoofer and drivers which handle higher frequencies, I don't see much advantage in actively bi-amping to split the signal between the mid-range and tweeter frequencies.

I actively bi-amp my three front channel speakers using subs as woofers in same cabinets as the midrange drivers and tweeters. The bi-amping configuration is between subwoofers and midrange drivers. I designed the passive crossovers for use between midwoofers and tweeters. I am getting an excellent performance with that arrangement. See my Post #10 in this thread for additional comments.

Cheers,

Thanks for pointing me to Elliott Sound site! Have visited there many times in the past, but he updated and actually had some discussion on active crossovers for mid and high frequency drivers. He pretty much verifies that the ROI for active crossovers is not as clear for much other than woofers.

I will add that if you have a driver with some ugly breakup, a dsp crossover is very good for creating custom notch filters to kill the peaking. It is much easier to nail precisely than with passive components .

 

[TLS Guy]

Thanks for pointing me to Elliott Sound site! Have visited there many times in the past, but he updated and actually had some discussion on active crossovers for mid and high frequency drivers. He pretty much verifies that the ROI for active crossovers is not as clear for much other than woofers.

I will add that if you have a driver with some ugly breakup, a dsp crossover is very good for creating custom notch filters to kill the peaking. It is much easier to nail precisely than with passive components .

That is true I think in the domestic situation, but not in high powered pro systems in large spaces. There is just too much loss and heat build up.

There are quite a few active speakers on the market now, especially in the pro monitor market that do have amps for each driver, including tweeters. My hunch is that once you put active crossovers in a speaker, it is cheaper to add a small class D amp now rather than passive components.

As to your last point about notch filters, there is one draw back that I ran into. The notch is removed ahead of amplification. If there is a peak in a driver, then any small amount of amplifier hiss is much increased. So I found I did have to put parallel notch filters in the speaker leads to get the S/N where it should be.

 

[Winkleswizard]

That is true I think in the domestic situation, but not in high powered pro systems in large spaces. There is just too much loss and heat build up.

There are quite a few active speakers on the market now, especially in the pro monitor market that do have amps for each driver, including tweeters. My hunch is that once you put active crossovers in a speaker, it is cheaper to add a small class D amp now rather than passive components.

As to your last point about notch filters, there is one draw back that I ran into. The notch is removed ahead of amplification. If there is a peak in a driver, then any small amount of amplifier hiss is much increased. So I found I did have to put parallel notch filters in the speaker leads to get the S/N where it should be.

Thanks for sharing, but I did not have hiss with my minidsp and emotiva amps. I will admit, I borrowed from Linkwitz’s LXmini design. I figured if the minidsp/emotiva combo was good enough for him, it should be good enough for my purposes.

My current implementation no longer requires a notch filter as I replaced the driver with one that has less breakup. This was another point made on the ESP site, but can say there really are limited driver choices with low breakup. The interesting takeaway was that this was also where he suggested that a 1st order filter might be suitable. Since this is the same as what is needed to minimally protect a tweeter in an active setup, this made me more comfortable with using passive crossover for it too.

Ww

 

[TLS Guy]

Thanks for sharing, but I did not have hiss with my minidsp and emotiva amps. I will admit, I borrowed from Linkwitz’s LXmini design. I figured if the minidsp/emotiva combo was good enough for him, it should be good enough for my purposes.

My current implementation no longer requires a notch filter as I replaced the driver with one that has less breakup. This was another point made on the ESP site, but can say there really are limited driver choices with low breakup. The interesting takeaway was that this was also where he suggested that a 1st order filter might be suitable. Since this is the same as what is needed to minimally protect a tweeter in an active setup, this made me more comfortable with using passive crossover for it too.

Ww

Let me tell you this makes a huge racket with even the quietest of electronics.

That has to be notched out after the power amp as any background noise is raised about 24 db!

 

[Winkleswizard]

While not pretty, not all that unusual. Depends on how much bandwidth you are trying to get out of the driver. If crossed low enough might not need a notch or less of one. If active, would be using 4th order filter.

Ww

 

[TLS Guy]

While not pretty, not all that unusual. Depends on how much bandwidth you are trying to get out of the driver. If crossed low enough might not need a notch or less of one. If active, would be using 4th order filter.

Ww

You have missed the point totally! That is a 10" bass driver. There are two in each speaker, one crossed at 60 Hz and the other fed a dual signal, the below 60 Hz and BSC which is first order then transitioning to second at 500 Hz. That is well below the 4K peak!

What you are missing is that any small amount of noise from the crossover output and amp is after the crossover. So that noise is NOT filtered by the crossover. So there is only one place that small amount of noise can be notched so you don't hear it from those drivers and that is the speaker leads. Just think about it!

 

[Winkleswizard]

You have missed the point totally! That is a 10" bass driver. There are two in each speaker, one crossed at 60 Hz and the other fed a dual signal, the below 60 Hz and BSC which is first order then transitioning to second at 500 Hz. That is well below the 4K peak!

What you are missing is that any small amount of noise from the crossover output and amp is after the crossover. So that noise is NOT filtered by the crossover. So there is only one place that small amount of noise can be notched so you don't hear it from those drivers and that is the speaker leads. Just think about it!

I was replying to your post with the driver graph. The racket I see there is driver breakup. If you are appyling a notch for some other purpose, then please share.

At this point, I have only a very limited picture of your system...

Ww

 

[TLS Guy]

I was replying to your post with the driver graph. The racket I see there is driver breakup. If you are appyling a notch for some other purpose, then please share.

At this point, I have only a very limited picture of your system...

Ww

You are still not getting it.

A lot of bass drivers have a break up mode like that. My point is that a notch in an active crossover does NOT obviate the need for one in the speaker lead.

The reason being is that amp and crossover noise is AFTER the notch in an active system. That means that without a passive notch in the speaker leads, the woofer break up modes effectively amplify the noise coming from the amp. I can assure you that you hear it especially during no program. The passive notch kills it.

 

[Winkleswizard]

Would agree there might be situations where a passive notch might get better results overall. I have been able to do so in my active designs with steeper crossover slopes or finding a driver with less breakup.

Ww

 

[yepimonfire]

That is true I think in the domestic situation, but not in high powered pro systems in large spaces. There is just too much loss and heat build up.

There are quite a few active speakers on the market now, especially in the pro monitor market that do have amps for each driver, including tweeters. My hunch is that once you put active crossovers in a speaker, it is cheaper to add a small class D amp now rather than passive components.

As to your last point about notch filters, there is one draw back that I ran into. The notch is removed ahead of amplification. If there is a peak in a driver, then any small amount of amplifier hiss is much increased. So I found I did have to put parallel notch filters in the speaker leads to get the S/N where it should be.

How much amplifier power is actually lost via a passive xover though? Enough to even add more than 1dB of sensitivity?

Sent from my LM-X210(G) using Tapatalk

 

[TLS Guy]

How much amplifier power is actually lost via a passive xover though? Enough to even add more than 1dB of sensitivity?

Sent from my LM-X210(G) using Tapatalk

The crossover in general wastes about half your amp power. It is generally reckoned to be about 1 db per order of the electrical filter. So a fourth order filter would take 4 db. However the lower the crossover point the more power is lost due to the massive series inductors. So a fourth order crossover at say 200 Hz might well loose you North of 6db.

Then the other issues is HF and midrange L-pads. Say you have a woofer 87 db sens and a tweeter 93 db. In the operating range of the tweeter you are loosing 6 db right there. However there is not that much power in the tweeter pass band.

However in a three way things are very different. Say you cross at 400 Hz and the mid is 3db more sensitive than the woofer. That would be usual. Say you have to pad for another 3db of band pass gain for a total of 6 db then you are loosing a lot of power.

In the last scenario an active design saves well over 50 % of the amp power.

That is another strike against passive three ways.

Domestically you have power to spare, but in large room applications these losses are serious and cause crossover failures by causing too much heating in the crossover components, to say nothing of the dynamic thermal compression from component heating.

That is why you no longer encounter passive crossovers in pro applications.

From my experience under domestic conditions when you do see crossover failure it is nearly always from a passive three way.

Hi-Fi sound tell me they see crossover failures in high powered three ways like B & W 800 series with significant frequency.

I think these issues are enough that these high end three way designs should have active crossovers on the woofer mid crossover point as a minimum.

For most domestic two ways this is not going to be a big factor, mainly because there is power to spare, and under most circumstances you do not get significant heating in a tweeter L-pad under domestic conditions. However I never specify anything less then 15 watt resistors even in tweeter L-pads.

 

[TLS Guy]

Would agree there might be situations where a passive notch might get better results overall. I have been able to do so in my active designs with steeper crossover slopes or finding a driver with less breakup.

Ww

These are the drivers in question. 7" and 10" SEAS Excel magnesium alloy cones.

These are very rigid cones, and so have excellent dispersion in their pass band. The downside of very rigid cones is abrupt and massive break up modes out of pass band which you have to deal with. The graph I showed was the 10" drivers.

 

[yepimonfire]

The crossover in general wastes about half your amp power. It is generally reckoned to be about 1 db per order of the electrical filter. So a fourth order filter would take 4 db. However the lower the crossover point the more power is lost due to the massive series inductors. So a fourth order crossover at say 200 Hz might well loose you North of 6db.

Then the other issues is HF and midrange L-pads. Say you have a woofer 87 db sens and a tweeter 93 db. In the operating range of the tweeter you are loosing 6 db right there. However there is not that much power in the tweeter pass band.

However in a three way things are very different. Say you cross at 400 Hz and the mid is 3db more sensitive than the woofer. That would be usual. Say you have to pad for another 3db of band pass gain for a total of 6 db then you are loosing a lot of power.

In the last scenario an active design saves well over 50 % of the amp power.

That is another strike against passive three ways.

Domestically you have power to spare, but in large room applications these losses are serious and cause crossover failures by causing too much heating in the crossover components, to say nothing of the dynamic thermal compression from component heating.

That is why you no longer encounter passive crossovers in pro applications.

From my experience under domestic conditions when you do see crossover failure it is nearly always from a passive three way.

Hi-Fi sound tell me they see crossover failures in high powered three ways like B & W 800 series with significant frequency.

I think these issues are enough that these high end three way designs should have active crossovers on the woofer mid crossover point as a minimum.

For most domestic two ways this is not going to be a big factor, mainly because there is power to spare, and under most circumstances you do not get significant heating in a tweeter L-pad under domestic conditions. However I never specify anything less then 15 watt resistors even in tweeter L-pads.

An lpad isn't going to dissipate all of the power it is attentuating though, would it?

Sent from my LM-X210(G) using Tapatalk

 

[TLS Guy]

An lpad isn't going to dissipate all of the power it is attentuating though, would it?

Sent from my LM-X210(G) using Tapatalk

No idea what point you are trying to make here.

 

[Winkleswizard]

Considering a relatively inexpensive minidsp is capable of performing all of the functions necessary to implement an active crossover, and REW makes measurements easy, outside of eliminating insertion losses, does an active crossover provide a significant benefit compared to a well designed passive one?

I've considered experimenting with it using my inexpensive R-15ms, I figure I could toy with using a slightly lower xover point (still within the operating range of the horn) but a steeper slope, such as an LR4, which is impractical to do passively. In addition, any sort or phase or time alignment problems can easily be compensated for in the minidsp. While the R-15m is hardly worth the trouble, it's inexpensive enough that blowing a tweeter during the process wouldn't be a big deal to me, and it'd give me some experience with it.

Anybody have experience with active crossovers, and do they actually provide a significant enough benefit to be worth the time involved in getting right and extra amplifier costs?

Sent from my LM-X210(G) using Tapatalk

Coming back around to your original post, I have a couple of other thoughts for you. If you have not already done, check out the existing passive crossovers. If they are high quailty and do not use large inductors, there may be less to gain on the SQ front going active.

If you do decide to proceed, knowing the original passive topology may help you more readily narrow down the active design. Also, as discussed, there may be parts of the passive crossovers that are worth keeping in place.

Better management of the woofer is where you have the potential for audible improvements in an active design. If you use these with a subwoofer, the improvement may be less, but hard to say until you know more. If you are still considering doing an active design, getting some measurements on the drivers without the crossovers would be worthwhile...

Ww

 

[highfigh]

An lpad isn't going to dissipate all of the power it is attentuating though, would it?

Sent from my LM-X210(G) using Tapatalk

Any energy that's not going to the tweeter will be wasted, as heat. That's the reason for 15W resistors.

 

[abrombo]

I have been built diy speaker systems with active crossovers since 1970 (the first one use 741 opamps). My current system is the diy version of the Linkwitz LX521.4 which was designed from the start using an active crossover. The LX521.4 and the Klipsch Jubilee are the two commercially available speaker systems that have active crossovers that I know off. I also built the CBT36K kit designed by Don Keele that also used an active crossover. The latest generation of dsp active crossovers (miniDSP Flex 8 for $600) are a significant improvement over what was available only 4 years ago. The active crossover puts much less strain on each amplifier channel driving it both with respect to power and bandwidth. The dsp crossover also allows you to compensate for irregularities in the individual driver responses. I have attached the plot of the dsp crossover used in the LX521.4 (note the commercial version uses an analog active crossover as opposed to digital). For the LX521.4 I am using 10 channels of amplification. Four channels at 130W per channel into 4 Ohms all channels driven for the woofers. Six channels at 125W per channel into 4 Ohms all channels driven for the lower and upper mid-ranges, and the tweeters. Having separate channels for each driver allows you to efficiently (without resistors) equalize the spl of each driver to account for differences in driver efficiency. Note in the frequency plot how much bass boost the woofers get in the LX521.4.