DIY speaker builders: Why not go active? (long) (1 Viewer)

[Wayne A. Pflughaupt]

I’m sure all DIY speaker builders are familiar with the concept, “the (passive) crossover makes or breaks your DIY speaker.” I’ve always wondered, why even use a passive crossover at all? Why not go active?

A little background:

I had an interesting experience with my speakers a few years ago, a somewhat unique two-way design from the now-defunct Genesis Physics brand. The tweeter has a huge magnet which allows it to run a very low crossover point of about 1200Hz. However, it wasn’t hard high-level, sustained use to burn out the tweeters, and after blowing a couple they gave me a diagram showing how to raise the crossover point to 1800Hz.

Well, immediately I noticed that more than the crossover frequency had changed. The bass sounded different – a little tighter, and the tweeter sounded a little shrill. I eventually traced that problem to a 3-4dB spike centered at about 6kHz.

Another thing: Years earlier I came across a white paper from Harrison Labs (the company that makes those “F-Mod” in-line filters) that said if you took a passive three-way speaker system that was being powered at about 200 watts and switched it to active, you could get the same output levels with a trio of 25-watt amps! Yup – a total of 75 watts active would perform like 200 watts passive!

These two things were enough to convince me that passive crossovers are basically so much crap: They suck out amplifier power and audibly change the sonic characteristics of a driver.

You can see the difference an active system makes at this on-going thread at the Home Theater Talk Forum. The originator of the thread is experimenting with modifying an Adire speaker kit to active, and reports the improvements:

• “My friend just came over and we watched the "Brain Damage + Eclipse" music. The MPEG compressed music... was apparently clearer and more open than [with] the regular Kit281 in the next room hooked to my nice CD player and amplifier. I'm starting to become more convinced that the active approach itself is responsible for gains in sound quality that don't have to do with frequency response.”
• “...even considering the slight difference in the crossover, the active speaker still has the basic tone [compared to the un-modified one]... of the other one. It's still a Kit281, just better. It has clearer treble and smoother midrange. The disadvantage is this smoother midrange pushes certain singer's high notes back from a prominence in the soundstage. I appreciate the more pleasant sound and lower distortion, but it's a little less "unrestrained" and I'm going to try and correct it towards more upper midrange excitement, despite maybe a little harshness. It's a defect I don't mind living with.”
• “Each range of sounds is improved in some way. The bass has more "punch" and transparency, the midrange seems smoother (though it still has some of the characteristic AV8 coloration) and the treble is way clearer. In general the speakers maintain the same unrestrained quality I liked so much, but are less distorted sounding than before.”

With reports like this, I personally can’t understand why so few DIYers try active. It seems to me hands-down to be the best way to go. It takes all the mystery and “voodoo” out of coming up with a passive crossover design that works. All you need to do is tweak the crossover point and perhaps the slopes.

Sure, the cost of hardware is a factor, but all the money saved by going the DIY route would cover most of that outlay. So that shouldn’t even be a concern. Why not just bite the bullet and go all the way? It only makes sense, if the desire is to get the best performance for the money. As we’re already seen, you don’t need expensive high-powered amps; new or (even better) used high-quality, low-powered amps sell pretty cheap. Note that the DIYer from the HTT thread is only using an Audio Source amp, not exactly the vanguard for top-flight fidelity.

Once you have the hardware, you can use it forever with any future DIY speaker venture you might want to undertake. Each new project would be virtually “plug and play” – just change the crossover frequency as needed. No endless futzing around trying to come up with an all-new passive crossover design.

So let’s hear it, all you DIY speaker builders. Why do you choose to mess around with bogus passive crossovers when you could go active and leave all that behind?

Regards,
Wayne A. Pflughaupt

 

[Mike Keith]

You know my position on active; I'm of the same mind as you about this subject. I believe that there are so many advantages to active, especially for DIY. The dampening increase alone is worth the extra trouble and cost IMO.

Passive components cause so many anomalies in the high power signal path that often compensation networks are required to combat cretin driver’s Q’s. With active you hardly ever (if ever) need these compensation networks, so a strait electrical slope and conveniently adjustable amplitude is all you have to deal with in most DIY active applications. The fact that you can so easily change X-point and amplitude for each individual driver is a godsend if you want to quickly compare different alignments and fine tune to the room.

What Wayne said about

is what I can verify, when I changed my old passive DIY's to active I could play them to far greater SPL's with the same amps and not clip them, where before the distortion and lack of control of the drivers (decoupling from passive components) caused clipping prematurely. I had no idea that this little 2-way 6.5 woofer and 1" tweeter could have so much clean clear sound, they would play at concert levels now!

Nice post Wayne, and I couldn’t agree more with the notion of Active being the best DIY tool ever, and with Digital finally becoming more popular it's finally price worthy to consider an all digital Active network, save a bunch of preset curves and switch (A/B) quickly for a true comparison, this would be great to dial in a 2/3/4 way system.

 

[ThomasW]

I feel it's a disservice not to share this 'epiphony' with a larger cross section of the DIY loudspeaker community.

Therefore I urge Wayne to share his insights with the DIY-loudspeakers mailing list. Tom Bennett, David Harris, Earl Geddes, Dan Wiggins, David Hyre, Thomas Danley, etc. certainly need to know this information.

In addition I'm sure that the unwashed masses such as Tom Nousaine, John Atkinson, Richard D. Pierce, Bascom King, et al, that post at rec.audio.high-end would benefit from these insights as well.

And lest we forget the likes of Siegfried Linkwitz, Charles Hansen, John Whittaker, and others at the dipole-list, and of course DIY Audio loudspeaker forum.

 

[Allen Ross]

I was just flipping through the PE catalog i got in the mail and that Behringer ultradrive pro sure looks nice. But what about the things you get with a nicely designed OX, the ability to correct dispersion, baffle diffractions, or are all those things "chasing tale" with problems going with a passive network? I haven't really looked into it, but i imagine that it would be rater nice to have a nice active setup, but i imagine that most are XLR and so that would add another component into the mix.

 

[SethSpeaks]

F-Mods are passive line level, fixed frequency, fixed slope attenuators primarily intended for car audio.

The attenuation point is determined by a combination of the F-Mod device and the input impedance of the power amp. Use a amp with a different input impedence, and instantly get a different attenuation point and slope whether you want one or not.

The correct way to create passive line level crossovers is explained on this webpage. I can't post active links so add the appropriate stuff in the front of the link below
t-linespeakers.org/tech/filters/passiveHLxo.html
And it's a lot more complicated than the misinformation posted on the Harrison Lab website

So if you want to forget about bafflestep compensation. Forget about zobels. If you want to forget about combining differing roll-off slopes to match the performance characteristic of the specific drivers. Actually if you want to forget about everything that makes a good loudspeaker, by all means use F-Mods

Ciao!
Seth

 

[Wayne A. Pflughaupt]

Yeah, I don’t see F-mods as the way to go; as you’ve noted, passive devices will be as problematic on the front end as they are on the back end.

I’ve toyed with the idea of modding my Genesis speakers to active. I’ve already got my eye on a couple of amplifiers that would be ideal for the project. For one, Yamaha makes a dandy little low-powered 4-channel amp that sells pretty cheap on eBay. Adcom also used to make a low-powered 4-channel amp that would also work well. I could sell my hunkin’ 250-watt Adcom amps and have money left over.

The guy at HTT is braver and more patient than I am, making his own crossovers from scratch. I’d just use off-the-shelf hardware. I’ve toyed with a few ideas, like:

• Using AudioControl Phase Coupled Activators. They’re a 24dB/octave design with Linkwitz-Riley alignment, which would probably allow me to safely push the Genesis tweeter down to 1000Hz (!). The PCA is designed to be a sub crossover, but they use plug-in chips for the crossover frequency, so functionally you can choose any frequency you want. I’m currently using one in my system between the subs and mains and I’ve been very happy with the sonics. The downside is that it would take a second PCA to separate out the lows for the subs. Fortunately the PCA’s specs are so-o-o excellent that even passing the upper frequencies through two of them shouldn’t have any affect on sound quality. However…
• I could probably just as well use an AudioControl three-way car audio crossover to get everything in a single unit.
• Another piece of hardware I’d like to try, just for the sake of comparison, would be an Ashly pro-audio crossover. They also have 24dB/octave slopes, but with a unique feature, a “Response” control that allows you to tune the response curve between the filters. Basically, the “Response” control allows you to utilize either Linkwitz-Riley alignment or Butterworth, or a combination of the two. As a practical matter, the control lets you compensate for the droop in response you usually get between the filters – very cool.

Regards,
Wayne A. Pflughaupt

 

[Ethan Winer]

Wayne,

> These two things were enough to convince me that passive crossovers are basically so much crap: They suck out amplifier power and audibly change the sonic characteristics of a driver. <

You're preaching to the choir. My JBL 4430s have been biamped since I bought them more than ten years ago. When I switched from the passive crossover to an active (a Rane AC22) two weeks after I bought them, the improvement in clarity was immediately obvious. And these are high-end professional loudspeakers, presumably having as good a passive crossover as can be made.

--Ethan

 

[Michael R Price]

Hi, I'm the guy with the active Kit281s. I'm not really ready to present the results of this project, but once I've tweaked the system to be as good as it can be, I will try to make PCBs available so it is easier for other owners to do the upgrade. You can read the HTT thread, but don't worry... I will post the whole deal here when I get farther along.

As far as I can tell there is just as much "voodoo" in the active design. Computer software has made it easy to predict the complex summed frequency responses of passive crossovers just as well as active ones. It's too bad that with regular speakers drivers it is often not a good idea to use symmetrical, consistent slopes (which most off the shelf crossovers have).

 

[Chu Gai]

Did you do any sort of response measurements post-active in order to compare then with what your final version will be?

 

[ThomasW]

Really?

Wayne,

How many optimized high-end passive crossovers have you designed and built? And how many comparisons have you made between those same crossover and active ones custom designed and built for the same drivers?

Actually, for that matter how many crossovers of any kind, have you designed and built? Some how I get the feeling that all this just pure speculation on your part.

The use of generic active XO's, even those with 24dB/octave LR slopes pretty much ignores all known fundamentals for how loudspeaker should be designed. Advocating the use of any generic active XO that lacks the ability to compensate for things necessary for proper levels of loudspeaker performance (baffle step compensation, impedance compensation, etc, etc,) is folly.

As interesting aside to this entire discussion is that really high-end design Wilsons, Genesis, etc, primarily use passive XO's. Now the people spending $100,000+ for a pair of loudspeakers can certainy afford whatever it takes for optimal performance. And my guess is that designers such as David Wilson, et al, would use active electronics if there were any sonic benefit.

So one could conclude that not all passive XO's are crap.

 

[Saurav]

But what are some of the benefits of passive crossovers over active, other than cost and number of amplification channels? Or, let's put it another way - what can be done with a passive XO that cannot be done with an active XO? The only thing I can think of is a series XO, and even that might be achievable by some complicated custom amp circuitry. On the other hand, so many things are easier to do with an active XO. BSC. Notch filters. Much greater accuracy in picking filter/EQ frequencies and slopes.

Impedance compensation and zobels... is that really that big of an issue with an active XO? A 2 ohm rise in a driver's impedance is a big deal when it's surrounded by passive components with impedances in the 1 ohm to 20 ohm region. But if the driver is directly connected to an amplifier's output which has say a 0.1 ohm impedance, the inductive rise in the driver's response will cause a much smaller change in the frequency response (because the voltage division will stay pretty much the same, since the amp's output impedance is so low). I think zobels become much less critical with line level XOs because the driver's inductive rise causes a much smaller change in the FR. But I could be wrong, so please educate me if I am And even here, one could build the required curves with an active XO too.

The downsides are adding more components to the signal chain and noise (which any active component will add). Some people tend to be more concerned about that than other people.

Anyway... just musing. I have both passive and active XOs in my setup. The passive is between the most difficult pair of drivers, so right now I'm definitely in the "actives are easier" camp. If I had to do an active XO for the same driver pair, maybe I'd feel differently on this

And I so need to build myself a measurement rig...

 

[VinhT]

There is much more to a proper crossover than just filter slopes and tweeter attenuation. Sure, slope and attenuation may be all you have to work with when using a simple active crossover, but those two aspects are merely the tip of the iceberg when it comes to crossover design.

 

[Michael R Price]

I did a measurement of the initial active speaker and I have it somewhere, but it looks really ugly. It was important to note that I had a dip around 1.4KHz (larger than the one in the original Kit281) because a construction error reduced the Q of the tweeter filter, probably causing the "pushed back" voices I mentioned.

The ones I am using now are so similar to the original Kit281s in tone (frequency response) that for awhile, I used my stereo with one speaker active and one speaker passive, and they sounded coherent (with almost-correct soundstage focus).

This week I'll pull one speaker into my computer room and do some measurements/listen to music while tweaking the resistor values in the filters. It should be fun but I have no idea if the best measuring setup will sound the best.

 

[ThomasW]

Saurav,

I'm certainly not "anti" active XO's. I am however "anti" nonsensical statements that all passives are "so much crap", and that people should be throwing away their passives and then drop in generic actives.

I have a 1/2 dozen Marchand XM-9's used in 2 different systems in my house. And have used custom designed active XO's for more than 25yrs.

So yes I certainly agree, custom built active XO's are wonderful tools for optimizing driver performance. One needs to look no farther than the Linkwitz site to see what can be accomplished with an active XO design.

Michael has learned that creating a good custom design isn't a piece of cake. And he certainly couldn't have gotten the performance he needs out of his K281's, with a generic 24dB/octave L/R design.

As for a comprehensive discussion of active vs passive designs, I think that's a topic for a new thread, probably on a different forum.

 

[Mike Keith]

It's logical that there well designed Active and poorly designed Active networks, just as there are with Passives. But all things being equal IMO directly coupling the speaker to the amp output is of great advantage that no passive could have. From what I've heard the most simple passives networks sound best, which tells me that the less components between your speaker VC and the Amp out the better, if we could throw the connecting wire out it would be better still.

My Marchand XM-44 has the possibilities of making any corrections quite easily with plug in modules, which IMO are a superior tool in designing, especially if you only have your ears as a reference.


I agree with Saurav about compensation networks, Impedance spikes are less a problem with direct coupling, and the driver control is phenomenally better IMO. I use 2 Marchand (slightly customized) units for my whole HT Music system, and much prefer them to my old Passive setup, in fact there is no comparison, the active is far superior to my ears. There is a bit of grunge on the top end and in the noise floor with all op-amp based units but the advantages far outweigh the disadvantages for me. When something like the DEQX Active crossover/correction processor gets price worthy I think this will be the best yet, do everything in digital right up to the amplifier inputs.

 

[Michael R Price]

Mike,

I would argue that even those Marchand units are missing a bit of adjustability that makes it somewhat "hit or miss" trying to use them for crossovers at higher frequencies. With well behaved drivers, especially if you had a lot of overlap you could probably get away with using one, without much difference from independently/continuously adjustable Q and frequency. I can tell you right now that a crossover like that wouldn't work for the Kit281s which push "value" drivers to the limits of their performance.

 

[AlexKunec]

I just bought a m-audio revolution soundcard. I have preouts for 8 channels. I was just wondering, is there any software out there that can make my computer act as an active crossover? Cause that would be perfect if i could use 1 channel for my tweeter, 1 for midrange, etc.... and adjust xo freq, slopes using realtime software. You could test and change everything in the matter of seconds, and when you finalize everything, u can an analog version.

 

[Mike Keith]

Michael R:

I can’t think of any adjustability that any passive can do that this Marchand (or some others) can't.

Marchand XM-44-4
Time Delay.
Notch Filters.
Shelving Filters (BSC).
1st, 2nd, 3rd, 4th, and 48db slopes.
asymmetric slopes (tweeter at one slope, mid at another).
independent level controls for each driver.

The DEQX can do much more, but it’s pricy..

Now any conjugant (series) or impedance networks can't be done, but as I believe and stated earlier they are not normally needed with active.

In fairness of my limited experience; I believe there are problem drivers that just don’t work well with cretin topologies, and I know nothing about the drivers you mention, so I can only offer opinion based on what I've used.

 

[ThomasW]

Michael P.,

The Marchand XM-9's are a simple but high quality 24dB/octave L/R unit. I use them for subs and for a quad-amped hybrid ESL/Magnetic planar/Leaf array. They're fine for those.

Mike K's Marchand XM-44 is much more customizable, and has plug-in modules for virtually any type of compensation, EQ, or whatever.

That being said since one has the option of choosing all the internal components used, absolutely nothing works as well as a dedicated custom designed active XO. Most people just don't have your patience .... :wink:

Alex,

You're probably better off with a Behringer UltraDrive Pro 2496. Several very knowledgable people have auditioned these on pretty high-end systems, and found their sonics to be quite good.

Currently SoundEasy 9.0 has the ability to integrate XO models and control those models in a 2496. This concept with additional features is being incorporated into the next version of LspCAD Pro.

Finally there's a HUGE thread somewhere on AV Science where a PC is being used to control almost everything with regard to a speaker system. Sorry it was months, and months ago when I last looked at it and my computer/programing knowledge isn't up to what was being done .........

 

[Michael R Price]

Right. I guess if you have control of the Q, frequency and slope of each filter then you're close enough. Then the only difference is independently controlling the Q and frequency of each of the 2nd order "sections," which is only useful sometimes.

I wish to eventually do it all digital also, but that's expensive because of the multiple DACs (and also, presently, above my head).