Active Crossovers

[ben.]

What does the Jub/Synergy comparison have to with it?

You said synergy horns are always better. I provided a case in the real world of a basic horn speaker that outperforms what is available in the synergy horn world. I brought up the footprint to acknowledge that they have two totally different design goals in mind and I would be curious to see what Danley could design in the footprint of the khorn since I think that's a good sweetspot for aesthetics.

You can save your energy on the line array data...I've probably already seen what you intend to show, and I'm not at all a fan of the line array approach anyway. You've missed my point twice about why I even brought them up in the first place.

What I said was that, in a comparison between "Classic" line arrays, conventional horn loaded, and Synergy horn loaded line arrays are better. I know that you know a Jubilee is not a line array, so I guess you should slow down and read more carefully with regard to context. Once again, Synergy horns were brought up only to illustrate the imperfection of electronic time alignment when implemented in designs where acoustic sources are more seperated.

Here it is, copied from a few pages back:
"Classic" line arrays using direct radiators are inferior to horn loaded designs since that behavior is strictly linked to array length. Off axis response is a mess as well. Horn loaded designs are better. Horn loaded designs utilizing Synergy technology are even better. google around and you might start to see where I am coming from. Hop a plane to Buffalo and I'll prove it to you and buy you dinner afterwards.

Please tell me why you brought up line arrays if I have indeed missed your point. I thought it was in an attempt to prove that sound waves can be joined or melded together.

[DrWho]

but the simple fact that you can change the angles between cabinets to change that behavior chould prove (even to an engineer) that the sound is not magically glued together.

If you violate the criteria for being able to treat it like a plane wave, then of course you shouldn't call it a plane wave. Demonstrating that you can violate the criteria is meaningless. I could create a huge long list of behaviors engineers refer to with "wrong terms" because it makes it easier to communicate what is happening, or what is important. There is always some layer of abstraction going on and you have to make sure you're operating within the same abstraction.

Again, put down the books, Mike.

Ok, I gotta get this off my chest...

First of all, discarding theory is not wisdom. True wisdom is knowing when the theory is valid and knowing how to compensate when it isn't. If there's anything I've learned in my short year and a half outta college is that the theory is far more valid than I ever thought it was...and way more valid than all the "real world" comments thrown around on the audio forums. In fact, I've been amazed by how the good engineers will be able to know all performance criteria ahead of time easily to within a dB or a few Hz (at least in the electronics world). I'd wager the speaker guys are within a couple dB of their predictions...in fact, at work I've already experienced tweaking in circuits to their theoretical maximums. Of course nothing worked entirely as expected right away, but it had to do with things outside the assumptions of the theory itself....so when those other things were addressed, the theory was hit exactly. If anything, my doubting of theory has become a thorn in my side and I've found myself brushing up on all the textbook equations because they actually work extremely well.

Second of all, nothing that I have brought up in this thread (except for my comments about assumptions behind the acoustics wave model) has had anything to do with "books". My comments on time-alignment have all stemmed from personal experience and many years of measuring more systems than I care to count. I'm still learning everyday and keep finding new gotchas and tricks, but it's incredibly frustrating being demeaned as one that lives only in textbooks when I'm trying to share real world experiences. But hey, I get the young guy card all the time so I'm kinda used to it...

[ben.]

but the simple fact that you can change the angles between cabinets to change that behavior chould prove (even to an engineer) that the sound is not magically glued together.

If you violate the criteria for being able to treat it like a plane wave, then of course you shouldn't call it a plane wave. Demonstrating that you can violate the criteria is meaningless. I could create a huge long list of behaviors engineers refer to with "wrong terms" because it makes it easier to communicate what is happening, or what is important. There is always some layer of abstraction going on and you have to make sure you're operating within the same abstraction.

Again, put down the books, Mike.

Ok, I gotta get this off my chest... d

First of all, discarding theory is not wisdom. True wisdom is knowing when the theory is valid and knowing how to compensate when it isn't. If there's anything I've learned in my short year and a half outta college is that the theory is far more valid than I ever thought it was...and way more valid than all the "real world" comments thrown around on the audio forums. In fact, I've been amazed by how the good engineers will be able to know all performance criteria ahead of time easily to within a dB or a few Hz (at least in the electronics world). I'd wager the speaker guys are within a couple dB of their predictions...in fact, at work I've already experienced tweaking in circuits to their theoretical maximums. Of course nothing worked entirely as expected right away, but it had to do with things outside the assumptions of the theory itself....so when those other things were addressed, the theory was hit exactly. If anything, my doubting of theory has become a thorn in my side and I've found myself brushing up on all the textbook equations because they actually work extremely well.

Second of all, nothing that I have brought up in this thread (except for my comments about assumptions behind the acoustics wave model) has had anything to do with "books". My comments on time-alignment have all stemmed from personal experience and many years of measuring more systems than I care to count. I'm still learning everyday and keep finding new gotchas and tricks, but it's incredibly frustrating being demeaned as one that lives only in textbooks when I'm trying to share real world experiences. But hey, I get the young guy card all the time so I'm kinda used to it...

Those comments had nothing to do with the value of theoretical thought or discourse. I say it because you don't seem to be able to communicate in a direct fashion. Still plenty of points in my argument you're ignoring...

Yes you can treat the output of whatever line array in certain conditions as a planar wave. You still seem to overlook the original context, which concerned whether or not sound waves can be "joined or melded together". Electronic time alignment is not a cure-all, because sound waves cannot be "joined or melded together". Over emphasizing its importance is a mistake. My arguments have only been to prove that point. If one could "meld' or "join" sound waves together, electronic time aligment would still work outside of the few points that are on the plane of points where the differences in source-listener distance remain equal. You can't and it doesn't. See, I'm trying to help people understand what's happening and what isn't. What are you trying to do?

I'm sorry if you're frustrated because you feel demeaned. I feel frustrated because you're obviously a smart guy but seem unable to communicate effectively in a reasoned argument. This stuff isn't in the shadowy realm of opinion and subjectivity that we normally deal with here. Maybe that's the problem...

[Rudy81]

Rudy81

One suggestion I've seen is choose a Fc one octave below your crossover frequency. (ie: 400hz crossover then choose 200hz to calculate your capacitor). You want your capacitor's impedance Xc to = your drivers impedance at 200hz in the above example.

It's assumed that you will also know the impedance of your driver around Fc. Since a drivers impedance varies with frequency unless you know it I guess your next best bet would be to use the nominal rating (ie: 8ohm, 16ohm...etc). This also assumes you are using amplifiers with low output impedances like solid state amplifiers typically have.

So you can use the calculator you found above until you narrow down the value you need.

If you intend to vary the crossover frequency by any substantial amount say between the Mid-Tweeter then you might want to pick up several different cap values for approximately the frequencies that you might want to try out.

This is the formula to calculate the capacitor value if you want it: C = 1,000,000 / (2pi) x (Fc) x (Xc)

So for example if you want a capacitor value for an Fc = 250hz and your driver impedance is 8ohm then:

C = 1,000,000 / (2 x 3.14159) x 250hz x (8 ohm)

C= 1,000,000 / (6.28318) x (250) x (8)

C = 1,000,000 / 12566.36

C = 79.577 mfd capacitor

Chose a standard value close ( ie: 80mfd ) to the calculated value because it's not that critical due to many variables.

Maybe someone with more experience will chime with any other protection methods that might be better but for expermenting this will definitly get you going.

mike tn

Mike: Thank you. I guess I'm on the right track. I'm glad I figured out that the capacitor chosen is not some fixed value someone suggested, but a variable depending on crossover etc. I see that most folks don't use any protective capacitors. I have been concerned since I am new to this type of hookup.

I use the amps, Parsound HALO A23 and A52 solid state amps, on a regular basis and they are totally quiet during the turn on procedure. I plan on using the EV DC-ONE via balanced connectors to these amps. I imagine there won't be too much danger of blowing drivers, but I was trying to be extra careful.

BTW, I have been studying the manual for the EV DC-ONE, and it seems like quite a capable and flexible unit for a 3-way application. I have also been reading up on active crossovers. I was unaware of the many advantages to using an active crossover.

Mike, do the capacitors need to be of any particular type or brand? Thanks again.

[jwc]

Miketn and Rudy.

I'm following the calculation of a capacitor in series with the mid/high....which is essentially like calculating a "High pass" but you are choosing a lower octave...I've read this explanation before.....such as if you want to biamp using a tube amp and use the internal passive Crossover in a speaker. Then you would put a cap as you say with a calculation one octave lower...

But why are you two talking about doing this behind an active crossover "for protection" to the drivers...please explain..

thanks

jc

[cfelliot]

2 Part article on bi-amping/tri-amping by Rod Elliott:

http://sound.westhost.com/bi-amp.htm

http://sound.westhost.com/bi-amp2.htm

[Rudy81]

JC: As you have gathered, I am new to active crossovers and have been reading as much as I can on the subject. Several articles, or posts, I have read indicate that it is prudent to 'protect' the mid and hi drivers from ampifier 'turn on thumps' by adding a capacitor just before the driver.

Now, I have pretty nice amps, and don't expect that to be an issue, but you never know. So, I had been asking about how to properly select said capacitors. It is purely a way to protect my drivers while I experiment and connect them directly to my amps.

Being new to the concept, I do not know if this is necessary, prudent, or advisable. I was looking for suggestions.

[jwc]

I see. Read some of that explanation in the above link provided. I've never given this consideration.

I have never put a cap in there. Used an Ashly, DBX and EV active XO in the past....so far...nothing blown up that I know of.

Hmm

jc

[Rudy81]

JC:

I would think that in the interest of 'sound purity' one would avoid putting anything in the way of the driver. I am guessing you may give up some of the advantages of going active, such as more direct amplifier control of the driver. But, since I try to anticipate what could go wrong, I looked into it. The trade off now is the cost of various capacitors just to try various frequency points. I'm sure putting a capacitor in line with the driver is not 'free' of problems. That's why I was asking about the process.

I may just wing it and try it without anything until I know if I am going to stay active or go back to passive and until I gather more information on the topic. I figured the EE types around here could shed some light on the practice.

[ClaudeJ1]

Miketn and Rudy.

I'm following the calculation of a capacitor in series with the mid/high....which is essentially like calculating a "High pass" but you are choosing a lower octave...I've read this explanation before.....such as if you want to biamp using a tube amp and use the internal passive Crossover in a speaker. Then you would put a cap as you say with a calculation one octave lower...

But why are you two talking about doing this behind an active crossover "for protection" to the drivers...please explain..

thanks

jc

I rember PWK, in the days of Mac 275's and Marantz Model 9's, did a controlled experment, which concluded that bi amping was a waste of time. Of course this was before HF Digital Domain Filters came along, it was strictly analog in those days, so it's an apples and oranges situation mostly.

But one of the points he made was that he would insist on having protective capactiors on tweeters and mids to prevent any square waves/DC sledgehammers from getting thought if by some accident the evil RCA plugs should come loose between the signal and power amp sections, espcecially in solid state/transformerless amplifiers. Since RCA plugs disconnect Ground before Positive, there is a loud squeal before the output slams to the 30-70Volt supply rail, which can open up a fragile voice coil in milliseconds.

[Don Richard]

My question is this: Aren't the suggested 20 uF and 4.7 uF caps going to limit my ability to do what I want???

Those values are for factory crossover freqs. For different crossover freqs, go an octave below your lowest estimated freq, using the calculations. This is a non-critical value and will not make much, if any, audible difference in the sound of the speaker if left in.

[DrWho]

The purpose of the series cap is to prevent any accidental low frequency energy from bottoming out the compression driver. Sources of low frequency energy could include turn on transients, dc offset on the output, ground loops, or even incorrect filter settings (which could be caused by user error / equipment failure / software glitches).

Do you know how a voltage divider works Rudy? Here's something I found after a quick search that should explain it relatively straightforward:
http://knol.google.com/k/electronic-circuits-design-for-beginners-chapter-6#

A capacitor behaves like a resistor that changes its resistance with frequency...and then when placed in series with the driver, it creates a voltage divider that limits the amount of voltage presented to the driver. As you go lower in frequency, the impedance rises until it hits infinity at 0Hz (DC). As the frequency goes higher, the impedance drops towards zero at infinity Hz. To calculate the effective impedance of a capacitor, use this equation:

Z = 1 / (2pi x f x c)

where c is the rated capacitance in Farads and f is the frequency of interest in Hz (divide uF by 1,000,000 to get it's value in Farads).

A speaker driver also has an impedance that changes with frequency. Here's the simplest a speaker impedance will ever look:

http://sound.westhost.com/tsp.htm

The important thing to note is the impedance spike at the driver Fs. In this example, the Fs is around 27Hz so it's probably a woofer. A compression driver Fs is usually going to be relatively close to your xover frequency, and you're going to have other impedance shifts due to the acoustic loading from the horn. The impedance can end up being quite crazy style, but almost always ends up higher than the nominal value at the frequency of interest...

All that to say, if you manually enter the effective impedance of your capacitor and the effective impedance of your speaker at various frequencies, then you can calculated how much voltage attenuation you're going to get at different frequencies. The important thing to note is that choosing a capacitor to be one octave below the xover frequency based on the nominal impedance is going to result in very little, if any, attenuation around the Fs of the driver because the driver impedance is high. Sadly, the Fs of the driver is generally where you're going to see an excursion maximum...

One of the reasons I bring it up is to point out that you can't guarantee protection of the driver with a series cap unless you allow it to interact with the passband of the driver.

The series cap will always protect against DC, and can usually do a good job against ground loops, which are usually the two biggest culprits. Turn on transients will be gear dependant and its always best to use gear that has a soft start feature (which most good amps and every active xover I've run across have).

The effective output impedance from your amplifier when using a DC block cap gets reduced by the ESR of the cap. Pretty much any film cap will be more than sufficient to not affect the output impedance. In fact, if the ESR is low enough, then the cap could have 100% distortion and not introduce any noticeable difference in distortion to the total system because its limited to the current flowing through the ESR and it gets divided against the impedance of the driver. Caps don't have 100% distortion to begin with so there is no need to spend crazy money on the capacitor. Just make sure its film and has a high enough voltage rating for your application - over 100V should be more than sufficient and going higher doesn't hurt. (I don't even know if you could find a film with a lower voltage rating anyway).

[Rudy81]

I rember PWK, in the days of Mac 275's and Marantz Model 9's, did a controlled experment, which concluded that bi amping was a waste of time. Of course this was before HF Digital Domain Filters came along, it was strictly analog in those days, so it's an apples and oranges situation mostly.

But one of the points he made was that he would insist on having protective capactiors on tweeters and mids to prevent any square waves/DC sledgehammers from getting thought if by some accident the evil RCA plugs should come loose between the signal and power amp sections, espcecially in solid state/transformerless amplifiers. Since RCA plugs disconnect Ground before Positive, there is a loud squeal before the output slams to the 30-70Volt supply rail, which can open up a fragile voice coil in milliseconds.

I hate RCA plugs for the same reason. That is one reason I will be going strictly with balanced connections for this project.

[Rudy81]

Those values are for factory crossover freqs. For different crossover freqs, go an octave below your lowest estimated freq, using the calculations. This is a non-critical value and will not make much, if any, audible difference in the sound of the speaker if left in.

Don: Thank you. I'm slowly getting the idea.

[Rudy81]

DrWho: Excellent response, thank you for taking the time to cover the basics....which is where I am at right now. I had been wondering about how much protection is afforded when you are selecting a capacitor that is one or two octaves lower since, as you correctly pointed out the frequency chosen is near the drivers minimus safe crossover.

I will be using good amplifiers that have soft turn on and I assume the EV DC-ONE will also have those type of safety features. Coupled with balanced cables, I should be ok without capacitors. I'm not so much worried about the tweeters since they are relatively cheap to replace. Actually, the tweeters cost about as much as some Solen caps for this application. The JBL 2470 phenolic diaphragms are another story. They are hard to come by and rather expensive. Unfortunately, they are necessary to get down near the Khorn bass bin.

I will have to think about the cost vs. benefit on this one. I might just put the K55's back in while I try this out.

[Rudy81]

My question is this: Aren't the suggested 20 uF and 4.7 uF caps going to limit my ability to do what I want???

Those values are for factory crossover freqs. For different crossover freqs, go an octave below your lowest estimated freq, using the calculations. This is a non-critical value and will not make much, if any, audible difference in the sound of the speaker if left in.

Don: Can you tell me what parameters you used for the factory bass to mid settings? Using the online calculator, a 20 uF cap with a 16 Ohm driver gives me a frequency of 496 Hz. In my case, the Khorn bass bin is unchanged, so I will need to cross the mid in basically the same place.

Based upon everything discussed here, I had guess I needed a capacitor crossing at about 200 Hz with a 16 Ohm driver which is roughly a 50 uF capacitor....correct. (This is if we go with the one octave lower theory.)

[DrWho]

Just to follow up on my previous post...

I guess I wasn't too clear about it, but because you're not gonna protect the driver at its Fs, the actual value of capacitor you choose is mostly irrelevant as long as you get the desired rejection at the much lower frequencies (like 60Hz and below).

Choosing a 200Hz corner should be more than adequate...you might still get a little attenuation around 400Hz or wherever you plan to xover, but you should be able to compensate within the dsp.

Do you know the impedance of your compression driver between 200Hz and 400Hz on the horn you plan to use?

[Rudy81]

Just to follow up on my previous post...

I guess I wasn't too clear about it, but because you're not gonna protect the driver at its Fs, the actual value of capacitor you choose is mostly irrelevant as long as you get the desired rejection at the much lower frequencies (like 60Hz and below).

Choosing a 200Hz corner should be more than adequate...you might still get a little attenuation around 400Hz or wherever you plan to xover, but you should be able to compensate within the dsp.

Do you know the impedance of your compression driver between 200Hz and 400Hz on the horn you plan to use?

No, I don't, but I will see what I can find. I will be using a JBL 2470 with original diaphragm. For the time being, It will be protected at 200 Hz and below......just in case.

Last night I picked up a Parasound HCA-1205A for $425 to replace my HALO 5 channel. So, we're ready to go as soon as the DC-ONE gets here.

[Rudy81]

And now, some completely different questions. The project is now underway. I am waiting for the DC-ONE and I'll be able to actively cross my Khorns.

In light of the discussions of driver alignment etc, it seems the best thing to do is to use a single driver as much as possible. I am limited in the woofer to mid crossover by the Khorn design. However, my mid Trachorn with its JBL 2470 and the eminence APT-150 have quite a bit of overlap. That will allow me to set the high crossover point in a variety of places.

According to the JBL data, the 2470 can go up to 12 kHz and the Eminence can go down to 3.5 kHz. Ideally I would set the crossover as high as possible to keep the majority of sound coming from a single driver.

So, what do you guys recommend? I am very optimistic about this experiment and the flexibility I will have with the active crossover.

[8-|]

[AltmanEars]

And now, some completely different questions. The project is now underway. I am waiting for the DC-ONE and I'll be able to actively cross my Khorns.

In light of the discussions of driver alignment etc, it seems the best thing to do is to use a single driver as much as possible. I am limited in the woofer to mid crossover by the Khorn design. However, my mid Trachorn with its JBL 2470 and the eminence APT-150 have quite a bit of overlap. That will allow me to set the high crossover point in a variety of places.

According to the JBL data, the 2470 can go up to 12 kHz and the Eminence can go down to 3.5 kHz. Ideally I would set the crossover as high as possible to keep the majority of sound coming from a single driver.

So, what do you guys recommend? I am very optimistic about this experiment and the flexibility I will have with the active crossover.

6K will probably do it...