3-Way Active - High WAF - Hi-Vi-Fas

[devinkato]

Here's some pics of a new build I've been working on.

Main complaint my girlfriend has had about floorstanders is that they're very big and bulky, so I've been mostly using small/mid sized cabinets sitting on speaker stands. I'm tired of the stands, and wanted to go to a floor stander.

I also miss my old 3-way system, so that was another design priority. The main downside to 3-ways is the expense of crossover components. I have a bunch of amps laying around, so I decided to go active using a Mini DSP.

Using the Hivi woofers made famous by the Overnight Sensations in a similar cavity size and porting setup. For the midrange, I'm using a 2.5" Vifa aluminum cone fullranger in an isolated chamber (used this in a previous design, and I really liked it). And for the tweeter, going with a .75" Vifa silk dome tweeter.

Woofer - http://www.parts-express.com/pe/show...number=297-429
Midrange - http://www.parts-express.com/pe/show...umber=264-1052
Tweeter - http://www.parts-express.com/pe/show...umber=264-1006

The speakers were designed to be very thin - I used a ton of internal bracing and will be stuffing it with Roxul insulation to absorb as much of the backwave as possible, along with deflex pads mounted behind the woofers. The internal cavity for the woofers actually stops about 6 inches short of the bottom. The bottom is filled with solid pieces of MDF glued together. It is very very solid.

I'm used to being able to reach through the woofer cavity and running wires to the terminal posts. Not the case with this guy, too tall and narrow. I cut an access point on the back. Will have to fabricate a nice cover for it that will house the 6 binding posts. Thinking a nice, thin, beveled piece of acrylic may look cool.

Also have to figure out what to do for feet/base.

[Geniuss_dad]

Re: 3-Way Active - High WAF - Hi-Vi-Fas

can you tell me the h-w-d measurements on these? I like the look so far.

[ReissM]

Re: 3-Way Active - High WAF - Hi-Vi-Fas

Cool design. What crossover frequencies did you choose for these drivers?

Active XO's and tri-amp setup? That should be wild.

[devinkato]

Re: 3-Way Active - High WAF - Hi-Vi-Fas

can you tell me the h-w-d measurements on these? I like the look so far.

6"x6"x42"

Tweeter is at ear level to my listening point. A tad higher if I'm slouching

[devinkato]

Re: 3-Way Active - High WAF - Hi-Vi-Fas

Cool design. What crossover frequencies did you choose for these drivers?

Active XO's and tri-amp setup? That should be wild.

None yet.

I usually use 12db electrical slopes to start. 400hz and 4500hz depending on piston size. The 2.5" mid shouldn't hit any beaming till ~5400hz, so I should be good there as a starting point.

[ReissM]

Re: 3-Way Active - High WAF - Hi-Vi-Fas

None yet.

I usually use 12db electrical slopes to start. 400hz and 4500hz depending on piston size. The 2.5" mid shouldn't hit any beaming till ~5400hz, so I should be good there as a starting point.

Well... sort of. I'm going to offer some personal opinions that you might disagree with, but here it goes.

I like that you're doing the math and figuring out how high you should go with the midrange driver. (In this case, a 2.5" diameter driver) A 2.5" wavelength would equal about 5400Hz, however there's more to the story. I did some research a few years back and found that calculating for one wavelength doesn't give you the dispersion/ beaming that you think it does.

Take a peek at the chart I created. It was derived from research done waaaaay back.




It looks like a confusing mess at first. Start with the diagram on the left. This is what your calculations are producing. When the diameter of the driver approaches the wavelength that you're trying to produce, then you get the first diagram on the left. This dispersion pattern is not very good. Off axis response is far worse than you might have expected when you were doing your "beaming" calculations.

If you were to limit the 2.5" driver to the frequency associated with half of a wavelength instead of one wavelength, your dispersion would look more like the second figure. Ahhhhh, now that's better! (See the middle diagram.) This is what people are trying to show when they talk about beaming.

Taking this to the next logical step, look at how amazing the dispersion is if you limit the driver to one quarter of a wavelength. See the diagram on the right.

Reality kicks in...
It's not quite practical to limit your 2.5" diameter midrange driver to an upper limit of 1356 Hz (see chart, green section), so some compromises need to be made. Having the midrange's upper limit be 1356 Hz would dictate having the tweeter ROLL ON at 1356 Hz, and that is rarely done for several reasons, most obvious is that you will burn the tweeter out fairly quickly.

So in general I try to use the green section (1/4 wavelength) of the chart whenever possible to get the best dispersion. That's fairly easy for low frequencies (woofers) but I usually end up with the yellow section (1/2 wavelength) for midrange drivers.

Let's get specific with your chosen drivers:
In your case, a 2.5" midrange can produce the dispersion shown in the middle diagram if you limit it's upper freq to about 2700Hz. (By pushing the 2.5" mid up to 5400 Hz it would produce the dispersion shown in the left diagram...not good) Now, the good news is that most tweeter's can easily handle the power when rolled on at 2700Hz. So crossing the mid to tweeter any higher than 2700 Hz just makes your dispersion/ beaming worse.

Does this make any sense? You were on the right track with your math, but I would respectfully suggest modifying your calculations by a factor of 2 (or 4) to give better results. This can sometimes mean using a smaller diameter driver to get better dispersion. In your case, your 2.5" mid is among the smaller midrange drivers already... so that's a bonus.

Try re-running the calculations with a six-inch midrange. They're not so pretty... with an upper limit of 1130 Hz in order to obtain the yellow dispersion shown in the middle of the chart. How the heck are you going to roll-on a tweeter at 1130 Hz? So picking a smaller diameter like 2.5" allows you to push it's upper limit to 2700Hz. Perfect!

This just goes to show that everything in speaker design is tradeoffs. I prefer using small diameter drivers for the midrange frequencies and picking low crossover frequencies to obtain the best dispersion I can (as compared to 5 or 6" midrange drivers) As long as the drivers can handle the power and excursion limits associated with low crossover points, it should improve the dispersion and reduce beaming problems.

A quick 3-way example:
An 8" woofer can easily play up to 200 Hz and still be in the green section of the chart. (great dispersion!)

Add to that -> a 2.5" mid playing 200 Hz to 1700 Hz (almost in the green section... great dispersion!)

Add to that -> a tweeter capable of playing 1700 Hz to 20 kHz and you've got a very wide dispersion design that minimizes beaming effects.

It's physics! Use this knowledge to your advantage to pick crossover frequencies that make sense and produce better results.

[devinkato]

Re: 3-Way Active - High WAF - Hi-Vi-Fas

ReissM - thank you absolutely so much for taking the time to share your research. It validates what my ears hear - going up to the "calculated" beaming points doesn't always sound best. It also sounds you, like me, like smaller drivers for dispersion - one of the exact reasons I prefer 3 ways both in car (my main background), and at home. I will definitely try and drop the mid/tweeter crossover point - the tweeter can definitely handle being played down low.

One thing, the attachment doesn't seem to load for me. I'd love to see it.

Last question/comment - I know a lot of people try to avoid having a crossover point in the middle of the audible midrange - you seen to not agree with this, as you mention crossing over right at ~ 2000hz. Is this something you've wrestled with, or think its a non factor in your experience? In car audio (worst environment) at least, having a point source driver handle as much midrange as possible has proven to improve imaging. Of course, this is dealing with driver placement, time alignment, etc, which can be controlled much better at home.

[ReissM]

Re: 3-Way Active - High WAF - Hi-Vi-Fas

One thing, the attachment doesn't seem to load for me. I'd love to see it.

Ahhhhhhhh. My entire post depends on you be able to see the JPG graphic. Let me try again...

[devinkato]

Re: 3-Way Active - High WAF - Hi-Vi-Fas

Ahhhhhhhh. My entire post depends on you be able to see the JPG graphic. Let me try again...
[ATTACH=CONFIG]30616[/ATTACH]

Thanks for that - While I don't think that I'll be able to get down the omnidirectional range of the 1/4th wavelength, I'll definitely keep that in mind, and get between there and the 1/2.

[ReissM]

Re: 3-Way Active - High WAF - Hi-Vi-Fas

Last question/comment - I know a lot of people try to avoid having a crossover point in the middle of the audible midrange - you seen to not agree with this, as you mention crossing over right at ~ 2000hz. Is this something you've wrestled with, or think its a non factor in your experience? In car audio (worst environment) at least, having a point source driver handle as much midrange as possible has proven to improve imaging. Of course, this is dealing with driver placement, time alignment, etc, which can be controlled much better at home.

In my experience, if properly implemented, it's fine. So what does "properly implemented" mean?

Well, let's take a step back for a second. In a perfect world we would have one transducer reproducing the entire audible range from 20 Hz to 20 kHz. A point-source. As you probably know, it's nearly impossible for a single driver to do that. The high cone excursions required for deep bass will mess up the delicate midrange. Also, in order to produce deep bass we typically need a larger diameter woofer... which will have beaming issues at higher frequencies. (We don't normally see a 12" woofer playing up to 20kHz right? Even if it could, the beaming would be terrible, like a flashlight shinning out light in a very narrow cone.)

So we typically split up the range into 2 or 3 pieces. Hence "2-way" and "3-way" designs. There are even 4 and 5-way designs but let's not get carried away.

So in order for 2 drivers to appear as though they are acoustically one, we try to locate them as close to each other as physically possible. Remember that at the crossover frequency both drivers are producing sound. Depending on the slope of the filters, the overlap can be 1 to 2 octaves above and below XO freq. Any time there are multiple sources producing the same sound there will be interference issues. This issue can be minimized by placing the acoustic centers as close together as possible.

Combining the knowledge of beaming that I discussed earlier with the physical proximity of the 2 drivers to each other is key.

2-way example:
We could push a 12 woofer up to 1800 Hz then have a tweeter take over from there up to 20 kHz. The problem is that the dispersion of the 12" woofer will be pretty bad at 1800Hz while the 3/4" tweeter will have a nice, wide dispersion at the same frequency. That wouldn't be a well designed system as far as driver integration goes.

A 6" woofer would be a better choice in the above design. It's dispersion would be much better at 1800 Hz and the two drivers would integrate better.

So to answer your question, I personally don't think it's a problem to crossover drivers at 2 kHz (or any other freq.) Just keep the two drivers located as close as you physically can, and cross them as low as you can safely do it without risking driver failure or high distortion. The lower the frequency, the longer the wavelength... and thus, the less chance that the physical offsets can create problems.

These are just my personal opinions and I'm sure there are many out there that will disagree. I'm open to hearing other people's opinions on this subject.

[devinkato]

Re: 3-Way Active - High WAF - Hi-Vi-Fas

Thanks again ReissM - I'm glad my little project has gotten some gears turning.

I do realize the importance of C2C in terms of point sourcing multi way speakers - Although I do believe it becomes less important the further away the listening position is. I used speed of sound / C2C spacing = MAX crossover point. This is MUCH easier in a 3-way than in a 2-way with a large woofer. This is the exact reason I am a HUGE proponent of 3-way front stages in car audio. There is no way that a 2-way setup with a woofer in the door, and a tweeter on the dash will EVER be okay acoustically.

As far as home audio - I tend to stay away from any personal 2-way builds, as there are a lot out there that have been proven to work really well and I'm perfectly content standing on the shoulders of giants for those projects. That said I tend to do more "random" projects such as this thin 3 way for my own use. I definitely tried to keep the drivers as close as cosmetically possible.

My question regarding crossing over in the audible range was more intended to discuss phase issues that occur around crossover slopes (not phase perfect crossovers). Are these phase issues also mitigated by C2C spacing, or are they simply due to the laws of physics in the crossover math? I know that many 2-ways crossover using crazy order slopes sometimes without issue smack dab in the middle of the range, yet I keep on reading (mainly from full range nut heads) how important it is to keep any sort of phase issues out of the mid range.

[devinkato]

Re: 3-Way Active - High WAF - Hi-Vi-Fas

Edges done.

Went with a 3/4" roundover on the sides, and a 3/4" chamfer on top. I didn't engage the chamfer all the way to the full 3/4" - woulda looked too deep. As always, I had (on the sides) the bit cut end right on the edge of the joint, so if it starts telegraphing the joint in the future, its not noticeable to the eye, as you already expect a line there.

[ReissM]

Re: 3-Way Active - High WAF - Hi-Vi-Fas

My question regarding crossing over in the audible range was more intended to discuss phase issues that occur around crossover slopes (not phase perfect crossovers). Are these phase issues also mitigated by C2C spacing, or are they simply due to the laws of physics in the crossover math? I know that many 2-ways crossover using crazy order slopes sometimes without issue smack dab in the middle of the range, yet I keep on reading (mainly from full range nut heads) how important it is to keep any sort of phase issues out of the mid range.

You touched on several things here... let me make a few statements and then try to answer your questions.

At this point I'd like to say that I'm not an expert and you can choose to dismiss any of my ramblings as nonsense.


With no XO network:
The raw driver will have an acoustic frequency response and acoustic phase response. In addition, there will be an electrical impedance curve plus an electrical phase curve.


The Crossover filter:
The choice of crossover filter type (Butterworth, Linkwitz-Riley, Bessel, Chevychev, etc..) and choice of slope (6, 12, 18, 24 dB/octave) will introduce a particular electrical response.


The trick is to be able to combine all of the acoustic stuff above (response and phase) with all of the electrical stuff above (response and phase) and wind up with something that works well. The biggest reason I use Jeff Bagby's Passive Crossover Designer (PCD) is that it shows you all of the above items and the results that you should expect when they are ALL combined. This is no small task, and I can testify that it is amazingly accurate. The fact that he gives PCD away for free is quite shocking. I feel like I'm stealing a $500 software package every time I use it.

So back to Phase. How does that relate to frequency response?

People are always talking about phase... both acoustic phase and electrical phase but it's one of those techie terms that many don't fully understand. Not too long ago (prior to my PCD days) I assumed that all 2nd order networks had to be reverse-phased on the tweeter. As I saw it, the two-pole filter introduces 180 degrees of phase rotation so I figured you always had to flip the tweeter's polarity. I oversimplified things and completely ignored the acoustic phase of the driver and the interaction it has with the filter. Sometimes the combination of driver phase and filter phase requires a reversal of the polarity, sometimes it doesn't. Curt was kind enough to explain it to me.

My point is, without looking at all of the phases and amplitudes, you don't have the full picture of what's going on. For example, if the phase is not tracking between a mid and a tweeter, then the frequency response can get a little mixed up. The mid and tweeter's responses won't add together smoothly and you might hear something that sounds a little "off". You might find that the two drivers are adding nicely below the XO freq but canceling above the XO freq. What the heck? I used to think that they would either add perfectly or cancel to produce a deep null. But there are dozens of variations on the XO filter designs that can produce a whole mess of in-betweens if you don't take everything into consideration.

It's possible that a 2nd order lowpass on the woofer will add nicely with a 3rd order highpass on the tweeter. Or maybe the the other way around... it all depends on the phase of the driver/filter combination. Those textbook formulae are only the tip of the iceberg... a starting point that needs to be tweaked for every driver that you use, in order to get the phase and frequency response correct.

So back to your comment about keeping any sort of phase issues out of the midrange... My opinion is that if a multi-way speaker is designed and optimized correctly, (utilizing all acoustic and electrical phase data) then it shouldn't matter what frequency you're mid-to-tweeter crossover is designed around. Just my opinion.

And yes, keeping the center-to-center distances as small as possible certainly helps matters. But as always, it depends on the wavelengths involved.

Example 1: Two woofers are separated c-t-c by 9". That 9" corresponds to a wavelength of about 1500 Hz. But if the woofers have a Low Pass roll off at 80 Hz, then the shortest wavelength they will reproduce is up around 170 inches, so the c-t-c distance of 9 inches doesn't really matter all that much in this case.

Example 2: Mid to tweeter c-t-c distance is 9" and XO freq is 3000 Hz. At 3000 Hz the wavelengths are about 4.5". So you've got 2 point sources separated by 9" trying to both reproduce a 4.5" wavelength. That's going to be a nightmare, creating massive constructive and destructive interference patterns. In this case the 9" separation REALLY matters!


Sorry that I wrote a novel in response to a simple question. Have fun!

[devinkato]

Re: 3-Way Active - High WAF - Hi-Vi-Fas

Sanded and filled in all the cracks and hit it with some high build primer. If anyone has any favorite high build primers that I should try, let me know. This is just the first one I saw and bought.

http://www.tcpglobal.com/restorationshop/itemdetail.aspx?itemno=KUS+KIT-KFP305-GL

[ReissM]

Re: 3-Way Active - High WAF - Hi-Vi-Fas

As always, I had (on the sides) the bit cut end right on the edge of the joint, so if it starts telegraphing the joint in the future, its not noticeable to the eye, as you already expect a line there.

I saw somebody here on TT do this trick a while back. I love this idea and think I might borrow it on some future builds. (I'm always learning something on this forum!)

The fact that your eyes already "expect" to see a line/ edge at that exact location is brilliant for hiding seams, and doesn't look fudged in any way. Nice job.