Coaxial Active Speaker Build

[IPunchCholla]

LaVoce was kind enough to send me their FRD and ZMA files, so I was able to eat some FR curves for my baffle at a few different volumes.

LaVoce’s recommendation of 14.5 liters tuned to 90hZ using 4th order Butterworth crossover at 2200 hZ:

Some SpeakerSim’s auto calculate feature, 9.2 liters tune to 76 using the same crossover:

And 10.6 liters tuned to 88hz:

And just for fun, throwing it int to REW and generating an eq:

And with the crossovers and eq FIR filters +sub:

 

[Wolf]

That is not xovered at 2.2k, it's more like 1.7k.

 

[IPunchCholla]

That is not xovered at 2.2k, it's more like 1.7k.

First full speaker build (second cross over). I just realized I assumed cross over referred to the settings on the cross over not where the speakers actually crossed especially since the 13 fir filters that follow would alter that significantly.

ETA. LaVoce recommends a crossover of 2200 hZ. I started with that as the HP and LP target for building the filters. I then tried every filter (1st through 4th Butterworth, Linkwitz, etc.) in various combinations to see what effect that would have. I then tried varying the frequency. The above was the best result I could find.

After your post, I tried varying the frequency much more in order to get them to cross over at 2200. I finally managed it using 3800 as the crossover frequency. It was not great in terms of FR response. I then built EQ filters and applied those. The actual crossover point dropped back to 1811 hZ and the FR response is far more variable.

The problem is, once you bring down the response of the tweeter to be in line with the woofer, either using a resistor or reducing power, the actual crossover point needs to be around 1700-1800 HZ to get a good response, which is exactly what happens when you use the recommended (2200 hZ 12dB or greater slope) crossover point from LaVoce.

 

[ChrisG]

It's worth checking your port output velocity at your intended power levels. Those ports look very small compared to the size of the driver, so I'd expect some compression (loss of output) as power levels increase.

NB - best to simulate with the intended/expected highpass filter in place.

Chris

 

[Grotti]

It's worth checking your port output velocity at your intended power levels. Those ports look very small compared to the size of the driver, so I'd expect some compression (loss of output) as power levels increase.

NB - best to simulate with the intended/expected highpass filter in place.

Chris

Right: the ports should be approximately 1/4 of the diameter of the bass driver, when two are used. Or 1/3 when it's only one....

 

[ChrisG]

I prefer to check the port velocity, usually with Hornresp. You need to consider excursion as well as driver diameter: a long-throw subwoofer driver will need a much larger port than a same-size midbass operated with a highpass filter.

FWIW, I've measured port compression with a port area of 1/3rd Sd, losing around 2dB of peak output. It was a compact subwoofer for a PA application, so I was squeezing everything I could out of a smallish box. Otherwise, I would probably have made the port even bigger, and/or morphed it into a quarter-wave resonator.

Chris

 

[IPunchCholla]

Thanks to both of you! My sims show I would need 4 ports this size to keep the air speed in them under 5% SOS. Or, of course, double the area.

Not having used multiple ports before, in the calculators I’ve seen, is the length for the tuning is for all port tubes combined or for each tube.

 

[fineMen]

Thanks to both of you! My sims show I would need 4 ports this size to keep the air speed in them under 5% SOS. Or, of course, double the area.

Not having used multiple ports before, in the calculators I’ve seen, is the length for the tuning is for all port tubes combined or for each tube.

That's not a valid criterion. Port noise will depend heavily on the port flare. There is a calculator available online, which still is a bit optimistic.
From my personal, anecdotal experience port noise is more distracting than low order distortion. So better take care ;-)

Regarding frequency response tailoring I would recommend to first e/q the individual drivers well beyond the intended bandwith and only after set the x/over in. Since e/q is 'minimum phase' (IIF) this will do wonders. It should the least prevent the destructive combination seen in your simulations in post #21 @2,2kHz. This includes the low end of the satellite. A Linkwitz biquad could shift the effective fs to e/g 20Hz as to not interfere with the x/o to the sub.

I always advocate a single test build because one never knows.

 

[Jukka]

Price (I might need 16 of them) and listed performance mostly. Kinda just winging it with this one to see what it can do Since I can’t find any measurements.

You choose the driver by price, but use FA122? You must be joking!

I'd suggest that you used a passive crossover and single amp channel for each speaker box. If you could fabricate active crossovers, you could do some pretty complex designs with less cost than FA122. If you must have DSP, but can manage crossover part with passive parts, something like Dayton Audio KABD-4100 will power 4 boxes. If you really need DSP for each driver, that Dayton will power 2 boxes and gives you more budget for driver selection.

There are even more options in PA amps if fan nose is not an issue.

 

[IPunchCholla]

You choose the driver by price, but use FA122? You must be joking!

I'd suggest that you used a passive crossover and single amp channel for each speaker box. If you could fabricate active crossovers, you could do some pretty complex designs with less cost than FA122. If you must have DSP, but can manage crossover part with passive parts, something like Dayton Audio KABD-4100 will power 4 boxes. If you really need DSP for each driver, that Dayton will power 2 boxes and gives you more budget for driver selection.

There are even more options in PA amps if fan nose is not an issue.

I might not be rational, but I’m not joking. The speaker specs are really good compared to more expensive competitors. I’ll see if they hold up in box.

These are going to be playing 24 hours a day for months. They will also be traveling to multiple locations. So single box convenience, reliability, built in DSP, and physical digital connection are all requirements. The only Dayton amp I saw that met those specs was more than the FA-122. Happy to hear other possible amps, but I’m definitely doing active.

 

[TNT]

@IPunchCholla - how is this going? Do you like the LaVoce driver?

//

 

[IPunchCholla]

@IPunchCholla - how is this going? Do you like the LaVoce driver?

//

Between work an d the holidays, I haven't been able to put any time into physical progress. I've just been filling in here and there simulating different combos in SpeakerSim. Below is the one for the LaVoce Coax using their FR files and REW to create PEQs and crossing over using simulated FIR 4th order Butterworth:

Then I remembered I have 16 Dayton DC160S-4 I could repurpose from an old audio installation, so I began playing around with cheap tweeters just to see what might be possible (again 4th order Butterworths in the digital domain):

So then I wanted to be able to look at how they might look in terms of horizontal and vertical dispersion, so I modeled up a variant with the two drivers. But, unfortunately VituixCad is only available on Windows, so I need to get a cheap PC to teach myself that...

Anyway, always something more to be done, it seems, before getting to the next step.

 

[ugman]

Ironically, I recently built a center channel speaker for myself using this driver. Few weeks back made some faital pro 10inch 2 ways with a peerless alum tweeter, crossed fairly low around 1600Hz - sounds good so far, having crossover mounted outside speaker on back for 'long term road test'. Then made center using the lavoce 8inch coax mentioned here as was very happy with the pro speaker 10 inch and wanted something also high efficiency. Center channels tricky, especially since have mine sitting in media cart under tv where baffle diffraction portions of the response difficult to quantify. But I knew if threw a random center channel type speaker down there had to much smearing in 200-600 range (guess). So my speaker was built focusing more on the crossover point and even response from 500-20k. While measuring, tweaking, listening and repeat, I knew bass would be thin but was focused on making sure no peakiness in ranges when listening to various genres despite this speakers main focus will be on dialog. Went through about 3 iterations of crossover approach. First laned at 2300-2400, but when put into watching a tv series for audition, it was plagued was often encounter. Knew I need more tweeter action as the 8 inch just goes blah after 1800Hz or so, rough guess. So pulled the xover point down to 1600-1800 with 3rd order on tweet. Was being finicky, but finally landed on something that sounded right in 'lab' that knew was thin sounding but knew in installed position that'd be good. And it was - very happy with currently. Will see over next days or weeks. Was getting frustrated thinking would have to resort to setting up measurement rig in room and dsp/active ing... Nope, nice and easy 92db+ speaker for center channel that holds audibility when walk around the room