[Kevin Haskins]

Ha... I've been playing with the Seas DXT tweeter on the Kepler design.   I went active for this version because it gave me some extra latitude in the crossover design (I can get better results).   

All of these are raw data, fed to LSPCAD for modeling the network.   No smoothing is used nor am I calculating the off-axis plots.   It is based upon a raw off-axis measurement data. 

Off-axis 22deg.



Off-axis 22deg null.  I added a 0-180 phase rotation to time align so the nulls are deep.  This typically only works well on one axis of measurement.   The closer you get the time alignment, and if you are using moderately steep crossovers, the null remains through a larger cross-section of off-axis measurements.   It goes to hell on the vertical axis though because of the physical path-length differences between drivers. 



Off-axis 35deg.



And the 35deg off-axis null.



45deg off-axis



60deg. off-axis



75deg off-axis



This is way out to the side.   The little response irregularities that you start to see are nothing compared to what you would typically see this far off-axis with a standard 2-way.   Keep in mind, most data you will see has been smoothed.   Using 1/6th octave smoothing would remove some of the fur in the measurement.     You really cannot do anything about the narrowing response of a 6.5" midwoofer but you can pick a crossover point that lines up the response as close as possible to the wider dispersion of the tweeter used above it.   


Much better than what I can do with the Peerless or for that matter any tweeter that doesn't have this kind of directivity control.   

[S Clark]

Congrates.  Very impressive FR curve.  I bet that it sounds very accurate.

[Kevin Haskins]

Congrates.  Very impressive FR curve.  I bet that it sounds very accurate.

Don't know yet... it should. I'm still playing with measurements and modeling.   I have a prototype PCB coming for the active crossover and I'll get that populated and start some listening trials and measurements with the network in place.   I'm also extending the bandwidth another 5-10 Hz deeper.   Having active control allows for a lot of adjustments on the bottom-end that you just cannot do with a passive network.    I also threw a couple notch filters on the board so I can deal with any room issues.   

[S Clark]

Congrates.  Very impressive FR curve.  I bet that it sounds very accurate.

  ... I also threw a couple notch filters on the board so I can deal with any room issues.   

I thought most you guys gated your measurements to cut out room response.??

[Kevin Haskins]

Congrates.  Very impressive FR curve.  I bet that it sounds very accurate.

  ... I also threw a couple notch filters on the board so I can deal with any room issues.   

I thought most you guys gated your measurements to cut out room response.??

For crossover design, sure.   I take the measurements outdoors with the speaker on a platform, lifted from my forklift so that I get a good 14mS window.    The data from these is all extremely clean from 150Hz up.    The couple extra notch filters are so that I can take a long 500mS window in the room, and EQ a few of the room issues once they are in the house.   They wouldn't be for an end-user because you have to be able to take good room measurements, populate the PCB with the right components, then re-measure to confirm that everything is right.   

[MaxCast]

I take the measurements outdoors with the speaker on a platform, lifted from my forklift

I'd like to see that on youtube.      Mommy, what's that man doing with that speaker way up there?   

[Kevin Haskins]

I take the measurements outdoors with the speaker on a platform, lifted from my forklift

I'd like to see that on youtube.      Mommy, what's that man doing with that speaker way up there?   

My neighbors stand around and stare as the speaker chirps.   They are always curious what they hell I'm doing and when I explain it, they don't look any less confused.   

[Danny Richie]

I keep looking at those tweeters and wondering what I could get out of them in a larger version of that face plate (more of a wave guide)? How much it would lift the bottom of its range and how low it could be crossed? Interesting for sure.

[Kevin Haskins]

I keep looking at those tweeters and wondering what I could get out of them in a larger version of that face plate (more of a wave guide)? How much it would lift the bottom of its range and how low it could be crossed? Interesting for sure.

The distortion measurements are safe now.   I'm running them to 1.5K with a fairly steep slope.    I've not tested them lower because the hump in the response of the tweeter where the horn gain + output of the tweeter is at a maximum makes it difficult to shape the response without adding more network.   I could hammer it into shape but it just isn't worth it for another couple hundred Hz.   The directivity is all gone by then anyway so there isn't much point to it other than I would have less of the directivity of the 6.5" midwoofer in the final response.   

This is raw data, no smoothing so if you consider the final one is off-axis 75deg, I'm pretty happy with this as a starting point.   

   

[Danny Richie]

I'm pretty happy with this as a starting point.

I would be too. It looks great.

I'd really love to see one with about a 6" wave guide that would allow a crossover point in the 1kHz range. The Fs would have to drop a little more though. I'd rather see it in the 500Hz range or less to be doing that to one.

Like you, I'd rather not have to use too many crossover parts either.

Very interesting still.

[Kevin Haskins]

I've seen 1" devices run that low but even with horn gain, you still have a LOT of excursion for a tweeter to handle.    We also did a 5" horn, not the DXT, but a shallow WG.   It had a little more directivity that reached a little lower but not much.     It maybe gave another couple hundred Hz of directivity control, and could have been run to 1K but distortion was rising quickly.    I would have crossed it over no less than 1.2K and then only with a 4th order or higher acoustic slope.   

When you think of the excursion demands, x4 the stroke for the same SPL for every octave you drop, the limit jumps up on you quick and you have to do drastic things to get a 1" device to run much lower.    Probably better to use a 3-way at that point and just deal with the extra cost and extra crossover.   

[Danny Richie]

Not necessarily.

If the wave guide gives you enough gain (acoustically) you can wind up with a lower crossover point with a smaller cap value on the tweeter than what you had to begin with when you had no wave guide and a higher crossover point.

Electrically the tweeter sees less low frequency signal sent to it (less exertion) than before.

[Kevin Haskins]

Not necessarily.

If the wave guide gives you enough gain (acoustically) you can wind up with a lower crossover point with a smaller cap value on the tweeter than what you had to begin with when you had no wave guide and a higher crossover point.

Electrically the tweeter sees less low frequency signal sent to it (less exertion) than before.

With the right horn, but I think you will find that the horn gets big quick.   You start trading-off other things like response smoothness and getting into fairly large devices.    You may be able to stretch a very capable tweeter to 1K in a 6" horn.    Wiggins has an XBL^2 tweeter with a big fat surround designed that could do it.   Even with that kind of tweeter though when you looked at the distortion profile, it is rising at 1.2K or so.    If you design a horn with gain down @ 1K or below it gets relatively deep and big.   Ours was good to about 1.6K.   You need a fairly large size for directivity control and of course as the device gets bigger, your increasing the distance between tweeter and midrange, which hurts your vertical off-axis performance.   

For all that work you would have to ask yourself what you gained.    You might get a little better horizontal off-axis control, give up vertical off-axis, and probably pick up some response issues from having a deeper horn.   Just a guess.... I'm going to play more with shallow horn profiles and see what I get.     The Seas unit is pretty darn good though if used properly.   For an off-the-shelf solution and it's size, it performs pretty good.   

[Danny Richie]

Everything is a little give and take.

You'd be surprised what you can get out of a properly shaped 6" wave guide though. This is a 6" wave guide that Bob Smith cut for one of my GR-T6 tweeters.



The only thing I didn't like was the cancellation in the top half of the top octave.

Note the gain in the low end though.

The Red line is the response with no crossover and the Green line is the response with a simple second order crossover on it.

I want to say that the cap value was only a 3.9uF or a 4.7uF and the gain from the wave guide could have allowed it to cross at about 900Hz without being worked hard at all.

I look forward to seeing what you do with the Seas unit. That thing looks really good.

[Kevin Haskins]

Everything is a little give and take.

You'd be surprised what you can get out of a properly shaped 6" wave guide though. This is a 6" wave guide that Bob Smith cut for one of my GR-T6 tweeters.



The only thing I didn't like was the cancellation in the top half of the top octave.

Note the gain in the low end though.

The Red line is the response with no crossover and the Green line is the response with a simple second order crossover on it.

I want to say that the cap value was only a 3.9uF or a 4.7uF and the gain from the wave guide could have allowed it to cross at about 900Hz without being worked hard at all.

I look forward to seeing what you do with the Seas unit. That thing looks really good.

That is a monster amount of gain.   I guess it depends on what SPL level you target, THD threshold, and of course how it looks when measured in the design.    The question is still what does it gain by crossing over an extra 500-600Hz lower?    For me the horn or waveguide is primarily a way to manage the power response.   If it doesn't gain me any better off-axis both vertical and horizontal, I'm not really interested in running it lower.    If it improved power response, I'd do it at the sacrifice of a little THD.   

Oh...that give and take!     

[Danny Richie]

The question is still what does it gain by crossing over an extra 500-600Hz lower?

Good question. Ever build a speaker two ways and listen to them both? One crossing at 2kHz and one crossing at 1.5kHz? You might be surprised.

Hmmm let's see, allow 1kHz and up to be covered by a big heavy diaphragm having to also play down to 50Hz or lower, with a lot more mass and stored energy? Or a smaller light weight diaphragm with much less mass and stored energy?

I was surprised at the resolution difference between the LS-6 and LS-9 that used the exact same drivers but the crossover point dropped from 1kHz to 850Hz.

Also the THD might be even lower with the small cap value on it crossed at 1kHz using a wave guide than it would with a big cap value on it crossed higher with no wave guide.

Now if I could have just not had the cancellation above 15kHz.....

[Kevin Haskins]

The question is still what does it gain by crossing over an extra 500-600Hz lower?

Good question. Ever build a speaker two ways and listen to them both? One crossing at 2kHz and one crossing at 1.5kHz? You might be surprised.

Hmmm let's see, allow 1kHz and up to be covered by a big heavy diaphragm having to also play down to 50Hz or lower, with a lot more mass and stored energy? Or a smaller light weight diaphragm with much less mass and stored energy?

I was surprised at the resolution difference between the LS-6 and LS-9 that used the exact same drivers but the crossover point dropped from 1kHz to 850Hz.

Also the THD might be even lower with the small cap value on it crossed at 1kHz using a wave guide than it would with a big cap value on it crossed higher with no wave guide.

Now if I could have just not had the cancellation above 15kHz.....

Yes...... I've used a number of crossover points.    I'd say it is an apples-oranges comparison though because you have different power response when you change the crossover due to the different directivity of the drivers.    So.... it is hard to say why the subjective differences are what they are.    Is it because of a lighter diaphragm, or because of the different power response?    All the research indicates the power response is the critical difference.   There isn't much to show that the mass of a diaphragm has any relevance to sound quality other than it's possible affect on the acoustic properties it lends to the device it is used within.    In other words, if it measures well, on/off-axis then it matters not what is producing the sound.   

In terms of stored energy, waterfall plots are almost useless (Toole & Olive 1988, Olive 1994).    They don't tell you enough about the character of the resonances to be tied to what is and is not audible.       
 

[Danny Richie]

Is it because of a lighter diaphragm, or because of the different power response?

I have my theories.

There isn't much to show that the mass of a diaphragm has any relevance to sound quality other than it's possible affect on the acoustic properties it lends to the device it is used within.    In other words, if it measures well, on/off-axis then it matters not what is producing the sound.

Then if my speaker measures exactly the same as your speaker (measuring only amplitude in all directions) then they will sound the same right? I think not.

In terms of stored energy, waterfall plots are almost useless (Toole & Olive 1988, Olive 1994).    They don't tell you enough about the character of the resonances to be tied to what is and is not audible. 

That is a very subjective statement and I do not agree with it. CSD are highly under rated.

If we can easily hear the stored energy differences (dissipation rates) of various capacitors having only slight measurable differences then how much more so the differences measured from driver to driver which by comparison are quite substantial?

[Kevin Haskins]

There isn't much to show that the mass of a diaphragm has any relevance to sound quality other than it's possible affect on the acoustic properties it lends to the device it is used within.    In other words, if it measures well, on/off-axis then it matters not what is producing the sound.

Then if my speaker measures exactly the same as your speaker (measuring only amplitude in all directions) then they will sound the same right? I think not.

Your ears are measuring it right?   They respond to stimuli and I'd say it is reasonable to say that if the stimuli is the same, in two different situations then the results should be the same.   Wouldn't you argee?    It would take more than frequency response measurement to fully characterize it but at the end of the day, our ears are a pair of microphones connected to a signal processing device.    They either measure the stimuli and process it or they don't.     The how is the interesting part.     

In terms of stored energy, waterfall plots are almost useless (Toole & Olive 1988, Olive 1994).    They don't tell you enough about the character of the resonances to be tied to what is and is not audible. 

That is a very subjective statement and I do not agree with it. CSD are highly under rated.

It is the exact opposite of a subjective statement.  It is one that is based upon objective data and that is why I made it.   It isn't my opinion.  It is based upon the research of the audibility of various resonances. 

I think we are heading down the same road we have been before Danny.   I understand we have our differences and I don't think we are going to resolve them this time.    Lets just agree to disagree about these things and not dig a hole arguing over it.   

[Danny Richie]

Okay, I'll just answer your questions and not go deeper into this. If it even looks like we are arguing then we are done. I had more questions for you but will withhold them.

Your ears are measuring it right?

I was referring to a microphone measurement.

They respond to stimuli and I'd say it is reasonable to say that if the stimuli is the same, in two different situations then the results should be the same.

If all else is equal then I agree. But is all else equal (not really meant as a question, just a provoking thought)? We were only talking about amplitude in all directions. If all else were the same then they would be the same speaker.

It would take more than frequency response measurement to fully characterize it but at the end of the day, our ears are a pair of microphones connected to a signal processing device. They either measure the stimuli and process it or they don't.  The how is the interesting part.

Ah, I agree. And aren't our ears more than a device that measures amplitude (again, not really asking a question)? I can play a 1kHz note on a piano and a 1kHz note on a flute to the same amplitude, but they do not sound the same. And yes, the how is a very interesting part.

Quote
In terms of stored energy, waterfall plots are almost useless (Toole & Olive 1988, Olive 1994).    They don't tell you enough about the character of the resonances to be tied to what is and is not audible. 

Quote
That is a very subjective statement and I do not agree with it. CSD are highly under rated.

It is the exact opposite of a subjective statement.  It is one that is based upon objective data and that is why I made it.   It isn't my opinion.  It is based upon the research of the audibility of various resonances.

I think anytime anyone says what is not audible then it is subjective. You might as well say "I" didn't hear a difference, and "I" don't think anyone will hear a difference. Or "I" don't hear a difference in "my" test, with "my" system.....

There are subjective differences even if we just look at hearing ability alone. When I had a group of the high school kids over we did some listening experiments. We humans are only "supposed" to only hear from 20Hz to 20kHz. But I had two girls here that could hear 22kHz immediately and repeatedly. The rest of the class could not.