Flint Acoustics: Custom High Fidelity Patio Speakers (future self-powered)

[Flint]

I do have one Frequency Response chart I can post - this is on axis indoors:


Chart: Frequency response of speaker, on axis at 1M, smoothed 1/6 octave

I want to do some ground plane measurements in order to have accurate bass data, but that's a good representation of the system.

Remember, these will be mounted directly on a wall and near the ceiling, so the bass below 200Hz will have plenty of reinforcement to increase the output and extension.

 

[Flint]

I thought I'd also share the phase response of this speaker, of which I am EXTREMELY proud.


Chart: Frequency Response and Phase, 1/3 octave smoothed

Take note of how the phase curve is very smooth. No insane shifts in direction throughout the midrange where crossovers tend to cause all sorts of problems. While phase response is not completely audible, a speaker with a smooth phase response reflects good driver alignment, good blending between drivers, an easier load on certain types of amplifiers, and typically a better polar response. All that effort I put into these crossovers paid off in spades.

 

[Haywood]

Hot damn. These may be the best sounding patio speakers in the history of patio speakers.

 

[PaulyT]

Nice, thanks for all the detail. Fun!

 

[Flint]

The weather was nice, on the humid side, but not terribly hot (which means every time I was in the shade I had to do fierce battle with mosquitoes), so I setup on the front sidewalk to do some ground plane measurements of the woofer.


Photo: Setup for Ground Plane measurements. Microphone is 2M away to calibrate to a 1M placement in an anechoic environment.

So, I took a bunch of measurements and got the following "dynamic performance" data:


Chart: Dynamic Ground Plane Frequency Response at 2.83V (black), 4.88V (brown), 28.3V (red) and 40V (yellow)

These voltages correlate to 1W, 10W, 100W, and 200W into 8 ohms and as such demonstrate what this one 6" woofer can do when being fed very excessive power. A few determinations can be made:

1. The woofer didn't blow at 200W, which I feared it might.
2. The 1W to 10W pseudo-anechoic bass extension is about 40Hz and the bass limit is about 31Hz.
3. The speaker can generate over 100dB SPL at 80Hz (wow!)

If I normalize the measurements and adjust for voltage gain you can see how the acoustic output compresses at higher power inputs:

Chart: Normalized Power Compression measurements at 1W (black), 10W (brown), 100W (red) and 200W (yellow)

This time you can see:
At 10W the compression is minimal above 50Hz and only about 3dB at 38Hz. Not bad for a small 6" woofer.
At 100W the compression is relatively smooth to about 60Hz, but isn't really having trouble until about 50Hz.
At 200W the woofer is struggling below 65Hz, but even shows significant stress in the 200Hz range.

Still, these are VERY good results from such a small mid-woofer.

 

[Flint]

I also grabbed THD measurements at various input levels:


Chart: 2.83V THD: 40Hz = 2.5%, 80Hz = 1.2%, 160Hz = 0.56%


Chart: 8.94V THD: 40Hz = 5.6%, 80Hz = 2.8%, 160Hz = 0.75%


Chart: 28.28V THD: 40Hz = 71%, 80Hz = 25%, 160Hz = 6.3%

I didn't make a THD chart for 200W because I feared blowing the woofer with the slower, long term sine sweep required for the test.

If you consider the size of the woofer and enclosure, the bass frequency acoustic gain from the placement, and the use of two woofers (3dB to 6dB additional output), then this is pretty good results.

 

[Flint]

Here are the gated "pseudo-anechoic" off axis response curves in 15 degree increments:

Horizontal dispersion:


Chart: On and off axis response curves at 1W/1M at 15 deg increments to the inside (toward listening position)


Chart: On and off axis response curves at 1W/1M at 15 deg increments to the outside (away from listening position)

Since the drivers are mounted off axis, the inside off axis response will be different from the outside off axis response. This is generally a good thing. As can be seen, the off axis coverage is VERY smooth across the midrange and treble with moderate and controlled dips and drops throughout the coverage area. This shows a smooth and natural power response which will sound balanced and natural across the listening area and as sound is reflected off side walls.


Vertical dispersion:


Chart: On and off axis response curves at 1W/1M at 15 deg increments downward (toward floor/seated listening position)


Chart: On and off axis response curves at 1W/1M at 15 deg increments upward (toward ceiling)

In these curves you see a more dramatic variation in the response, especially in the crossover range since the two drivers' output in equal and cancelling each other out do to differences in coincidence (arrival time). By placing the woofer above the tweeter, the overall significance of the drop in output around the crossover and the range below is greater as the mic is moved upward. This means the reflections off the nearby ceiling will have less output in the critical listening range and thus have less impactful comb-filtering. I often get asked why tweeters are typically placed above the midrange drivers, and this is one of the main reasons (at least in high end speakers). The floor is closer to the mid/tweeter combination and the phase cancellations are greater as you move in the direction of the woofer than in the direction of the tweeter. Since these will be near the ceiling, I flipped the arrangement over to help with the sound reflected off the ceiling.

 

[Flint]

Also, since these are hand-made, I need to ensure they both perform the same. I don't really want to run all of the measurement tests on both speakers, so I run samples of the tests on both to ensure key starting points are the same between the two.

Here's the basic on axis 1W/1M measurement for both the left and right speakers:


Chart: Left (black) and right (red) speakers comparison. Smoothed 1/3 octave.

below about 800Hz they are virtually identical, so the black curve blocks the red. Above that there is more variation, but it is very moderate and inaudible in real life. The difference is more likely from the driver variation from the factory than from my builds.

 

[Flint]

After all the crazy measurements were completed and now that I feel I am done with these speakers, I mounted them high on stands on my own porch, acoustically similar to how they will be installed on Haywood's patio:


Photo: Speakers on porch

They are reversed left and right, but that's because they are farther apart than ideal for stereo listening if they were the other way around.

I then measured the response at 1M to show how the gain from the walls and ceiling might impact the bass:


Chart: Left & right speaker response on porch at 1M, smoothed 1/3 octave

As you can see, the bass extension is pretty great! The speaker closer to the side wall has extended bass to below 30Hz and to the other is showing gain of about 6dB SPL from "anechoic". Of course, there are other nulls caused by the floor and ceiling, but look at the midrange!!!! Whoop!!!!

Here's both speakers with a mono test tone with the microphone in the yard about 15 feet away:

Chart: Mono signal, both speakers, about 4M from speakers (about 5W input per speaker)

Even at 4M distance the bass extends to below 30Hz quite easily and the gain is a little higher than the midrange - the classic house curve.

I am very pleased!!!!

 

[Haywood]

Like I said. Best patio speakers ever. Take that, Speakercraft.

I am dying to hear them.

Does this mean you got the amps in?

 

[Flint]

Does this mean you got the amps in?

I am struggling with how I can transmit audio to the amps. There isn't a simple way to connect the HDMI switch/transmitter to any of my gear, but at some point I am obviously going to figure it out.

 

[Haywood]

I am struggling with how I can transmit audio to the amps. There isn't a simple way to connect the HDMI switch/transmitter to any of my gear, but at some point I am obviously going to figure it out.

There is a USB transmitter with an analog input on its way with the other amps.

 

[Flint]

For the data nerds out there, here are the dispersion polar plots:


Chart: Horizontal Polar Dispersion Plot. Negative Numbers for inside the on axis response, Positive Numbers for outside.


Graph: Horizontal polar plot Normalized to On-Axis.


Graph: Vertical Polar Plot, negative numbers for downward angle, positive numbers for upward (plot should be inverted to reflect real life).


Graph: Vertical Polar Plot normalized to on-axis curve.

Most of this is just informational, but the vertical plots which show how sound is projected upwards (+) and downwards (-) are interesting as to how these will perform. Notice in the bottom, normalized, plot, the downward dispersion has much more energy than the upward dispersion. This means that less energy will be projecting towards the ceiling and reflecting back and causing cancelations and nulls. Ceiling reflections aren't eliminated, but they are reduced and will improve the experience for the listener.

 

[PaulyT]

Interesting so why is it that there's such a dip in response in the vertical off-axis plots around 2-3k? Both in the polar graphs in the last post, and the line charts up above. Is that a mathematical consequence of the speaker shape or ... ?

 

[Flint]

Interesting so why is it that there's such a dip in response in the vertical off-axis plots around 2-3k? Both in the polar graphs in the last post, and the line charts up above. Is that a mathematical consequence of the speaker shape or ... ?

Quite simply, the crossover is around 2100Hz, so at exactly that frequency the two drivers are putting out the exact same amount of energy and as you go up in frequency the woofer rolls off. As you go down in frequency, the tweeter rolls off. Since both driver are equal loudness at 2100Hz, as you move up and down, one driver moves closer to the mic while the other moves away, and that determines how much phase difference there is between the two signals and it cancels out.

 

[PaulyT]

Ah, I see, thanks.

 

[Randy]