Your techniques make very good sense to me.
And your question "would I make different decisions" makes even more sense to me....
I think that question gets to the heart of what i was trying to express.
Measurements are tenuous at best, and I take them to be reality.
Because of the difficulty in knowing what they are and how to use them, I have to hold them at arm's length in terms of what decisions I can make based on them. So I try to find out what about them is something i know i can act on.
Sometime it seems like far-field are the answer, but sometimes near-field stitched seem to surprisingly work better. Crazy.
Simulations on the other hand, generally give "definitive outputs", providing comparisons where it's easy to say this one is better than that one.
So decisions look easy. But really? The assumptions embedded in sims are inevitably huge.
I guess bottom line is, I don't think we should ever take sim's minor differences seriously...and we should continuously question measurements as well, probably ignoring minor differences in them too.
Last edited:
Kimmo indeed states in his documentation that for circular drivers the vertical measurements are not really needed, a small bump in di only impact.
But why NOT take ALL of the measurements 0-180 degrees in 10 degree increments horizontal and vertical? In ARTA and REW I can set multiple measurements with a specific pause in between. I use 7 seconds because that gives me time to walk to my manual turntable and rotate it and get back out of the way. Taking 19 measurements instead of 10 takes an extra 4 minutes. Yes, the actual verticals may not make a BIG difference, but the cost of all the extra measurements is probably an extra 15-20 minutes of walking back and forth to the turntable. Compared to weeks spent on building the cabinet and figuring out the crossover.
For a small speaker that can be laid sideways on the speaker stand, yes it is certainly not much effort to do the vertical polars.
For a large speaker, it is not so simple to elevate it 40+ inches from the floor, in the horizontal position, and then rotate it about an axis for each driver. When I did this for the tower speakers I built, I had to build a rotating table. With a weight of 88 lb, I was not able to safely lift this beast onto the table by myself. Overall, it was a pain.
For this current project, I can balance the speakers horizontally on my stand with the axis of rotation at the midrange. But to make vertical polars of the woofer or tweeter, the speaker will fall off the stand. I would need to make some kind of structure or mechanism to hold it in position.
The above rotating table had a diffraction effect that was visible in the measurements. In fact, any speaker stand will have some diffraction effect. For horizontal measurements, the stand is underneath, so it is more or less representative of how the speaker will be used... but for vertical polars, any stand or support structure adds a non-representative diffraction to the measurement.
So in my opinion, making vertical polars can be quite difficult, and there can be an added level of uncertainty with the measurements.
j.
For a large speaker, it is not so simple to elevate it 40+ inches from the floor, in the horizontal position, and then rotate it about an axis for each driver. When I did this for the tower speakers I built, I had to build a rotating table. With a weight of 88 lb, I was not able to safely lift this beast onto the table by myself. Overall, it was a pain.
For this current project, I can balance the speakers horizontally on my stand with the axis of rotation at the midrange. But to make vertical polars of the woofer or tweeter, the speaker will fall off the stand. I would need to make some kind of structure or mechanism to hold it in position.
The above rotating table had a diffraction effect that was visible in the measurements. In fact, any speaker stand will have some diffraction effect. For horizontal measurements, the stand is underneath, so it is more or less representative of how the speaker will be used... but for vertical polars, any stand or support structure adds a non-representative diffraction to the measurement.
So in my opinion, making vertical polars can be quite difficult, and there can be an added level of uncertainty with the measurements.
j.
Yes in arta and manual rotating trntable does not take lot of time. About 10 seconds per measurement. The question for me is : what is the added value of measurements beyond 90 degrees? They are heavily conaminated by reflections, and if i remember well in VCad the use for CTA2034 is limited to 90 degrees.
Actually, I think we need measurements out to 180 degrees to adequately calculate power response. Without power response, we don't have a valid directivity index, or a valid PIR curve. For the way I design speakers, both of those curves are important.
The 90 - 180 curves look pretty messy. There can be reflections within the gate window, and certainly there are all kinds of diffractions and cancellations from the sides and back of the cabinet. But all this gets averaged together when making the power response calculation.
The 90 - 180 curves look pretty messy. There can be reflections within the gate window, and certainly there are all kinds of diffractions and cancellations from the sides and back of the cabinet. But all this gets averaged together when making the power response calculation.
I think I have seen some tests that show both (1) using 20 degree increments and (2) not going past (or much past) 90 degrees make a small (something like 0.5 db) difference in power response. If I was making some budget bookshelf speakers maybe I'd save the time. But my point was, if you are trying to design the best speaker you can, and are going to probably spend weeks designing it, why save 30 minutes when taking measurements.
Yes they are, but getting a proper indication you most certainly need to go all the way to 180 degrees.
It's still a tool, not the goal, the goal is to make a good performing speaker.
What have been already said before, many things can be already seen from preliminary simulations.
Things like diffraction issues to give one example.
So you can measure that again, but it won't give you any new information.
Just only confirmation and sometimes results that are just slightly different.
But again, a 180 degree measurement is not gonna change that.
Also don't forget that function follows form in many cases.
Or in simple words, the size and shape of the cabinet is basically set.
Staring at a power response is fine and all, but just looking at the off-axis response will give you the same idea.
Just not in one nice looking graph.
In the end it's looking at exactly the same thing, just on many different ways.
Again, it's a tool, not the goal.
If you would wrap turntable legs and microphone stand in felt or something used for speakers damping would you see difference in measurements?
Especially this leg close to the tweet and mid....
I am sure it would help, particularly at higher frequencies. The problem is the table makes the speaker baffle seem wider than it truly is. This is going to affect the diffraction hump in the 500 - 2k region, even with a felt or foam absorber.
Back in post 438, @xrk971 warned me of the potential for creating an unwanted resonance between two chambers. I have a lower chamber behind the woofer that is 11 liters, and upper chamber behind the tweeter that is 6.9 liters, and a narrow duct between them that is 2.5 liters.
https://www.diyaudio.com/community/threads/compact-low-cost-active-3-way-speaker.402812/post-7518083
My initial stuffing of the box at 16 g/l resulted in too much damping of the bass. My next step was removing 60% of the wool to get a density of 6.5 g/l. I removed it uniformly from all three spaces. The result was an odd bump in impedance at about 150 Hz, and I also saw a +3 dB bump in near field output at 150 Hz. If xrk971 had not made me aware of this, I might have wasted quite a bit of time trying to figure out what was happening. But since I was aware of the possibility, I understood the situation right away.
The damping strategy that worked was to stuff the tweeter chamber and the "duct" chamber at 16 g/l, and then lightly stuff the woofer chamber at 3.5 g/l. This produced a usable bass response. Thank you @xrk971
j.
https://www.diyaudio.com/community/threads/compact-low-cost-active-3-way-speaker.402812/post-7518083
My initial stuffing of the box at 16 g/l resulted in too much damping of the bass. My next step was removing 60% of the wool to get a density of 6.5 g/l. I removed it uniformly from all three spaces. The result was an odd bump in impedance at about 150 Hz, and I also saw a +3 dB bump in near field output at 150 Hz. If xrk971 had not made me aware of this, I might have wasted quite a bit of time trying to figure out what was happening. But since I was aware of the possibility, I understood the situation right away.
The damping strategy that worked was to stuff the tweeter chamber and the "duct" chamber at 16 g/l, and then lightly stuff the woofer chamber at 3.5 g/l. This produced a usable bass response. Thank you @xrk971
j.
You can also try putting some higher loss stuffing like open cell foam in the duct that connects them. Essentially isolating the two chambers but then you lose the volume that was supposed to be part of the woofer bass chamber. So I think your solution to add extra damping in the smaller chamber worked out well.
On measuring maximum output at 40Hz
Following on from comments earlier in the thread about maximum output at 40Hz, I decided to measure my compact speakers at 40Hz.
To cut a long story short, I still don’t know what SPL they can produce!
I spun up TrueRTA and set the freq to 40Hz with my preamp volume on minimum. I slowly turned up the volume intending to note the level at which I could start to hear audible distortion from the two pairs of 5.25” drive units. Unfortunately at much above 80db everything in my shed was rattling and shaking so much and making so much other noise that I didn’t stand a chance of hearing any distortion. Cone excursion was visually massively higher than I have ever seen when playing music, even when I play loud.
All I can say after my test is that 98db at 40Hz must be epic!
Following on from comments earlier in the thread about maximum output at 40Hz, I decided to measure my compact speakers at 40Hz.
To cut a long story short, I still don’t know what SPL they can produce!
I spun up TrueRTA and set the freq to 40Hz with my preamp volume on minimum. I slowly turned up the volume intending to note the level at which I could start to hear audible distortion from the two pairs of 5.25” drive units. Unfortunately at much above 80db everything in my shed was rattling and shaking so much and making so much other noise that I didn’t stand a chance of hearing any distortion. Cone excursion was visually massively higher than I have ever seen when playing music, even when I play loud.
All I can say after my test is that 98db at 40Hz must be epic!
Hi, did you measure distortion while doing the test? I bet you heard a lot of distortion and little actual 40Hz, except what brain makes you hear the 40Hz just fine, it builds the fundamental for you from the harmonics. Non-linearities in the driver moving parts and motor would start to compress the lows as excursion increases, while you could still crank the volume and hear and even measure things get louder. Even though the fundamental doesn't increase much the harmonics will and make impression of loud bass.
If you are still in the shed, try do another test: put some other tone with the 40Hz, like 500Hz or 1000Hz for example. I speculate you'd hear that higher tone distort while the bass sounds "clean". If you'd now use high pass filter or reduce SPL of the 40Hz tone to reduce excursion, the high tone would clean up, and this is the distortion you'd hear with small speaker, bass signal distorting the midband while bass itself could sound just fine.
Another test with just a low tone would be to sweep a little, say at 47Hz there could be almost no sound due to room modes. On the other hand, if the fundamental isn't too loud, the harmonics being higher up in frequency could even things out a bit.
Perhaps your driver is good and it sound clean as you describe!
Anyway, this kind of tests are fun stuff!
If you are still in the shed, try do another test: put some other tone with the 40Hz, like 500Hz or 1000Hz for example. I speculate you'd hear that higher tone distort while the bass sounds "clean". If you'd now use high pass filter or reduce SPL of the 40Hz tone to reduce excursion, the high tone would clean up, and this is the distortion you'd hear with small speaker, bass signal distorting the midband while bass itself could sound just fine.
Another test with just a low tone would be to sweep a little, say at 47Hz there could be almost no sound due to room modes. On the other hand, if the fundamental isn't too loud, the harmonics being higher up in frequency could even things out a bit.
Perhaps your driver is good and it sound clean as you describe!
Anyway, this kind of tests are fun stuff!
Last edited:
This was more of a “finger in the air” test to see which way the wind was blowing….
I’ll do some more testing in due course, but I am very confident having watched the massive excursion that it will not get close to 98db SPL at that frequency.
Like you say it was super fun 🙂. I’m away for a while so no more testing for some time yet…
I’ll do some more testing in due course, but I am very confident having watched the massive excursion that it will not get close to 98db SPL at that frequency.
Like you say it was super fun 🙂. I’m away for a while so no more testing for some time yet…
With my main system, I have a pair of 12" SB34NRX75-6 drivers (one per side). During testing, the loudest I have taken the system is 105 dB SPL at 32 Hz to 48 Hz. This is with both woofers, measured at approximately the listening position. The woofers could have gone louder, but the glass in the stereo cabinet was rattling out of the frame, and there was a lot of buzzing/rattling coming from all over the house, from windows and furniture. Based on that testing, I estimated that 98 dB SPL capability at 40 Hz, with one driver radiating into 2 pi space, at 1 m distance, would be adequate to produce 98 dB SPL at the listening position with both drivers playing. In most music, there is not much L-R stereo signal in the deep bass... it is almost all L+R mono.
These Dayton drivers will not do what the SB drivers will do... they have smaller Sd and smaller Xmax, which results in about half the Vd. But I think they will be enough.
Hey All,
I have been keeping close tabs on this project since it's inception but I am generally a lurker around here trying to gain as much knowledge as possible. This speaker design is fabulous and the price to performance ratio seems near unbeatable.
I tend to use headphones or IEMs for much of my musical listening, but these may be able to push me into making a set with some slight modifications for musical listening. Based on reading almost all discussion here, off axis response and diffraction problems were obviously a driving force behind much of the design decisions. As I am not as technically up to date with current speaker design trends and advanced simulation using VC2, I was wondering whether people think that doing a 5 channel surround using similar driver concepts would be possible.
My current plan is to use almost the exact speaker design for the LR channels but adapted to use a passive XO. This will obviously entail doing some modifications to the box sizes as I won't be able do do any fancy active XO tricks. Additionally, I would construct the LR surrounds using the 6" version of the mid driver used in this design (SB17NBAC35) with the same tweeter used here to allow for similar XO points and aesthetics utilizing the compound bevel cuts.
However, I am struggling coming up with a competent design for a center channel due to their general low profile nature. My current though is to use two SB15NBAC30's and a Tectonic TEBM46C20N BMR in a horizontal MTM setup with a lower XO ~250-300Hz to minimize off axis distortion.
I generally see that centers are not particularly popular here due to most people preferring standard stereo listing as opposed to surround, but was wondering whether anyone would have any suggestions for/against my current plans.
Thanks,
O
I have been keeping close tabs on this project since it's inception but I am generally a lurker around here trying to gain as much knowledge as possible. This speaker design is fabulous and the price to performance ratio seems near unbeatable.
I tend to use headphones or IEMs for much of my musical listening, but these may be able to push me into making a set with some slight modifications for musical listening. Based on reading almost all discussion here, off axis response and diffraction problems were obviously a driving force behind much of the design decisions. As I am not as technically up to date with current speaker design trends and advanced simulation using VC2, I was wondering whether people think that doing a 5 channel surround using similar driver concepts would be possible.
My current plan is to use almost the exact speaker design for the LR channels but adapted to use a passive XO. This will obviously entail doing some modifications to the box sizes as I won't be able do do any fancy active XO tricks. Additionally, I would construct the LR surrounds using the 6" version of the mid driver used in this design (SB17NBAC35) with the same tweeter used here to allow for similar XO points and aesthetics utilizing the compound bevel cuts.
However, I am struggling coming up with a competent design for a center channel due to their general low profile nature. My current though is to use two SB15NBAC30's and a Tectonic TEBM46C20N BMR in a horizontal MTM setup with a lower XO ~250-300Hz to minimize off axis distortion.
I generally see that centers are not particularly popular here due to most people preferring standard stereo listing as opposed to surround, but was wondering whether anyone would have any suggestions for/against my current plans.
Thanks,
O