Most of those that provide a default or defaults typically use a maximally flat (one of Small's, QB3 or whatever) for the sake of simplicity, with the ability to adjust as desired. Those formulas are also usually the ones employed by simple on-line calculators. They have to use something, and max-flat is a logical place for people to start from.
A lot of agreement in recent posts that sims are only as good as the folks who use them.
In fact and speaking of bass (which is all we bother with when T/S is the topic), FR is today of substantially less importance than in the past.
First, we all measure our net final sound at our chair since anybody with a $60 mic and REW can do so. That net result is monumentally affected by lots of factors in the room and speaker operation that aren't determined entirely in the T/S model.
Second, nostalgic audiophiles excepted, DSP-EQ is available today (and since 2003) for anybody with $150 and would like to try it. So FR, as modelled as if the speaker was hanging from a crane in your backyard, is subject to parametric EQ to a fare-thee-well even at real low frequencies.
So what should we prioritize in design? Transient behaviour? Distortion (hey, go find that in your T/S model)?
B.
Hey you could design a mechanical system to do that, fed from an analyzer that determines which is most advantageous tuning frequency moment-by-moment and stoppers the right holes instantly!
Well, on this device I can't read the one you posted, but what JBL recommends used to not be based on home use at all. I remember for some of the pro woofers' parameters, calculation for a sealed box with Qtc=0.707 would yield a box the size of, oh, putting the woofer face down on the floor. When I talked to their engineers about it, the assumption was in a real use situation (concerts, discos) the woofers would heat up significantly and drive up the effective Q, meaning you needed quite a different enclosure from what a calculator would tell you.
I think it goes back to Small's thesis, and since that is based on electrical filter theory, some baggage carried over. He told me that yes he just used -3 dB because it was mathematically convenient (and IIRC rather usual for electrical filters as a reference point). But yes "at the cost of low performance" because in actual rooms you have room gain, and the -6 dB and -10 dB points are likely to be more important than F3.
Tell me more please. I used to run MLSSA and export the data into LEAP, but that was long ago and I'm out of touch with current options. And I cannot connect my LEAP parallel printer port key to my MacBook Air ha ha. Yet I plan to get back to designing a project and need more modern software.
I also very much would like to be able to simulate tone bursts. Even if just from a simple electrical filter model would be interesting. I just find it a very visual representation. I could once upon a time do this by exporting a simulated response backwards into LMS and it would do a transform and then I think I could simulate a tone burst...or was it just the impulse response, hmmm.
Yes, if I understand you since you can juggle specs to get the same results, ergo conversely juggle them to get radically different ones, so seems better to solve for one or more specs to get the lowest practical total impedance and of course there's the box Q and any vent losses to account for to get the desired system Qt [sysQ].
GM
Hello,
I am soon able to get away to my hideout by the beach by myself for a couple of weeks.
This time I am attempting to build a pair of aperiodic vent woofers. Most of the stuff that I read says that this type of build is largely experimental. One thing I have seen in multiple sources is that the enclosure should be about the size of sealed enclosure tuned to near 0.71Q with a smallish port tuned to be largely resistive in nature.
Reading the reviews over on ASR indicates that tuned reflex ports often interferer with the speaker output. Other sources talk about stuffing a sock into the reflex port to tame the upper impedance peak that interferers with the speaker output.
This time I am looking at JBL 2226H drivers in about 2 net cubic enclosures with aperiodic vents.
There will be iterative steps.
The port will be 4 inches in diameter;
First step will be a 4” dia. by 2 inch long tube.
Second step will be a 4” dia. by 4 inch long tube.
Third step will be a 4” dia. by 6 inch long tube.
Next, Yes, I will stuff a sock in it, three different socks. The socks will be lengths of nylon stocking, donated by my wife, stuffed with increasing amounts of Dacron pillow stuffing.
Yup there will be three ports with three different value resistive plugs for a total of nine different test conditions.
Thoughts?
Thanks DT
I am soon able to get away to my hideout by the beach by myself for a couple of weeks.
This time I am attempting to build a pair of aperiodic vent woofers. Most of the stuff that I read says that this type of build is largely experimental. One thing I have seen in multiple sources is that the enclosure should be about the size of sealed enclosure tuned to near 0.71Q with a smallish port tuned to be largely resistive in nature.
Reading the reviews over on ASR indicates that tuned reflex ports often interferer with the speaker output. Other sources talk about stuffing a sock into the reflex port to tame the upper impedance peak that interferers with the speaker output.
This time I am looking at JBL 2226H drivers in about 2 net cubic enclosures with aperiodic vents.
There will be iterative steps.
The port will be 4 inches in diameter;
First step will be a 4” dia. by 2 inch long tube.
Second step will be a 4” dia. by 4 inch long tube.
Third step will be a 4” dia. by 6 inch long tube.
Next, Yes, I will stuff a sock in it, three different socks. The socks will be lengths of nylon stocking, donated by my wife, stuffed with increasing amounts of Dacron pillow stuffing.
Yup there will be three ports with three different value resistive plugs for a total of nine different test conditions.
Thoughts?
Thanks DT
Sounds very interesting. What measuring setup will you use? And if you're shooting for Q=0.7 surely check the JBL parameters. Some of their woofers used to have crazy weird parameters, I guess because once they heating up they would be "right" or something.
Oh I'd add a test with panty hose material across the port because...well, because. Once you're set up, why not? I'd also be curious to see a test with actual socks stuffed in there, and maybe an angled "port bung" if you can make one out of foam or such. The point of the latter two being to see how well DO they really seal the port-or not
I will stretch a coffee colored nylon across each of the test ports
Hello,
The setup will use the AP APx555 analyzer, APx1701 Transducer tester (test amplifier) and AP555 speaker test software.
I will cut some 6 inch square pieces of ½ inch plywood, drill a hole in each square piece for the various length test ports and attach each one in turn to the test enclosure. I will leave one piece of ½ ply intact to seal the test opening in the test enclosure. Thanks for the reminder.
The JBL 2226H data sheet says that the driver is warmed up for two hours with 600Watts pink noise prior to testing the TS/P.
Me, I am going to sweep the speaker several times at 2 or 3 Watts until the data looks stable and call the data from the last sweep good. I did read somewhere that JBL tests the TS/P at -20dB’s less than 600Watts.
Yes I will stretch a coffee colored nylon across each of the test ports and record the data as well, thanks good idea.
Thanks DT
Hello,
The setup will use the AP APx555 analyzer, APx1701 Transducer tester (test amplifier) and AP555 speaker test software.
I will cut some 6 inch square pieces of ½ inch plywood, drill a hole in each square piece for the various length test ports and attach each one in turn to the test enclosure. I will leave one piece of ½ ply intact to seal the test opening in the test enclosure. Thanks for the reminder.
The JBL 2226H data sheet says that the driver is warmed up for two hours with 600Watts pink noise prior to testing the TS/P.
Me, I am going to sweep the speaker several times at 2 or 3 Watts until the data looks stable and call the data from the last sweep good. I did read somewhere that JBL tests the TS/P at -20dB’s less than 600Watts.
Yes I will stretch a coffee colored nylon across each of the test ports and record the data as well, thanks good idea.
Thanks DT
DT,
2226H data sheet says :
"Thiele/Small parameters are measured after 2 hour exercise period using a 600 W AES power test and will reflect the expected long term parameter values once the driver has been installed and operated for a short period of time."
The driver is not "warmed up" for two hours, it is "exercised" to break in the suspension so that Fs will not read too high. If it were still "warmed up" it would also exhibit 4.6 dB of power compression (sensitivity loss) due to the heat of the voice coil having made the impedance rise considerably- more than double.
Also, the parameter values may take a long time to change (break in) if not operated at levels making excursion approach Xmax- the AES "exercise period" has the driver driven with pink noise limited to 6dB crest factor, open air, at a voltage that results in the rated power at the rated impedance, though the actual power is far less, as the impedance is far higher than the nominal rating. Plenty of excursion though, does not take much voltage for excursion to exceed Xmax down low, open air.
Although a single 4" port will be adequate for low power, if you were to use the driver near it's rated power and Xmax, the slug of air in the port that is supposed to wiggle back and fourth will be "blown out", also known as "port compression".
The 4" port at 2" should result in an Fb around 60Hz, 4" 50Hz, the 6" 45Hz.
The 2" & 6 " stacked together would get Fb down to around the 2226H Fs of 40Hz.
Stuffing the port will throw away much of the relatively little low frequency gain provided from the lower tuning frequencies in the "too small" box.
Have fun at the beach!
Art
Last edited:
In as much as there is considerable opinion ridiculing the notion of warm up (some say drivers are not horses), it would be very nice of you to do some basic measurement(s) before and after the "exercise".
I'd say just noting the driver resonance (using any old uncalibrated impedance set-up) would be illuminating. Does it shift more than a couple of Hz and did that happen to be up or down?
B.
I'd say just noting the driver resonance (using any old uncalibrated impedance set-up) would be illuminating. Does it shift more than a couple of Hz and did that happen to be up or down?
B.
Ben,
In all the examples of suspension "break in" I have heard of, the Fs shifts down. The changes can be surprisingly large, "before" and "after" exercise.
As an example, when testing the resonant frequency of a brand new B&C 8NDL64-8 (8 ohms) measured the Fs at 87 Hz, 7 Hz higher than the spec sheet rating of 80 Hz.
I then “broke in” the speaker by pushing the cone in and out by hand to near Xlim (as far as the suspension would allow, around 15mm peak to peak), then repeated the test, and the Fs dropped to 78 Hz.
I had used around 4V for the two initial tests, after a few minutes decided to try 1V (”small signal”) and read around 82 Hz, 4 Hz higher.
Bringing the voltage back up to 4V, the Fs dropped to 80Hz. The suspension probably was still not fully "broken in", but was already pretty close.
I've been using the 80 Hz Fs drivers in boxes with a 35Hz Fb for almost a year, they seem to be "broken in" now ;^).
Art
Last edited:
loud enough for my wife to tell me to turn it down
Hello,
I am not getting in too deep into the break-in warmup discussion. I am sure that there are valid considerations on both sides of the discussion.
The pair of JBL 2226H’s I have on the shelf in the garage were purchased used. If there was any break-in it was accomplished a long time ago.
TS/P for these drivers will be tested at 4Volts input or 2Watts. Later the impedance and SPL plots of a speaker installed in the test enclosure will also be swept at 4 volts input. This is near typical operating conditions in the listening room at my house, possibly loud enough for my wife to tell me to turn it down. Not a disco.
I have tested TS/P on a number of drivers previously. I have found that the TS/P test quite a bit differently between the storage shelf in the cold garage and after being in the 65 degree “Lab” overnight, then there is still more delta in the tested TS/P after a dozen or so test sweeps in the “Lab”. There is not a magical set of TS/P. TS/P are all over the place depending on test conditions.
Thanks DT
Yes I will post some before and after exercise plots.
Hello,
I am not getting in too deep into the break-in warmup discussion. I am sure that there are valid considerations on both sides of the discussion.
The pair of JBL 2226H’s I have on the shelf in the garage were purchased used. If there was any break-in it was accomplished a long time ago.
TS/P for these drivers will be tested at 4Volts input or 2Watts. Later the impedance and SPL plots of a speaker installed in the test enclosure will also be swept at 4 volts input. This is near typical operating conditions in the listening room at my house, possibly loud enough for my wife to tell me to turn it down. Not a disco.
I have tested TS/P on a number of drivers previously. I have found that the TS/P test quite a bit differently between the storage shelf in the cold garage and after being in the 65 degree “Lab” overnight, then there is still more delta in the tested TS/P after a dozen or so test sweeps in the “Lab”. There is not a magical set of TS/P. TS/P are all over the place depending on test conditions.
Thanks DT
Yes I will post some before and after exercise plots.
Last edited:
All good posts.
But like I said, opinions are all over the place about breaking-in and "exercise"*.
But for the here-and-now and curious-minds-want-to-know, could you please just give us the scores, say at 4 volts, room temperature, out of the box, again later after one push on the cone, and after some good "exercise".
Some test methods are crude, even if your digital meter reads to three decimals places. Much of the enterprise of doing T/S "at home" is fraught with inexactitude. But folks take the numbers** and plug them into your favourite sim and get out the circular saw to build the box. All other reasons aside, it is one reason to like sealed boxes since they have only a single parameter and it just doesn't matter much if you get it on the button.
B.
* truly super new book on fitness: "Exercised" by Harvard paleo-anthropologist Daniel Lieberman.
** figuring out what numbers have to be right-on and what numbers can be just good guesses is called "sensitivity analysis"
But like I said, opinions are all over the place about breaking-in and "exercise"*.
But for the here-and-now and curious-minds-want-to-know, could you please just give us the scores, say at 4 volts, room temperature, out of the box, again later after one push on the cone, and after some good "exercise".
Some test methods are crude, even if your digital meter reads to three decimals places. Much of the enterprise of doing T/S "at home" is fraught with inexactitude. But folks take the numbers** and plug them into your favourite sim and get out the circular saw to build the box. All other reasons aside, it is one reason to like sealed boxes since they have only a single parameter and it just doesn't matter much if you get it on the button.
B.
* truly super new book on fitness: "Exercised" by Harvard paleo-anthropologist Daniel Lieberman.
** figuring out what numbers have to be right-on and what numbers can be just good guesses is called "sensitivity analysis"
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.