The speakers I'm building have a triangular port in the lower corner, i.e., the port shares two enclosure walls. However, they are monitors not subwoofers, and as such will be elevated on stands rather than be placed on the floor. My question is what k (end correction) value to use in calculating port length.
I came across this, but can't locate its origin to explore the conditions: http://www.ctc-dr-weber.de/speaker/PortCorrection.jpg
A regular poster here, bjorno, has previously commented that the k value of 2.227 (lower right) includes the influence of the floor plane outside the enclosure. If so, then I would assume this is also the case for all three lower examples.
My triangular port correlates with the lower center example in that two cabinet wall are shared. However, if the illustration assumes the speaker is sitting on the floor then there is a continuing plane at three port exit points, two inside the enclosure and one outside (the room floor).
I understand that the total k value is arrived at by adding up the conditions at each port exit, and have tried to extrapolate from the three lower illustrated k values for a sum for "port wall plane continues beyond port exit." No success. The consistent sum appears to be ~ 0.498 per shared wall.
Can anyone shed the light of knowledge into my conundrum? If in fact k = 1.728 with two cabinet walls shared plus the influence of the room floor, then what is the correct k when placed on a stand?
Thanks for all help,
David
I came across this, but can't locate its origin to explore the conditions: http://www.ctc-dr-weber.de/speaker/PortCorrection.jpg
A regular poster here, bjorno, has previously commented that the k value of 2.227 (lower right) includes the influence of the floor plane outside the enclosure. If so, then I would assume this is also the case for all three lower examples.
My triangular port correlates with the lower center example in that two cabinet wall are shared. However, if the illustration assumes the speaker is sitting on the floor then there is a continuing plane at three port exit points, two inside the enclosure and one outside (the room floor).
I understand that the total k value is arrived at by adding up the conditions at each port exit, and have tried to extrapolate from the three lower illustrated k values for a sum for "port wall plane continues beyond port exit." No success. The consistent sum appears to be ~ 0.498 per shared wall.
Can anyone shed the light of knowledge into my conundrum? If in fact k = 1.728 with two cabinet walls shared plus the influence of the room floor, then what is the correct k when placed on a stand?
Thanks for all help,
David
Last edited:
David,
For what it is worth, using triangular corner ports and slot ports I have found the difference between being on the floor and on a stand to only change fB by around 1Hz, less than I would have thought.
Since port calculators seldom agree with each other or actual measured response, I find the best solution is make the port a bit longer than predicted, measure and cut down if needed.
You can easily determine Fb by visually measuring excursion using sine wave tones, the excursion minima is at Fb.
If the Fb is lower than predicted, reduce port length.
Art
For what it is worth, using triangular corner ports and slot ports I have found the difference between being on the floor and on a stand to only change fB by around 1Hz, less than I would have thought.
Since port calculators seldom agree with each other or actual measured response, I find the best solution is make the port a bit longer than predicted, measure and cut down if needed.
You can easily determine Fb by visually measuring excursion using sine wave tones, the excursion minima is at Fb.
If the Fb is lower than predicted, reduce port length.
Art
That's reassuring to know. I also thought it would make more of a difference.
Since I lack test equipment I'm trying to optimize my math, so to speak.
So, dumptuck, you're suggesting to start with the k value indicated for one shared wall (1.23) and then shorten as needed to hit Fb. Regarding fitting the port length, a nice thing about the type of port in question (customized triangular slot port) is that I can vary the area of the port until the corresponding length fits most comfortably in the enclosure's depth (as large as possible provided there's sufficient breathing room at the back).
Getting close is about all you can expect of "math".
The method of finding Fb I mentioned in #2 requires no test equipment other than your eyes and a ruler.
You can find sine wave files on line if you don't have a generator.
Thanks for this suggestion. As stated, I don't have test equipment. Is there freeware that includes a sine wave generator? If so, in addition to the visual check, I believe I could measure Fb with the mere purchase of an inexpensive DMM, could I not?
This works just fine: NaTCH Engineering - SigJenny an Audio Signal Generator for Free
P.S. You're right, of course, that the closer you get the math/simulation, the better decisions you can make from the start. It is what it is
.
P.S. You're right, of course, that the closer you get the math/simulation, the better decisions you can make from the start. It is what it is
.
Last edited:
A digital multi meter will not directly measure Fb, though it is possible using dummy resistors to measure the speaker's impedance at various frequencies and infer Fb at the impedance minima.
Because meters are generally most accurate at 60 Hz, voltage has to be measured at each frequency to insure accuracy of the results, very time consuming, and lots of math.
I find it far easier to sweep a tone and look (or even feel) for the least cone movement, which is Fb.
Just curious, Weltersys, when you sweep what specifically are you using?
Thanks for your help
I use a BK Precision analog function generator making sine waves. I set the level so at the maximum excursion, which occurs about 1/3 octave above Fb, the speaker is moving less than rated Xmax.
Break in of the suspension is helpful, as the cone will move further than it will before break in with the same power, and make visual movement easier to see.
That said, Fb is Fb- it does not change even if the speaker's Fs or compliance does.
Art
A couple of years ago I did this same experiment with triangular corner ports. I built four identical boxes with new JBL 2235's and measured them oriented in all possible configurations. With all the ports centered together compared to all on the outside (ports and drivers vertical) so max to no boundary loading the difference was just under 3Hz. I measured with TEF driving a single Techron 7570 with a LinearX VI box.
Yes I was surprised.
Yes I was surprised.
The "true" length of a port is such a complicated thing that any of the numbers are going to be gueses. Some will likely be better than others, but which ones? Thats why it is essential to test the result in the end and modify from there. The idea of looking for minimum cone motion is a good one, but actually measuring the impedance curve is also quite trivial with a PC these days and that is far more accurate. The PC has put in our hands capabilities at home that exceed those of the very best a few decades ago. Why not use those, like HolmImpulse.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.