Hello,
I'm building two subwoofer boxes. I'm using 18NW100 8ohms. When I measured the impedance curve of the set (box/subwoofer) I noticed that the tuning was lower in frequency than the subwoofer resonance. So I decided to measure it alone. And it gave me a resonance at 41.7hz when the datasheet says 31hz. I don't understand. Any ideas? thanks!
I'm building two subwoofer boxes. I'm using 18NW100 8ohms. When I measured the impedance curve of the set (box/subwoofer) I noticed that the tuning was lower in frequency than the subwoofer resonance. So I decided to measure it alone. And it gave me a resonance at 41.7hz when the datasheet says 31hz. I don't understand. Any ideas? thanks!
Optimistic data sheet? Change in manufacturing? Lack of break in maybe? I'm inclined to believe your measurements first (assuming they're done well), over data sheets.... In my limited speaker design/build work, I've always measured the drivers, and used my numbers in final design....
The Fb (frequency of box tuning) is independent of the driver's Fs, in other words, it will be whatever it is regardless of the driver used.
Although preconditioning the driver, testing at higher levels and suspension break in will lower the measured Fs, there seems to be a problem with the current 4 ohm version of this particular B&C driver as linked in post #4.
I'd suggest contacting B&C Speakers Sales & Operations Director, Bennett Prescott, he likes to keep customers happy.
Art
Thanks. I'm new to this and didn't know that Fs reduces with use. The speakers are very lightly used. But 31 (datasheet) to 41 (measurement) Hz difference in Fs seems like a lot.
I have bought many woofers. Most of them have measured parameters that are different from the manufacturer's data sheet. It usually comes down to the suspension being stiffer, so the Fs is higher. If your box is ported, then the port tuning will be what you design it to be. I expect a few minutes with WinISD and loading the parameters you measured will allow you to sort things out. It could be that after a few months of use they will loosen up. It is super frustrating to size the box to get low bass and then get the woofer only to find that you have no low bass and a midbass hump in the response. This happened with my latest build using Tympany 10" woofers. I had to put a notch filter in the crossover to tame the 50 Hz bump. I usually build sealed woofer boxes and then just use the Linkwitz Transform ( 2nd order asymmetric shelf filter ) to get the in box F3 and Q I desire.
Last edited:
I tuned the enclosure to 37hz, as recommended by the manufacturer.
I should have built the enclosure after measuring. I trusted the datasheet. I will modify the ports taking into account that the Fs is 41.7hz. What Fb do you recommend? Will 45hz be fine? Thanks
I should have built the enclosure after measuring. I trusted the datasheet. I will modify the ports taking into account that the Fs is 41.7hz. What Fb do you recommend? Will 45hz be fine? Thanks
You could leave it as is, play the speaker for a while. Measure again in few months. To explore the response of different tuning I always use software. I like the freeware WinISD. http://www.linearteam.org/ You can add your woofer into the database and type in the measured TS parameters from Limp, DATS or REW. I found the closed box or added mass methods worked well. I was using the cone mass from data sheets and got wildly wrong results because some data sheets have the cone mass way off. Read the help file to see the procedure. They involve using the set reference command.
With WinISD you can compare several projects with different tuning at the same time.
With WinISD you can compare several projects with different tuning at the same time.
Fs is not a fixed parameter, the value typically depends on the load. Here is the JBL 2216Nd factory measurements for example:
Large signal parameter values are more representative of real-world usage than small signal values.
If you want extension down to 37 Hz (and slightly below) with the 18NW100, you won't get that if you tune to 41.7 Hz. Fb of 37 Hz would be fine even if the Fs is higher IMO. By the way, the factory recommendation is just one of many possibilities.
Large signal parameter values are more representative of real-world usage than small signal values.
If you want extension down to 37 Hz (and slightly below) with the 18NW100, you won't get that if you tune to 41.7 Hz. Fb of 37 Hz would be fine even if the Fs is higher IMO. By the way, the factory recommendation is just one of many possibilities.
Last edited:
Whenever I measured Fr of a driver, it was significantly higher than the date sheet states.
B&C do their measurements after some burn-in with high excursions.
i tried the same here and after an overnight burn-in came to similar results.
One day later I repeated measurements -
and Fr was a high as it ever was.
Since then I do not believe any Fr data from manufacturers.
B&C do their measurements after some burn-in with high excursions.
i tried the same here and after an overnight burn-in came to similar results.
One day later I repeated measurements -
and Fr was a high as it ever was.
Since then I do not believe any Fr data from manufacturers.
If one were to measure the T/S parameters with the typical measuring jig + PC, these would be the small signal values.
That is simply one alignment of any that could be chosen.
The recommendation is based to provide a reasonably flat response (-3dB at 38.7 Hz) in a relatively small 160 dm3 (5.65 ft3) enclosure suitable for PA use, where driving to Xmax/Xvar (or more) is common.
If you use the speaker with over 550 watts (excursion limit in the 40Hz range at around 123 dB), in use it's Fs will drop to near the published specifications.
Most PA operators tend to to use far more than that much power, hence the stiff suspension that reads high without preconditioning.
If you want usable response to a bit below Fb, leave the ports as built.
You will want to high pass the enclosure a few Hz below Fb, the impedance minima.
If you want more output around 45Hz, shorten the port to raise the Fb, but the cabinet won't reproduce the low fundamentals of a 4 string bass guitar (~41Hz), much less a 5 string bass (~31Hz).
I made this transfer function measurement with Smaart Live 8 at a venue with an EMC8000. Without filters, do you think it's okay?
For an in room response, your response looks reasonable, 30Hz is ~-8dB from the average level of 60-200Hz, similar to this chart:
That said, as well as the obvious room deviations, your ECM 8000 may add another 3dB of variation:
Have you run the driver past Xmax (18mm peak to peak) yet?
That said, as well as the obvious room deviations, your ECM 8000 may add another 3dB of variation:
Have you run the driver past Xmax (18mm peak to peak) yet?
Last edited:
You can push the cone by hand or with amplification.
Music, pink noise, or sine waves may be used.
Below Fb, at around 27Hz, won't take much voltage to exceed Xmax, and won't be very loud, nor will there be much power dissipation, since the impedance rises from around 8 ohms at Fb to above 25 ohms.
Note that at Fb, the impedance minima, the excursion will be least- the voice coil will burn up before you see much movement.
Music, pink noise, or sine waves may be used.
Below Fb, at around 27Hz, won't take much voltage to exceed Xmax, and won't be very loud, nor will there be much power dissipation, since the impedance rises from around 8 ohms at Fb to above 25 ohms.
Note that at Fb, the impedance minima, the excursion will be least- the voice coil will burn up before you see much movement.
There is a way to use amplification with a PC and REW to measure large signal parameters.