Converting ports

There is nothing wrong with a round port, and if you can get it to fit, it should work OK. There are successful loudspeakers with all manner of shapes, so it is possible to make any shape work.

The problem is that ports are inherently nonlinear (unintuitively, a flared port can generate more distortion at high levels than a straight port), and unless you have some knowledge of acoustic design, by which I don’t mean being able to copy the T/S parameters into some software package and click the go button, it is probably best to stick with shapes that have been tried and tested. Round is a tried and tested shape.

The tuning of a port, flared or otherwise, is a function of the port cross sectional area to port length ratio. Predictive software simply works out the diameter, or wall lengths, if rectangular, based on the cross-sectional area of the port. You can do this manually for any other shape. The end correction modifies the effective length by some fixed amount and is needed because the radiation impedance of a port is not zero. Finding the end correction for flared ports is difficult, because of the many possibilities to vary the flare amount which changes the inertia of the flow at the port exit. Even if there is a box to tick in the software choosing a flared port it more than likely is just an estimated/guessed amount.

One advantage of round ports is that there are a lot of plastic pipes available, and it is relatively cheap and easy to change them out to fine tune the resonant frequency. With a 3D printer it should be equally easy change port designs for any port shape and/or flare amount.

On a positive, note while slight variations in tuning might make response changes look severe in software simulations, in real life you might not hear much difference. So, if having a rounded end to your port makes you happy, then go for it. It has been shown that with flared ports the tuning is mainly dependent on the minimum cross-section rather than the maximum. Because you are essentially only rounding the end of the port ignoring the routed round over, and treating it as though it is straight, should get you close enough as a starting point.

FOO

Instead of worrying about the port, you should confirm if the woofer will give you low frequency extension you are aiming for. When port noise is louder than the actual loudspeaker propagating sound, it usually means the user is trying to achieve frequencies in which, the woofer's voice coil cannot produce.

The Eighteen Sound 18 LW 1400 is a PA Woofer. Based on the woofer's (TS Parameters) calculations, 55 Hz (-3dB) is lowest you can achieve before the woofer's voice coil stops producing sound... At least anything that offers some means of significance. The piston will still be active nonetheless using every bit of the xmax trying to make up the loss. However, the sound is determined by the voice coil not, the xmax. This is why, xmax is not accounted for in the TS Parameters calculations.

Best Regards,   

You will need to start measuring the dB versus Frequency levels of woofers in free air.

All loudspeakers have a frequency limit in which, the dB level will decrease. You can feed the woofer 20 Hertz. The xmax will move 20 cycles per second however, it does not mean the voice coil is producing anything significant dB wise @ 20 Hertz. The end result is hearing all types of mechanical noises and very minimum sound at 20 Hertz. Then, wondering why the port noise is louder than the speaker once placed in a reflex box at the given frequency.

I've conducted such tests decades ago and invested heavily in a lot of raw speakers that confirmed what Harristech Software predicted. From my experience, Harristech is the only software company that offers such a feature that allows you to determine the -3 dB roll off point of any TS Parameters calculated in it's software.

Remember -3dB is half the output you will get out of the amplifier. So 1000 watts is now 500 watts. The remaining 500 watts is wasted in heat all courtesy of the voice coil in addition to kissing your efficiency goodbye.

Best Regards,

Elliot Thompson wrote: You will need to start measuring the dB versus Frequency levels of woofers in free air.

fudge22 wrote:   While measuring a loudspeaker in free air can be useful to determine the electro mechanical parameters of a loudspeaker driver, if the manufacturer has already provided the parameters, I’m not sure of any benefit in doing so (unless you don’t trust them). I’ve never seen a practical loudspeaker that has not been mounted on a baffle or in some sort of enclosure. Given that the enclosure will affect the response, surely it is better to measure the response of the combined system; especially as the total acoustic output from a reflex loudspeaker combines both the output from cone and the port.

None of that can attribute when making a comparison of frequency response vs dB. If the speaker cannot reproduce a frequency with great efficiency, no amount of xmax can make up the loss without mechanical noise being the overriding objective. The benefit of measuring a loudspeaker frequency response in free air is no frequency restraints due to the size limitation of the enclosure.  

Elliot Thompson wrote: All loudspeakers have a frequency limit in which, the dB level will decrease. You can feed the woofer 20 Hertz. The xmax will move 20 cycles per second however, it does not mean the voice coil is producing anything significant dB wise @ 20 Hertz. The end result is hearing all types of mechanical noises and very minimum sound at 20 Hertz. Then, wondering why the port noise is louder than the speaker once placed in a reflex box at the given frequency.

fudge22 wrote: I’m not sure whether you don’t understand how xmax works, or whether you are trying to explain it in a way that I don’t understand. First you say that xmax tries to make up for the lack of output below the -3dB point and now you say that xmax moves at 20Hz. It comes across that you believe xmax is only relevant below the -3dB point on the transfer function, and is used to compensate for the drop off in response; thus in the bandpass region of the response it is irrelevant. Xmax is a single value, with the units of mm; it is the maximum amplitude that the cone can move before the output becomes significantly nonlinear. The value is either calculated or measured.

The key word is tries! However, it cannot produce a frequency in which, the voice coil is incapable of delivering with great efficiency without adding mechanical noise in the equation. The confusion may lie as you are looking at a theory whereas, I am looking at things from a real-world scenario.  

fudge22 wrote: As stated previously, xmax is not a TS parameter. When calculating a response, either manually or using software, the TS parameters don’t consider the signal level. It is a bit like the formula used to calculate the frequency response of an electrical filter. You can calculate the values of the components to give the desired response or cut off frequency, but the values tell you nothing about the voltage and current needed. A graph just shows the output response, relative to a reference level, with a constant input signal.   If you have two identical drive units, except for the value of xmax, the calculated response for a given enclosure will be identical. In real life, at low levels, the measured response will be the same, however, the loudspeaker with the lower xmax will not be able to get as loud. If you take the 18sound drive unit mentioned in this thread, and limit xmax to 4mm instead of 9mm, the transfer curve will be identical, however the maximum output obtainable will be less up until 125Hz. At 65Hz the maximum output will be 7dB down.

I am not talking about two identical drivers utilizing identical TS Parameters with different xmax. I am talking about finding the right driver for the job based on the TS Parameters. If one wants 30 Hz, it would be better to use a driver that can achieve 30 Hz flat based on it's TS Parameters and not a driver that is - 3 dB @ 55 Hertz based on it's TS Parameters.

Elliot Thompson wrote: I've conducted such tests decades ago and invested heavily in a lot of raw speakers that confirmed what Harristech Software predicted. From my experience, Harristech is the only software company that offers such a feature that allows you to determine the -3 dB roll off point of any TS Parameters calculated in it's software.

fudge22 wrote:   Congratulations on your decades of heavy investment and testing. However, I suspect that you are in a minority if you believe that the Harristech software is the only software that allows the response, including -3dB point to be calculated from TS parameters. There are other programs that can model the response quite accurately.

As I stated, from my experience. If there is something on the market today that offers such a feature, I am more that willing to make a comparison to Harristech. However over two decades ago Harristech was the only one to offer such a feature. The free programs used by many today does not offer the feature as Harristech I am referring to.

Elliot Thompson wrote: Remember -3dB is half the output you will get out of the amplifier. So 1000 watts is now 500 watts. The remaining 500 watts is wasted in heat all courtesy of the voice coil in addition to kissing your efficiency goodbye.

fudge22 wrote: I’m not sure there is any easy way to tell you this, but the maximum efficiency for an 18” speaker at 40Hz is 4% (see Keele -Max efficiency of direct radiator), so very little efficiency to kiss goodbye. Of the 1000 Watts coming out of your amplifier 960W are wasted as heat in the voice coil. Fortunately, it doesn’t take much acoustic power to get loud.

In loudspeaker frequency response measurements, there is a starting point. That is what many use as reference or a guideline to determine what is the starting point (insert dB here) to the -3db roll-off point. So, if you sweep a frequency starting @ 100 Hz (The majority do not cross reflex subs higher than 100 Hz) downwards, @ 1 watt in which, it is -3dB @ 55 Hz, the speaker is delivering 0.5 watt @ 55 Hz, despite the speaker delivered 1 watt @ 100 Hz.   

Best Regards,