Seeking Help from the Speaker Box Experts

[tstone]

I decided to make a very light, very compact 1x10 cabinet for use in low-volume applications with double bass. I selected an Eminence Basslite S2010 Neo for a driver. I input the speaker parameters into WinISD Pro, and it came out with an “optimum” cab of 27.7 liters, tuned to 46 Hz. According to the program, this would require a 4-inch round port 13.5 inches long.

I thought the cone excursion curve was a little unbalanced, so I decided to tweak the tuning to 48 Hz. I also felt a slot port would be more space-efficient, so I drew up a box with a rectangular port of the same area as a 4 inch circle.

I knew the port length would require some tuning, so I knocked together a quick prototype out of ½ inch MDF, screwed together. I tested the cab by running constant amplitude sine wave signals into it, using the signal generator in WinISD. I found the box frequency by putting a 4.7 ohm 5 watt sense resistor in series with the speaker and measuring the voltage across the resistor. I looked for the frequency at which I got maximum voltage, hence maximum current, hence minimum impedance.

I was very surprised when the box frequency turned out to be, not 48 Hz, but 53 Hz. I fooled around with bent ports and such for a while, but even with a very long port, I could never get the frequency below about 50 Hz.

At this point, I decided to go back to basics. My slot port was very narrow, and I thought that could be part of the problem. I figured the best-understood ports are straight round ones, and if I mounted the port external to the cabinet – remember, this is just a mockup – that I could vary the port length and volume without changing the box volume.

So I assembled the setup shown in the photo, using a 4 inch diameter 36 inch long mailing tube for a rear-facing port. I figured I would start with a ridiculously long length, then cut the port off a few inches at a time until I found the optimum length.

But with a 36 inch long port, I got an Fb of…wait for it…49 Hz!!! According to WinISD Pro, a port that long should put me at 29.7 Hz!

So I am really scratching my head wondering what the heck is going on. Some possibilities that have occurred to me:

* WinISD Pro is just not that accurate
* My box is not airtight – although I would think that would lower Fb if anything
* There is something wrong with the way I am making my measurement
* My speaker is grossly out of spec
* ???

I would welcome hearing any ideas anyone has on the subject, because at the moment it has me stumped.

[billfitzmaurice]

The problem is that's not how you measure impedance. You need to use a two resistor comparator set up, or an impedance measuring device such as the Woofer Tester 3.

[gumtownbassman]

I might be giving bad advice here, but i think the box really needs to be air tight around the joints and edges.
As for the port area inside of the box, doesn't that substract from the internal box volume/area calculations?
Does the box design require the port on the same side as the front of the speaker?
maybe i'm wrong, i just found your post interesting.

[tstone]

Originally Posted by billfitzmaurice

The problem is that's not how you measure impedance. You need to use a two resistor comparator set up, or an impedance measuring device such as the Woofer Tester 3.

OK, let me back off and say I'm not really interested in measuring impedance per se, I just want to determine Fb. How about an ammeter in series with the speaker, would that be better?

[Robby Hoinsky]

Originally Posted by tstone

OK, let me back off and say I'm not really interested in measuring impedance per se, I just want to determine Fb. How about an ammeter in series with the speaker, would that be better?

Originally Posted by billfitzmaurice

The problem is that's not how you measure impedance. You need to use a two resistor comparator set up, or an impedance measuring device such as the Woofer Tester 3.

Robby

[billfitzmaurice]

Originally Posted by tstone

OK, let me back off and say I'm not really interested in measuring impedance per se, I just want to determine Fb.

To do that you must accurately measure impedance.

[qts]

You have to add the volume taken up by the port to the netto volume of the cab (27.7 liters) that will make the box bigger.
To tune, you can also run the sine wave of 48Hz trough the speaker and watch the cone movement.
When it almost stands still at that frequentie, the cab is tuned

[tstone]

Well, based on the input I got here, I thought about it some more and decided that where I was going wrong was in treating the power amp as a constant voltage source, which it isn't. I didn't want to buy a $100 piece of software/hardware that I might use only once, so I went back to the original definition of impedance, the ratio in the frequency domain of voltage to current. I came up with the circuit shown below.

First, at a given frequency, I measure V2, the AC voltage across the driver. Then I move one lead of the voltmeter and measure the AC voltage V1 across a sense resistor. Since R is known, I can calculate current i=V1/R. From those values, I can calculate the impedance at that frequency, Z(f)=V2/i. The process is repeated at each frequency of interest.

It's a little time-consuming, but I'm not in production, so that's fine. I'm getting much better values now. With my super-long external port, I'm measuring an Fb of 31 Hz, very close to the 29.7 Hz predicted by WinISD. Now I can start chopping down the port to tune it up.

If anyone sees any flaws in this approach, please let me know. I'm a mechanical engineer, not a EE, so I could still be missing something.

Meanwhile, thanks to all who tried to help.

[rad bassman]

interesting thread, good luck with your design, keep us posted with progress

[billfitzmaurice]

Originally Posted by tstone

If anyone sees any flaws in this approach, please let me know. .

It won't work either. You need a dual resistor rig, one at least 1kOhm, one 4 to 10 ohm. The design for it can be found in many speaker sites and books.

[Mr. Foxen]

I think you might be better off figuring your port tuning by the driver excursion method, face the driver up and put some rice grains on it so you can see how it is moving an sweep a frequency through it around your estimated tuning till you find where it moves least, and that is where the port is loading it.

[bgavin]

Originally Posted by tstone

OK, let me back off and say I'm not really interested in measuring impedance per se, I just want to determine Fb.

You can do this the mechanical way.
Put the speaker on its back, face up.
Drop salt granules into the valley where the dust cap meets the cone.
Run up your signal generator and watch for the least bounce.

A more accurate method uses an RTA (True Audio: Audio Spectrum Analyzer and Loudspeaker Design Software).
Take a simple nearfield measurement and look for the dip.
Place the mic a mm or two away from the cone and run the sweep.
There will be a decided notch in the response, where the driver stands still.
The mic is so close, it only hears the cone, not the vent or anything else.

Link to JBL SF15 Nearfield

Note the width of the trough.
This is also visible in the WT3 testing device.