Tuning a Base-Reflex enclosure with a impedance curve

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Tuning a Bass-Reflex enclosure with a impedance curve

I recall back in an earlier day reading a Popular Electronics periodical about early base-reflex enclosures. The article published the impedance curve of a speaker in a sealed enclosure. Then the author began drilling a number of half inch holes in the enclosure. With each added hole the original impedance peak lowered in amplitude and increased in frequency. Also with the addition of new ½ inch holes there was a new lower frequency peak added to the impedance plot. With the addition of additional holes both impedance peaks increased in frequency. The lower frequency peak increased in amplitude and frequency. The higher impedance peak continued to reduce in amplitude and increase in frequency with addition holes in the enclosure.

This past Friday I took out a base-reflex enclosure, about 1.5 cubic feet with a JBL 2204H driver installed. The enclosure has a fair amount of polyester fill and two 1-3/4” holes drilled through the 3/4” front baffle. I plugged the holes and plotted the impedance. Then I unplugged one of the holes and appended the plot, then unplugged the second hole and appended the plot again. The AP plot is attached.

It looks like the enclosure could use one more 1-3/4” port to make the impedance peaks close to equal.

Thoughts

Thanks DT
 

Attachments

  • JBL2204 in 1,3 Ft3 enclosure, Impedance Magnitude (1).png
    JBL2204 in 1,3 Ft3 enclosure, Impedance Magnitude (1).png
    56.2 KB · Views: 811
Last edited:
Interesting the addition of each port seems to increase the resonant frequency, drop the impedance amplitude with some lower mode now opening up. Keep going - raise the amplitude of the lower resonance with a drop and shift of the higher one, bringing them closer together in level.
 
Hello,

Yes it is interesting.

Now years later with the simulators and design software I do not know what assumptions are built into the software.

I also strongly suspect that the larger manufactures do not stop the design process with software. I expect there is extensive lab testing to tune the design.

Looking at the impedance plots of a few commercial designs the higher Bass-Reflex impedance peak is most often a bit higher in magnitude than the lower frequency impedance peak. Rarely the lower frequency impedance peak is the higher magnitude one; interesting. Whatever their design goals are, I do not know.

Before it gets too damp and rainy in Nor Cal I will drag this Bass-Reflex enclosure out on the back deck to test impedance curves and frequency response.

Also just for fun I ran an impedance plot on a JBL2119H mid-range driver in a sealed 0.55 Ft3 enclosure. Plus there is a Theile Small / Parameter measurement of the speaker, that is driver installed in the box.

Later I will post the impedance plot of an JBL D2 driver mounted on a M2 Horn / Waveguide.

Thanks DT
 

Attachments

  • JBL2119H in box Impedance Magnitude .PNG
    JBL2119H in box Impedance Magnitude .PNG
    41.7 KB · Views: 715
  • JBL2119 in box Thiele-Small.PNG
    JBL2119 in box Thiele-Small.PNG
    27.2 KB · Views: 680
Last edited:
diyAudio Moderator
Joined 2008
Paid Member
Some look at the twin peaks as being side by side. Would it be better to see them as (something like) upper and lower as they act upon each other?
Also with the addition of new ½ inch holes there was a new lower frequency peak added
I think I didn't understand this one, surely you mean one changing lower peak, not one for each hole..
 
BR is a good example of how you can have an elaborate and sophisticated analysis and sim and the result is nonsense.

The classic theory is to create a box that counter-acts the driver resonance and kills it right in the middle. Therefore, the impedance curve was a perfectly good way to nail it.

But it is now obvious that the sonic output is a mishmash of direct and port emanations and the way they work together (and importantly, how they oppose one another below the tuning) is a matter of art. And that's all even before considering the how the cab sits in the room or what your "house curve" is like.

B.
 
Some look at the twin peaks as being side by side. Would it be better to see them as (something like) upper and lower as they act upon each other?

I think I didn't understand this one, surely you mean one changing lower peak, not one for each hole..

Hello,

Each additional hole incrementally adjusts the tuning of the enclosure. There are twin peaks, one measurement at a time. The effect of 20 holes sum together to make a single twin peak curve.

I suppose that you could put you fingers over the holes and play it like a finger flute.

finger flute guy - YouTube


Tin Whistle Lesson - Britches Full of Stitches (polka) - YouTube

Seriously we can tune the enclosure by varying the number of ports, the diameter of the ports and the length of the ports.

Thanks DT
 
Last edited:
It indeed appears that you'd need a fourth vent to equalize both resonant peaks. In this case I'd recommend a single port of double the diameter, but of the same length, to decrease possible chuffing noise. Why? As length and area would be the same in both cases, vent resonances and air velocity also would be the same. But as a cylinder of double the dia has only half the barrel surface, air friction will be reduced.
Best regards!
 
It comes out of bass reflex (not 'modern' T/S vented box) design, where a symmetrical double-saddle impedance curve theoretically indicated Fb = Fs. Whether you always wanted that was another matter of course and depended on your target alignment and box volume, just as it does today, and you adjust Av (and / or Lv) as desired to achieve that. Classic rule of thumb from bass reflex design: holding Lv static, reducing Av by a factor of 4 lowers Fb by approximately 1 octave.
 
Hello,

This all just for fun.

I had the old seat of the pants speakers that I used in the garage for years. Sitting on the bench there is a nice analyzer so I ran a couple of impedance curves for the fun of it.

I had the JBL 2204H’s because they were cheap. The “sub-woofer” boxes came from the car parts store. They were about the right size. The drilled ports are 1-3/4 inch because that is the size of the drill bit that was handy. The boxes have played for years and do not sound all that bad. Now I think that I will drill one more hole to see what difference that makes. That double saddle impedance curve reminds me of the old Girl Scout song, “Alice the Camel, has two Humps”.

Except this time I will use a calibrated measurement microphone that came from Audio Precision.

Guys thanks for all your thoughts.

Thanks DT

From a previous post you can see there is a mid-range and a wave-guide loaded HF driver to be added to the measurements.
 
Why the effort to make them equal??
It comes out of bass reflex (not 'modern' T/S vented box) design, where a symmetrical double-saddle impedance curve theoretically indicated Fb = Fs..
Yeah, that was out of my old boss Jim Novak's era. Then Thiele and Small came along (with Benson!) and recognized you could model speakers like electrical circuits. Then a lot of folks got way too minutely carried away with that, forgetting it was a MODEL. Then folks like Doug Rife (MLSSA) and Chris Strahm (R.I.P. :( LinearX (LEAP)) and Wolfgang Klippel and others gave us tools to examine things on a deeper level.

The shape of the peaks has to do with losses as well. Running tests on a sealed box, like making an impedance curve to find T-S parameters, any prescence of a lower peak indicates the sealing is not good. That's why I'd run curves way lower than "necessary"-something like an LMS would run down to 10 Hz but I preferred MLSSA running to I forget 1 or 0.1 because if the curve even started to lift at the bottom I would check my gaskets and clamps.
 
head_unit,
You put your finger on several of the topics I was thinking of. Thanks for bringing them up. The largest problem that I see so far in using Theile Small Parameters for calculating the Parameters of a vented woofer is that the TS/P are measured / calculated using low or small level signals rather than listening level or large level signals. Driver Parameters change as the test voltage increases.

Robh3806,
I am changing the way that I read your reply, question, just a little.
If the acoustic curve is "right" what does the twin saddle impedance curve look like?
No worries, this is for fun.

Scottmouse,
Thanks for the theoretical; equal impedance curve humps is: Fb = Fs .

I see a little fun coming up over the cool fall and winter months.
There is a 2 Ft3 box sitting here to mate with the JBL 2204H driver. I am thinking of bolting on an assortment of different size and shape tuning ports and looking at frequency response compared to the impedance curve shape, and or humps. Perhaps even stuffing some polyester pillow stuffing into the port for resistance. (Thinking of time constant stuff.)

An added thought;
Most of the impedance curves of commercial woofers I have seen show the lower frequency impedance hump a bit lower in magnitude than the higher frequency hump.

Thanks DT
 
It comes out of bass reflex (not 'modern' T/S vented box) design, where a symmetrical double-saddle impedance curve theoretically indicated Fb = Fs. Whether you always wanted that was another matter of course....

My intuition is the motion of the driver is better held in check with the "classic" tuning since that is tuned for driver resonance resulting in better impulse response. If so, SQ might be better.

What do you think?

Sims, like robo-advisors for investments don't do a good job determining what the customer values. Always seem to solve for flat response (which is not what anybody really wants for a sub).

B.
 
Last edited:
My intuition is the motion of the driver is better held in check with the "classic" tuning since that is tuned for driver resonance resulting in better impulse response. If so, SQ might be better.

Not as far as I know. Classical bass reflex design did not particularly target Fb = Fs, that was just a part of the design process, albeit a significant one, after which you generally tuned 'to taste' by adjusting Fb with the vent CSA or with vent damping via the old click test. An optimally sized box (note caveat) for given driver and drive impedance should have a slightly faster settling time than an acoustically undersized enclosure tuned to the same frequency, but that sets aside the actual frequency / amplitude alignment, and came with a bunch of other caveats. Most classical BR boxes (as opposed to the ducted vent type derived from electrical filter theory, first by Novak, and latterly by Theile, then Small) tended to require the aforementioned additional damping as they were otherwise known for ringing.
 
Hello,

If you have a tuned vented woofer on your bench I do not think that it knows that it was designed with filter software or not.

The click test that I recall reading was a 1.5volt D-cell battery switched across the terminals of the woofer. There were no capacitors in a crossover sending the high frequencies other places.

The focal topic I see is resonance, stored energy and damping. Higher frequency ringing need not apply. The power amplifier output impedance is a key factor in calculating the system Quality Factor.

Ideally the speaker should follow the input signal. It should stop moving when the input signal stops. Unfortunately the speaker has a certain amount of inertia making it slow to start moving and a certain amount of momentum to keep it moving when the signal stops. This inertia and momentum accounts for the measured phase shift and if you will, group delay.

I suspect that an impedance plot, if we take a close look, will tell us much of what we want to know about a tuned port woofer.

Thanks DT

An afterthought:

I am glad that mid-range speakers and tweeters do not have resonance peaks smack in the middle of their pass bands. (Yes, I know, horns and wave-guides are an exception.)
 
Ringing refers to the settling time / step / impulse response / GD of the combined LF system load, and has nothing to do with capacitors, HF drive units etc. That was what the click test was intended to help assess, and as necessary correct via judicious enclosure / vent damping. Amplifier output impedance, along with wire loop, connection resistance & any relevant DCR from series components in the LF load, notably inductors, should be incorporated into a quality enclosure alignment as a matter of course.
 
I picture ringing as ringing like a bell.

Like would likely be seen on an oscilloscope as oscillation at the edge of a square wave. In the case of the click test, oscillation when the 1.5volt D-cell is switched off.

A low output impedance power amplifier will negate much of the need for port damping.

In line inductors may help if there is ringing, as parallel capacitors may also help damp ringing. Now we are talking about a crossover.

I am not ready for a crossover yet.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.