Standing waves, or something else?

[nullspace]

Hi All,

I'm working on a new speaker, a two-way with Great Plains Audio 414-8B AlNiCo 12" in a vented box plus Great Plains Audio 802-8G Series II AlNiCo 1" compression driver on an Acoustic Horn Co. AH!1000 50deg conical horn.

I got a trial box built but had an unwelcome testing result, and am wondering if it's from an undamped standing wave or something else.

Here are the particulars:

• ~5cu ft tuned to 27Hz.
• Interior dimensions are 36.375" high by 21" wide by 12" deep.
• Driver is 7.25" from top and port is 4.625" from the bottom.
• The port is one of the PSP 4" dual-flared ports, and is located on the rear of the box.
• 1" of 8.5# 100% wool felt on top, back, and one side of the cabinet.
• All measurments with ARTA.

Here's a comparison between the impedance of the driver free-air (sitting on my dinner room table, facing up) vs. in-box. Box is measured in the center of the room, away from the walls/corners. Red vs. black, respectively.


Here's a close-up of the impedance glitch.


Lastly, here's the near-field response for the driver and port, smoothed 1/6 octave.


I would quickly chalk it up to standing waves, if there wasn't already something going on in the free-air impedance in the same ballpark. But, that measurment creates enough ambiguity for me that I thought I'd throw it out to the diyAudio audience. Any thoughts on this?

In cae anyone's interested, here's what the box looks like.

I got the idea for the construction here: Bracing overkill yay or nay ? I'm using 3/4" x 1.5" poplar. I chose poplar because it's almost as stiff as plywood, but is 3/4 the weight and relatively cheap.

Regards,
John

[flexus]

did you go through the room during measurement?
your parquet is made of wood, during impedance testing - tapping on the floor can cause these kind of ripples if its made with noise measurement, not swepped sine.

[nullspace]

Nope. I was seated in the next room the entire time. And, the fact that the frequency response chart correlates with the impedance ones leads me to think that it's not a result of measurement error.

Regards,
John

[speaker dave]

That looks real and normally I would blame it on the cabinet. It generally can give issues when putting a driver at the very top of a long cabinet. The confusing issue is that you have a milder version of it in your table top measurement.

If we assume that the table top measurement and its minor glitch is a fluke then the 180 dip would be a top to bottom standing wave. Your cabinet is lightly lined rather than stuffed and its dimensions aren't helping.

I'd recommend you try a fairly thick partition of fiberglass that bisects the cabinet half way up its height. This would put absorptive material at the velocity peak and give the best effect.

Try it and if it doesn't work then it is something else.

David S.

[nullspace]

Thanks. Too late I realized that having the height almost exactly 3x the depth was not going to be a good thing. However, I never imagined the impact could be of this magnitude.

I'll try your suggestion.

Regards,
John

[nullspace]

I had some time today, so I cut a piece of wool felt to fit horizontally in the interior. It's at about 60% the total height as I couldn't get it any closer to midway. The latest data is in blue.



Based on the impedance sweep, I'm comfortable chalking up the orginal resonance to standing waves. However, I suspect the 8.5# wool felt is a bit too dense based on the increased resonant frequency of the box with the added damping. I might try some of the Bonded Logic #6 acoustic cotton next. Comments?

Regards,
John

[Flaesh]

nullspace, have you another freeair measurments?

IIRC such issue (surround resonance) and its fixing by speaker manufacturer was mentioned somewere on AEspeakers forum.
and in some German magazine similar glitch was described as "Sikenresonantz" (like this term - sound funny on Russian)))

generally manufacturers ignore this issue. you can find a lot of curves in datasheets with peak on impedance and deep on FR..

it's interesting how it increase in box.

[nullspace]

No, I don't have any updated near-field frequency responses. The latest impedance sweep shows the absence of the resonance at ~180Hz, and that was all I was looking to establish.

The prototype has been a worthwhile exercise, and several issues have come to light. In regards to the issue discussed here, I'll be changing the internal dimensions to 34" x 21" x 13" for the next iteration. Approximately the same internal volume, but a better distribution of standing wave modes I suspect.

Regards,
John

[DBMandrake]

Originally Posted by nullspace

I had some time today, so I cut a piece of wool felt to fit horizontally in the interior. It's at about 60% the total height as I couldn't get it any closer to midway. The latest data is in blue.



Based on the impedance sweep, I'm comfortable chalking up the orginal resonance to standing waves. However, I suspect the 8.5# wool felt is a bit too dense based on the increased resonant frequency of the box with the added damping. I might try some of the Bonded Logic #6 acoustic cotton next. Comments?

Wool felt is very dense compared to the fiberglass suggested by speaker dave, also unlike fiberglass its non-porous and air-tight (more or less) so if you've bisected the cabinet horizontally with it and, I assume, have it largely suspended in mid air tacked just at the edges, it will be moving up and down like a bladder with air pressure changes with bass excursion. This will cause a lot of energy loss at bass frequencies, so I'm not surprised the Q has dropped so low.

Generally a dense non-porous material like wool felt would only be used attached to the walls of the cabinets as a lining - never suspended in mid air or used as a stuffing. In fact it needs to be tacked quite firmly to walls so that it doesn't flap about at bass frequencies and unnecessarily lower the Q.

Using a light, porous material like fiberglass or open cell polyester foam as a stuffing material is a different thing altogether, although you can't use too much in a bass reflex design or the efficiency of the port will be lowered a lot. (In fact I usually don't use any stuffing in a bass reflex box, only lining, keeping well away from the end of the port)

One thought I had about breaking up the vertical standing wave modes in the box without adding so much stuffing/lining as to ruin the bass response, is what if a couple of shelf partitions were added ?

For example 1/3rd of the way down from the top of the box a shelf is fitted in the front half, the full width of the box, but only half the depth of the box so the back half is an opening. Then at 1/3rd up from the bottom the same is done but this time the shelf is at the back wall with the front being a gap. Both shelves are lined with the same material as the cabinet walls, for example firmly attached wool felt.

Now there is no direct path from the top to bottom of the box, which should break up or at least redistribute the vertical modes, however there shouldn't be too much loss of Q at the box tuning frequency as the openings beside each shelf are still very large.

Note: I haven't tried this, but I have thought of trying it, has anybody actually tested an idea like this out ? Spacing the panels both 1/3rd from top and bottom may not be ideal, it might be better to make the spacing a little more uneven to further distribute the standing wave modes.

Just an idea

[DBMandrake]

One further comment - I just noticed the cabinet depth is only 12", that's pretty shallow for the height, if the cabinet is tall but very shallow that will tend to make the vertical standing waves worse, just like a long narrow pipe will have a "pipe resonance" that is stronger than an equal length pipe that is much fatter. See if you can google the equations/look up tables for pipe resonances that show the effects of different proportions on pipe resonances - I'm pretty sure they're applicable to the internal standing waves of cabinets as well, if the cabinet is long and skinny like a pipe.

Probably the best proportions for a cabinet (length/width/height) to minimize standing wave resonances would be to stay fairly close to the "golden ratio" between each axis.

It doesn't matter which axis is which, as long as the shortest length to middle length ratio is near the golden ratio, then middle length to longest length is also near the golden ratio.

This ensures both that standing wave frequencies of each axis are equally spaced, (so that they don't stack up at the same frequencies) and also that you don't end up with proportions that are too long and skinny. (The golden ratio will give a fairly squat cabinet shape)