Closed end TL experiments, impedance flattening

[Dr.EM]

I'm sure there's abundant research online on the subject but I much prefer to experiment and discuss it myself, so I have been playing around with closed end transmission lines.

I used a small speaker:

Visaton - Lautsprecher und Zubehör, Loudspeakers and Accessories

chosen for its high and pronounced impedance peak. Of course it'd never sound great but all I'm interested in is taking impedance measurements from it.

So, I measured the free-air Fs at 292Hz. To build a 1/2 wave TL I cut some 2" poster tube (so cross section = Sd, or very close) to 589CM length. The driver is sealed into one end with blu-tack and the other end accepts the plug/bung.

I linked the driver to a little LM386 battery power amplifier (enough power for this) with a current sense resistor in series (1 ohm). The measurements you see were made with HOLMImpulse and are representative of the current through the driver (inverse of impedance magnitude).

Here's the graphs I came up with:



Here you see:

RED = the driver in free air
BLUE = the driver in the closed end 1/2 wave tube, unstuffed
GREEN = the driver in the 1/2 wave tube, stuffed

What I'm seeing is in the 1/2 wave TL without stuffing the original impedance peak has been flattened out, but replaced by 2 new peaks each of half the magnitude either side of the original. Why does this happen? When I had the MJK sheets I remember seeing something like this but I never really understood how to use them properly.

Adding stuffing flattens the overall impedance substantially. What I could not achieve though is a completely flat impedance, is it possible through this method? Adding too much stuffing seems to block the TL pathway and stop the original peak being nulled properly, the green line shows the best balance I could find. The green line, with stuffing, uses a slightly shorter tube as without making it shorter the null occurs slightly lower and is less effective, perhaps owing to slower sound propogation through the stuffing.

I can add some pictures of the setup if required, though it's not pretty! Hoping some people will find this interesting and can add some more input here. It's a topic of some interest on here but not greatly explored that I've seen. Thanks for looking

[speakerdoctor]

The test jig you've built seems something like a Kundt tube. Another experiment you can try is to play sine wave tones thru your speaker and slowly pass a sound level meter along it's length. You'll find dips and peaks that correspond to the sound wave's peaks and valleys that are present with the standing waves.

You might want to check out these links related to Kundt tubes:
Standing Waves in a Tube
Kundt's tube - Wikipedia, the free encyclopedia
YouTube - Kundt's tube ~287Hz

[MCPete]

Hi,
Are you saying that the end of the line opposite to the end where the speaker is attached is plugged? If so, that is contrary to the conventional practice of having the line closed at one end only by the driver. Perhaps that is the source of the strange impedance vs frequency graph? I haven't worked with the transmission line, but have some interest in it.

Regards,
Pete

[speaker dave]

I'm not sure what the benefit of a closed in line would be. You've lost the bass enhancement that an open ended line output would contribute.

Either way, as you overstuff you end up with no impact from the line. The 1/2 wave effect will be gone and you will have a damped sealed box.

Flat impedance was never the goal. Smooth and extended bass response with the minimum of cone motion is what you want, no matter what enclosure type you prefer. That won't come with flat impedance.

David S.

[speaker dave]

Originally Posted by MCPete

Perhaps that is the source of the strange impedance vs frequency graph? I haven't worked with the transmission line, but have some interest in it.

Regards,
Pete

Only strange in that it is upside down. Otherwise it is as expected.

David

[MCPete]

David,

Can you explain why the impedance peaks of the blue curve are expected? I would assume that Dr. EM would be interested to learn that also. -Don't know if that would be lengthy. I'd guess that it has to do with harmonics of the line.

Regards,
Pete

[planet10]

Originally Posted by MCPete

Can you explain why the impedance peaks of the blue curve are expected?

I found that curve unexpected. I'd expect a closed 1/2-wave line to have an impedance similar to a sealed box.

Mybe there was a leask somewhere.

The point of a closed TL is to be able to ensure that any reflection from the far end has passed thru all the damping at least once, the idea being to totally kill any reflections & timesmear back thru the cone.



dave

[SteveVaiRules]

Would the nautilus tubes that B&W use have a similar curve?


Matt

[planet10]

The picture i posted is a Nautilus test mule. The shipping product uses 4 tapered 1/2-wave TLs to damp internal reflections to close to nothng.

dave

[wcwarriner]

You may recall an article in the December 2003 issue of audioXpress by Cornelius Morton entitled "A Hegeman Subwoofer." In this article Mr. Morton describes the design and construction of a sub based on the multiple closed tube design of Stewart Hegeman. The purpose of the staggered tuning stubs was precisely to counter the reactance of the woofer at resonance. Mr. Morton's published measurements show a fairly flat resistive impedance through the resonance region. This not only presents a simple load to the driving amp, but also produces a well damped acoustic response. His report may be of interest here if you have that issue. Morton corresponded with Don Morrison on the original Hegeman design. Morrison utilizes the design in his commercial speakers Audio Products
Bill