Tapped Horns - Simulation VS Reality

[bjorno]

Hi,

FYI:More CSI observations It looks like the designer of this TH by a mistake? used a 2P = 2 drivers in parallel simulation seen at the TH input screen here above, but the finished TH-box has only one driver? See picture 1(3) where the suggested 2P driver TH is compared to using only one driver.

The termination (load) from the pipe looks ending too early and not fully(acoustically) enclosing the driver. Actually the implemented speaker looks like a slightly positive taper OD(offset driver) TL terminated at a driver diameter less than the total internal length = 259 cm and thus tuned above fs. This is IMO not a TH.

As can be seen at the excursion plot from about 50-60 Hz: only a very steep(> 24 db/octave) HP filter would protect the 2P drivers from over excursion: See picture 2(3) where also plots of more realistic max SPL is shown for the 2P driver TH and the one driver closed box at the left side of the picture.

The driver in use is IMO almost a text box example of a driver primarily well suited for a small closed box . As Qts is about 0.42 a box volume should stay within 2xVas at max.


Picture 3(3) shows a 2P (two drivers in parallel)T-QWP (tapped quarter wave pipe) compared to the proposed 2P driver TH. Note: the pipe diameter must then be less than 2xSd for both sides of the drivers in order to minimize pass band ripple when tuned at fs.

b

[Brian Steele]

Originally Posted by bjorno

Hi,

FYI:More CSI observations It looks like the designer of this TH by a mistake? used a 2P = 2 drivers in parallel simulation seen at the TH input screen here above, but the finished TH-box has only one driver? See picture 1(3) where the suggested 2P driver TH is compared to using only one driver.

Good catch. There actually is only one driver in the box.

Originally Posted by bjorno

The termination (load) from the pipe looks ending too early and not fully(acoustically) enclosing the driver. Actually the implemented speaker looks like a slightly positive taper OD(offset driver) TL terminated at a driver diameter less than the total internal length = 259 cm and thus tuned above fs. This is IMO not a TH.

Furthermore, the rear of the driver is actually loaded by a chamber. See diagram below:



How would you model this in HornResp?

[djk]

Did the spider to former glue joint fail?

With no deformation of the back end of the former, why did the coil come apart?

I see a hard rub, and a coil delamination from improper adhesive and/or cure.

Every system should have a cut-off filter below the tuning frequency.

Amplifiers need clipping eliminators, or soft clipping on their inputs (to prevent amplifier clipping from producing loudspeaker dynamic off-set).

I'm not sure about the 'tapped horn' being the blame here.

[davygrvy]

David, How does HR calculate max SPL? It doesn't happen to do sqrt(P*Z) where Z is a calc'd impedance does it? If so, that's a bit odd as power amps are constant voltage systems, not constant power. What I mean is, if at 110Hz Z is 9 Ohms @ -42 degrees, but rated impedance is 4 Ohms and is being driven by 500W amp, that's still 45V max output. To get 500W across 9 Ohms is 67V. The amp is already clipped, you aren't going to get there. There really would be a voltage ceiling as well to maxSPL.

[Lord Baccus]

Originally Posted by djk

Did the spider to former glue joint fail?

With no deformation of the back end of the former, why did the coil come apart?

I see a hard rub, and a coil delamination from improper adhesive and/or cure.

Every system should have a cut-off filter below the tuning frequency.

Amplifiers need clipping eliminators, or soft clipping on their inputs (to prevent amplifier clipping from producing loudspeaker dynamic off-set).

I'm not sure about the 'tapped horn' being the blame here.

No the Glue did not fail / looks as if the coil started to unravell from the bottom hitting something in the gap...

[David McBean]

Originally Posted by davygrvy

How does HR calculate max SPL?

Hi davygrvy,

Hornresp calculates normal SPL response using a constant input voltage Eg. Maximum SPL response is calculated using a constant input power Pmax (except when the Xmax limit is reached at any frequency).

The attached screenprint of the sealed box example discussed earlier, illustrates the difference.

The gray trace shows the SPL for a constant input power of 500 watts. (Xmax set to 99.9 mm to ensure that the response is not displacement limited at any point). The black trace shows the SPL for a constant input voltage of 40 volts.

Note that the diaphragm displacement at 40 hertz for an input voltage of 40 volts is 13.4 mm, whereas the diaphragm displacement at 40 hertz for an input power of 500 watts is 29.0 mm.

As I said in my previous message, it is important to understand the difference between input power and input voltage :-).

Kind regards,

David

[djk]

Thanks for the close up of the coil damage.

I agree, it looks like it bottomed out on the backplate.

Dynamic off-set can destroy woofers in the blink of an eye. It is very important to hard limit (or clip) the input signal to the amplifier so the amplifier feedback loop never clips.

I used a 1KW amplifier on a sub tuned to 23hz with about 6dB of boost. It would bottom out the instant the amplifier clipped. The same sub could be bottomed out by a 100W amplifier driven into clipping. You could see the cone jerking back-and-forth in an irregular maner (it lookeed like it was between 2hz~8hz), even though there was a filter at 23hz (part of the EQ for the sub).

[Lord Baccus]

I have the signal set with an O-Scope / NO SQUAREWAVES
using an Alpine 9887 ( Imprint ) & Kove Audio Aphex Ax1 Signal Processor, have a CLEAL 6v through the Rca's / Clean Signal on the Oscope...
Gain on the amps is set on the floor maybe 1 notch up from Zero (0) / no bass boost

please keep in mind that i am only a Weekend SPL Warrior / during the week i have an SQL Enclosure loaded

[bjorno]

Originally Posted by Brian Steele

How would you model this in HornResp?

Hi,
I would model this outlay almost like a straight pipe where S1 and S2 have constant equal CSA, L12+L23 set to the first chamber length using the TH Wizard.

As the termination opening is IMO too large for a 'tapped purpose',I would end the length calculations as for an 'OD' system model ( See the picture) that is invoked at the main input screen.

A length reduction, about 9" i.e. Total L = L12+L23+L34+L45 -9" is then necessary as seen in the picture.

After returning from the TH Wizard screen to the main input screen in order to
calculate an 'OD' speaker the red colored label 'TH' must be changed to a red 'OD'.

Then if the calculation button is pressed, the driver to 'terminus' distance must be inserted pressing 'Tools' and from the drop down list select the 'Combined
Response...' and enter for this case a distance of about 4.25" or about 11 cm.

b

[Brian Steele]

Originally Posted by bjorno

Hi,
I would model this outlay almost like a straight pipe where S1 and S2 have constant equal CSA, L12+L23 set to the first chamber length using the TH Wizard.

As the termination opening is IMO too large for a 'tapped purpose',I would end the length calculations as for an 'OD' system model ( See the picture) that is invoked at the main input screen.

A length reduction, about 9" i.e. Total L = L12+L23+L34+L45 -9" is then necessary as seen in the picture.

After returning from the TH Wizard screen to the main input screen in order to
calculate an 'OD' speaker the red colored label 'TH' must be changed to a red 'OD'.

Then if the calculation button is pressed, the driver to 'terminus' distance must be inserted pressing 'Tools' and from the drop down list select the 'Combined
Response...' and enter for this case a distance of about 4.25" or about 11 cm.

b

Hmm... will try that when I have a few minutes to spare...

I forgot to include that the internal height of the box (not given in the picture) is 11".