Design Review: Tang Bang W5-704s Woofer

[69stingray]

Hello Everyone-

I have gotten a design I think I like for Tang Bang’s W5-704s with a Dayton PT2B-8 Planar Tweeter. I already have the tweeter and was looking for a woofer that would be able to be crossed over at 4,000 Hz. I used the FRD Consortium tools. Traces are from Zaph's measurements.

I have to go back and re-run all of my simulations again, just to double check everything. I have played with some many internal volumes, tunings, baffle sizes and crossovers, I like to “rebuild” from scratch with what I think I settled with.

Here is where I am at. The internal volume is 17 Liters net. Baffle is 9-12/16”x14-11/16” with 3/4” edges. Woofer has an LR contour network for baffle step, a Zobel network and a 4,000 Hz lowpass filter. The tweeter has an approx. 15 dB L-Pad and a 2nd order electrical high pass filter at 4,000 Hz.
I have included some graphs from PCD and an Excel version with +/-1.5 & +/- 3 dB limits for reference.

THOUGHTS?

Excel Plot of Predicted Freq. Response:


PCD Plot:


Reverse Connection on the Tweeter:


Crossover:


REFERENCE THREAD:
Tang Band w5-704s

[454Casull]

Not going to notch out the peak at ~1.7kHz?

[FullThrottleRic]

It looks like the box is too big for the woofer looking at the droop followed by a big peak.
I'd put a notch filter on the tweeter's resonance and increase the slope of the tweeter, the phase tracking isn't too good looking at the tweeter reversed simulation.
Try oversizing the 0.37mH woofer inductor and adding a resistor in series with the 8.42uF capacitor to take care of the baffle step, less inductors in series with the woofer that way

[sreten]

Hi,

The box is far too big for the Vas of the driver, the driver does not suit vented loading.

If your going to use it you should go sealed and add an active subwoofer.

Your crossover is far too complicated and does not work very well
in controlling the ripple issues caused by the box, generally you do
not need a separate baffle step element in the crossover.

Wait for Zaph to publish his W5-704S/SD1.1 design for some pointers.

/sreten.

[69stingray]

The more I look at it, the more concerned I am that there is going to be a lack of bass in 100-400 Hz range. We are only talking a 1.5 dB though…

It looks like the box is too big for the woofer looking at the droop followed by a big peak.

The box tuning has a ~3 db peak at tuning. I have decided to smooth it out a bit by lowering the tuning from 60 Hz to 50 Hz. If it had a perfectly flat response, it would be around 79 dB, so the rise you are seeing is the baffle step raise combined with the rise in the transducer itself.

Try oversizing the 0.37mH woofer inductor and adding a resistor in series with the 8.42uF capacitor to take care of the baffle step, less inductors in series with the woofer that way

I twill try that. What is the drawback of having inductance inline with the woofer? I am also going to try a 2nd-order slope. With the second order, I am hoping to smooth out some of the roughness in the 1-3 kHz range.

[quote] The box is far too big for the Vas of the driver, the driver does not suit vented loading.[quote]

I used Zaph’s measurements and the “design” volumes for both sealed and vented are huge. Sealed (w/ Qtc=0.707) is around 30 Liters (if I recall). At 17 L, a sealed enclosure at Qtc-0.74. Zaph send he is going to measure some more to see if there is a consistency issue. Zaph's measurements do differ from mfg. spec by a good margin.

If your going to use it you should go sealed and add an active subwoofer.

The speaker is going to be for my brother-in-law for his bedroom (he is finishing a large addition with a new master bedroom). They probably will not be played loud, but he is a fan of the bass, so I do want to try to stick with a vented enclosure. There is no plan for a sub.

Thanks for the pointers. I will give some a try and report back.

[augerpro]

If these FR's are not from the driver measured on the actual baffle how do you account for baffle step?

[69stingray]

I have been using the free FRD Consortium tools.

Start with Baffle Diffraction Simulator and then input into Frequency Response Combiner that combines the predicted box response, actual freq response on an open baffle, with the baffle diffraction to get a overall frequency response.

Here is a good writeup on the procedure: RJB Audio

[sreten]

Hi,

recommended alignments :5L to 10L sealed, 10L to 15L vented.

Anything that looks flat will be a serious boom box.

I reccommend 10L tuned to 40Hz, experiment with stuffing to tune bass.

/sreten.

[69stingray]

Hi Sreten,

Anything that looks flat will be a serious boom box.

Why do you say that? Room gain? Woofer excursion below tuning frequency? All of the sims you posted have humps in the response, will those be boom boxes also? Just curious.

I have been playing with Unibox and have gotten what looks like a respondably flat response. The response does gradually raise approx. 2 dB from ~200 Hz to ~60 Hz.

[sreten]

Originally posted by 69stingray
Hi Sreten,

Why do you say that? Room gain? Woofer excursion below tuning frequency? All of the sims you posted have humps in the response, will those be boom boxes also? Just curious.

I have been playing with Unibox and have gotten what looks like a respondably flat response. The response does gradually raise approx. 2 dB from ~200 Hz to ~60 Hz.

Hi,

None of them are going to sound wonderfully tight, the best 10L sealed.

Any attempt to get flat response below 100Hz will be a boom box
due to group delay at the port frequency if its say 50 to 60Hz, go
for 40Hz or below tuning in my book.

Use of room gain will bring up low bass if you keep the speakers
away from the walls, corners and floor, i.e. freespace mounting.

/sreten.