The Small Bangs: plans for a little 2nd order MTM speaker

[fjhuerta]

At first sight, the curves seem to be 2nd order already. We'll see.

Here's the impedance for both woofers in series.



I assume this means no copper shorting rings, etc. Not a good sign, huh?

[fjhuerta]

The usually well behaved SEAS 27TBFC/G.



So far, I've managed to get pretty good results with a simple 4th order acoustical crossover, and a more complex 2nd order acoustical. I believe the 2nd order is a better option due to its on and off axis FR (a bit flatter). But, with less components, and a lower X-over point, the 4th order isn't looking bad at all.

4th order acoustical, L-R @ 1.6 KHz.

[fjhuerta]

Transfer function.



FR (with a little flare at 6.1 KHz, couldn't get rid of it without a more complex circuit)

[fjhuerta]

2nd Order

This is the second order crossover. Far more complex than the 4th order one.



There was no way a simple coil and cap would do the trick. The woofer's rollover is already a 2nd order slope up to a certain point, then it has a dip, then a break-up. After trying a couple of options, I ended up with a series shunt filter, which acts as a BSC circuit. From there, an LR circuit with a resistor in parallel shapes the response from about 2 KHz on down. The end result is a practically perfect 2nd order woofer slope.



I believe distortion testing proves I can use a 2nd order on the woofer at 2.6 KHz without that many issues, so I may do it.

Anyway, the tweeter does behave very nicely (as always), and will do a 2nd order 2.6 KHz acoustic filter with a simple 2nd order electrical filter. BUT - phase isn't correctly aligned. By using a ladder delay network, like Zaph's (actually, with the same values!) this problem is fixed.



This is the FR. The flare at 6.1 KHz is gone.



And the transfer function.



Pretty strange looking, but that's because the woofer's FR is strange to begin with!

Any suggestions on which filter to build? The 4th order seems to be very close to the 2nd order (I'd assume because of the woofers' size - they are very small and crossed very low). But then again, apparently, 2nd order filters sound better.

[fjhuerta]

I've been playing around with the designs.

I'll be building the 2nd order version. I optimized the filters a bit more. I'll be posting them later.

[fjhuerta]

There has been no interest in this design, but FWIW - it's finished. I'll be listening to them in a while. Simulations are excellent, and even though the x-over is quite complex (for a 2 way) it fit inside the smallish box.

I'll take some pics when they are up and running.

[fjhuerta]

In the end, everything worked out perfectly!

The speakers are amazing. Tang Band every now and then delivers a great driver - I love these W4's! I don't know how many people would pay +$65 for them considering their sensitivity and size, but since s/h was pretty low, they may be an option for customers outside the USA.

I'll let measurements speak for themselves... other than a suck-up at around 2 KHz due to room effects (visible only on the in-room measurements) the speakers measure pretty good.

I'd love to have some help regarding their distortion profile, though.

[fjhuerta]

Steps

[kevinp.]

nice work Javier! What is that, felt around the tweeter?

[fjhuerta]

Hi Kevin,

It's synthetic felt (aka "polyester", I suppose ) I got it from Madisound. It hurts more than it helps, though - I noticed a bit more diffraction after I applied it. But ripping it out is out of the question now - I'd damage the baffle finish. Live and learn.

http://www.madisound.com/catalog/product_info.php?products_id=770

[fjhuerta]

I tweaked the 2nd order filter a bit. I'll be posting the design in a few moments.

This will be the definitive version. The project was a success!

[fjhuerta]

By lowering the value of the cap on the tweeter crossover, and lowering the resistor before it, I managed to tilt the tweeter response upwards, therefore flattening it as much as possible up to 20 KHz. this has an effect on impedance that will be seen later.



I took various measurements (1m, 1.5 m, 15° and 30° off axis) and optimized the ladder delay network manually until I got the closest phase match on woofer and tweeter.



Since I basically decreased resistance on the network on the upper response of the tweeter, in order to flatten the tweeter frequency response, impedance drops the higher the tweeter goes. It won't send any amplifier running for cover, though - the impedance is very bening.



The transfer function. Kinda strange looking, but it works.



The reason why there are so many parallel caps and resistors is a) for added peace of mind regarding power handling, and b) somehow, ordering smaller caps lowered the price of the crossover significatively.

[Eric89GXL]

There has been no interest in this design, but FWIW - it's finished...

I am interested in it. For some reason, the forum just let me know today that there have been new posts for the last month. They look great!

The FR is surprising, and it doesn't look particularly close to the one on the TB website. Is your mic calibrated?

The distortion test looks strange, as well. How are you measuring the distortion? John Krutke of www.zaphaudio.com has outlined several of the considerations/techniques he uses in his measurements, for example.

Eric

P.S. Hopefully next time the forum will let me know /promptly/ when there's a new post in this thread...

[ThomasW]

Hi Kevin,

It's synthetic felt (aka "polyester", I suppose ) I got it from Madisound. It hurts more than it helps, though - I noticed a bit more diffraction after I applied it.

The hole in the felt, it shouldn't be a circle it should be a square or diamond.

[fjhuerta]

I suppose it should, Thomas. FR is a bit worse (not much, but noticeable in the upper registers).

The sad thing is, taking the felt out would damage the speaker finish at this point. So I'll just live with it.

They do look nice, though.

[fjhuerta]

I am interested in it. For some reason, the forum just let me know today that there have been new posts for the last month. They look great!

The FR is surprising, and it doesn't look particularly close to the one on the TB website. Is your mic calibrated?

The distortion test looks strange, as well. How are you measuring the distortion? John Krutke of www.zaphaudio.com has outlined several of the considerations/techniques he uses in his measurements, for example.

Eric

P.S. Hopefully next time the forum will let me know /promptly/ when there's a new post in this thread...

Hi Eric,

The microphone is calibrated. Actually, the FR of the TB woofer is very close to TB's own specs. But they choose their vertical range so that any driver of theirs will look flat on it. Every TB driver I've used (W3-1335sb, W4-1337, W4-1658, W4-657, W4-1320SC) does this to some degree.

The distortion measurement isn't too surprising at this point, either. I did it in a noisy environment with a so-so sound card. I'll repeat it at a much quieter spot with an SB Audigy soon.