Flex Your PCD Mettle:

[Zilch]

Re: The Bloom

More proof diffraction off of the baffle edge affects both woofers and tweeters. In this case, the woofer especially. (Move those KDT baffles forward guys.) The directivity is getting so progressive and smooth throughout the frequency range... what remains to be done?

Flush-mounting the waveguide and rounding over the edges themselves. I believe we have incrementally demonstrated that while commonly known "best practice" may only minimally affect on-axis response of waveguide designs, its value in achieving smooth power response is clear -- getting the directivity "right" is fundamental....

I'm not ready to throw in the towel quite yet. ;)

[Zilch]

Re: Flex Your PCD Mettle:

Oh...verticals.

Yah, those:



Midpoint is the red curve; forward axis is presently +4.55°, the yellow curve, supra, approximately. Here's the forward axis response:




This driver combination is a challenge, what with a 16-Ohm woofer and a high-efficiency 8-Ohm (113 dB spec) compression driver, a wide disparity, as illustrated in the raw driver response curves:



Achieving those nice system response curves in the first couple of attempts came with a price in yanking down the HF hard. Look at the PCD sim system impedance here:

http://techtalk.parts-express.com/showthread.php?p=1669327#post1669327

It dropped below 4 Ohms at 5.2 kHz and ended up at 3 Ohms at 20 kHz.

I've since reworked it, and here is the measured result in comparison:



It just took an additional resistor and some component value tweaks to keep it above 8 Ohms across the entire spectrum. I'm gonna try to move the forward axis down a couple of degrees before calling it a "wrap." Extra effort with these old Altec midwoofs, 'cause they're sounding mighty tasty....

Footnote: 16-Ohm 4555s and diaphragms are available; I just don't have any.

[NEO Dan]

Re: Flex Your PCD Mettle:

Interesting Zilch,
Availability being what it is on the altec, are there any other candidates for mid-woofer that you would recommend?

[Skywave-Rider]

Altec

I recently called in to GPA for prices on new Altec (GPA) 414 woofers:
Pricing from GPA minus shipping:
414 Alnico: $330.
414 Ceramic: $240.
Both 8 or 16 Ohm.

[Zilch]

Re: Flex Your PCD Mettle:

Interesting Zilch,
Availability being what it is on the altec, are there any other candidates for mid-woofer that you would recommend?

Vintage alnico ones like the ones I'm using here seem to be going for ~$300/pair. Altec made both consumer and pro products with pairs in them and 811B horns. There was even a HeathKit clone. These are not rare woofers.

I'm not suggesting stripping them, rather, merely observing that there is ongoing availability from those who do that. GPA can also recone damaged ones.

We've worked through several options already in this thread, and there are plenty more to try. Let's not ignore the vintage EconoWave candidates, if that's our pleasure, is all; PCD can coax even difficult combinations such as this one to happen....

[bruson]

Re: The Bloom

[QUOTE=Zilch;1672274]Is all but gone after moving the baffle flush front:



Beautiful polar.

[Zilch]

Re: The Bloom

Beautiful polar.

The speaker itself is somewhat of a work of art at this point.

[I ain't sayin' what "movement," tho.... :D ]

[bruson]

Re: Flex Your PCD Mettle:



Still a work in progress.

Used the 1 cu ft PE sub box on sale for the 10" B&C 10PS26.

This fall I'll do some measuring for the xover and pretend I'm Zilch (HA!)

Xover is now a simple LR around 1200hz.
Sounds good but I expect some measurements will improve that this fall.

I routed a 1/2" roundover on the box and I'm back to using the 3" yellow swim noodle for the WG roundover, not pictured.

[Zilch]

Re: Flex Your PCD Mettle:

I've linked your project in the first post and uploaded your pics here so they don't get lost if they're deleted from PhotoBucket, including your earlier version:

http://techtalk.parts-express.com/showthread.php?p=1663194#post1663194

[bkillen]

Re: Flex Your PCD Mettle:

OK I'm giving PCD a try and I would appreciate any tips. I am using a 15" woofer with the D220ti on a QSC waveguide. My vertical C-C spacing is 12" and WG is 3/4" forward of the 15. I don't quite know how to set the offsets in PCD. the modeled response does closely match the measured so wegot something right.
The crossover I'm working with currently is the ewave SR variant minus the LCR notch on the woofer. I have a small bump in the crossover region I would like to improve. all measurements are on axis @1meter.

[bkillen]

Re: Flex Your PCD Mettle:

The pic.

[Zilch]

Re: Flex Your PCD Mettle:

Here's how I determine the offsets (just the tweeter, actually) empirically:

Invert tweeter polarity and measure, moving the mic up or down until you find the deepest notch at crossover. That'll tell you your actual crossover frequency, and where your drivers are summing maximally at your "normal" polarity, i.e., the true forward axis of your design, within 1/8", typically.

Calculate the difference between your mic height there and the height "on-axis" which is midway between the drivers in my designs.

Put that in PCD as your design/measurement height offset.

Invert tweeter polarity in PCD and adjust the tweeter offset until PCD also generates a maximum depth notch. Increments as little as 1mm make a difference; approach it iteratively, and ask yourself if the value makes sense, knowing the physical offsets of where you assume the drivers' acoustic centers might actually be.

If I like where that forward axis falls, generally between 0° and 5° above center, corresponding to a typical listening position, that's where I finalize the design.

Easier said, of course, because any change made in the filters moves the actual forward axis up or down. You can watch what any change does by inverting the tweeter polarity in PCD and comparing how the depth of the notch changes. I usually have that stored as the overlay for this purpose.

Changing the height then tells you where the forward axis has moved. As I hone in on my final design having the forward axis AND the response where/how I want them to be, I repeat the process via measurement and adjust the offset accordingly. Successive approximation ultimately gets it all matching up; PCD and the actual measurements should track each other closely if everything is correct.

[That happens surprisingly often, actually.... ;) ]

[Zilch]

Re: Flex Your PCD Mettle:

The pic.

That's gonna work, looks like. Get the system up off the floor for your measurements. As you see in the ZilchLab pics, I have the mic and axis of device under test at ~1/2 ceiling height to minimize reflections from the floor and ceiling, and a similar distance away from other objects and boundaries in the space. For convenience, that means inverting the speaker on my relatively high test stand, and thinking "upside down."

From my own experience, you have too much overlap of the drivers in the crossover region. Look back at the final crossover design plots of the projects completed earlier in this thread to get more of a feel for this.

Since you have a wider center-to-center distance using a 15" driver than I typically show with 12s (10.5"), aim for the lowest crossover frequency that'll work to maximize the usable height of the forward lobe. That suggests moving your lowpass successively downward, or increasing its rolloff slope, leaving your highpass where it presently is for starters; the response hump in the crossover region should disappear when you have it "right."

Pete and others may have additional suggestions....

[Skywave-Rider]

Re: Flex Your PCD Mettle:

Thanks for those concise, clear posts, Zilch.

Sorry, no "thumbsup" emoticon here on Techtalk.

[bkillen]

Re: Flex Your PCD Mettle:

Thanks for those concise, clear posts, Zilch.

Yea, What he said.

thanks, that clarifies a lot, I will let everyone know how it comes out.