Flex Your PCD Mettle:

[carlspeak]

Re: Flex Your PCD Mettle:

Hey Z, do you have an automated system for taking those 36 measurements for the polar maps? If not, it's got to get pretty boring tweaking that turntable in 5 deg. increments over and over again. :D

BTW, thanks for the wavelet links!

[bkillen]

Re: Flex Your PCD Mettle:

Build a sub!:D

or three!

[Zilch]

Re: Flex Your PCD Mettle:

Hey Z, do you have an automated system for taking those 36 measurements for the polar maps? If not, it's got to get pretty boring tweaking that turntable in 5 deg. increments over and over again. :D

Oh, the EXCITEMENT! :p

It takes about an hour to do one, but we sure have learned a bunch from them.

[It's 19, actually; I just click the "mirror" button with these symmetrical designs.... ]

[carlspeak]

Re: Flex Your PCD Mettle:

Oh, the EXCITEMENT! :p

It takes about an hour to do one, but we sure have learned a bunch from them.

[It's 19, actually; I just click the "mirror" button with these symmetrical designs.... ]

I thought it was 90 deg each side of zero deg. to map the entire shape.
90/5=18X2=36.
Is your turntable automated? or do you move it manually ea. 5 deg?

[Zilch]

Re: Flex Your PCD Mettle:

I thought it was 90 deg each side of zero deg. to map the entire shape.
90/5=18X2=36.
Is your turntable automated? or do you move it manually ea. 5 deg?

Plus the 0° = 37. HOWEVER, if the speaker is symmetrical, I just do one side and use the same data for the other direction. AR3a is asymmetrical, so I did all 37 measurements for that one. Stock AR4x, same deal. Allison, well, you know what happened -- thankfully, you independently confirmed the finding.

I do it manually. I have a motorized turntable, but haven't engineered it into the measurement system yet. I'm not doing anything any other DIYer with the requisite interest can't do....

[Dennis H]

Re: Flex Your PCD Mettle:

We all appreciate your hard work, Zilch. One small quibble with the polar measurements: your axis of rotation appears to overstate the directivity a bit, i.e. the AC is moving farther from the mic as it rotates. You're showing 'constant directivity' on the polar maps down to 300 or so when we all know the woofer is nearly omnidirectional down there. Like I said, a small quibble.

[Zilch]

Re: Flex Your PCD Mettle:

We all appreciate your hard work, Zilch. One small quibble with the polar measurements: your axis of rotation appears to overstate the directivity a bit, i.e. the AC is moving farther from the mic as it rotates. You're showing 'constant directivity' on the polar maps down to 300 or so when we all know the woofer is nearly omnidirectional down there. Like I said, a small quibble.

That's just to tease the dipole-heads. ;)

The conventional polar plot says it's only -6 dB at 90°, 250 Hz, which the normalized -10dB polar map confirms:



It's a dilemma how best to measure directivity; do we rotate about the acoustic center or center of the face of the baffle, somewhat forward of it? I asked about this in the "Geddes on Waveguides" thread at diyAudio and was ignored.

As may be seen in the pics, I rotate about somewhere in the region of the acoustic center. Once I get to ~75° the cab edge begins "shadowing" the direct path, and the mic is looking more at the side of the cabinet. It's even more pronounced with edge overhangs in the way -- we're just looking at diffraction out there.

What's my rationale? I'm more interested in what's happening in the region of the 30° - 45° beamwidth pattern edge, and having the acoustic center stay put better portrays that.

[Also, rotating about center front is not so easily accomplished.... ;) ]

[Zilch]

Re: Flex Your PCD Mettle:

There's always concern, as mentioned above, that variability among drivers will render the crossover design an academic exercise. In this case, the woofers are very old (date code 391742?) and of indeterminate provenance. The moment of truth comes with swapping in a different driver set (the 4555 also) and clicking the "Start Measure" button:




L-pad's at 1:00, and as in other recent designs, I have further provided means to adjust the HF "tilt," i.e., alter the power response. It's running the red curve right now:



In this instance the new baffle has a 3/4" rounded over woofer cutout, as well. We'll see if that has significantly altered the directivity, next....




Toggle:

http://techtalk.parts-express.com/attachment.php?attachmentid=10907&d=1284011969

http://techtalk.parts-express.com/attachment.php?attachmentid=10947&d=1284178267

Toggle:

http://techtalk.parts-express.com/attachment.php?attachmentid=10859&d=1283925332

http://techtalk.parts-express.com/attachment.php?attachmentid=10948&stc=1&d=1284179200

[Skywave-Rider]

Re: Flex Your PCD Mettle:

Those woofers are tight!
Are both red and green from the top graph mounted in the round-over baffle(s)?

[Zilch]

Re: Flex Your PCD Mettle:

Those woofers are tight!
Are both red and green from the top graph mounted in the round-over baffle(s)?

No roundover in the red one; tryin' to see the difference.

Click the toggles, two sets, now....

[Skywave-Rider]

Re: Flex Your PCD Mettle:

On the horizontal polars, toggling between round-over and not, I believe the round-over distributes the energy better. If you can post an image of the rounding, that would be great.

Looking at the second set you just posted makes it easier to see. Not sure what you did with the gain.

[Skywave-Rider]

Re: Flex Your PCD Mettle:

Zilch, if you choose to go with a round-over does that mean you would take new base measurements for both drivers on the baffle and bring those into PCD? I am asking because it seems to affect the on-axis response.

[dantheman]

Re: Flex Your PCD Mettle:

Zilch, could we do less smoothing on those? That might make the difference more easily visible.

Dan

[Skywave-Rider]

Re: Flex Your PCD Mettle:

Did you change the core filter in the second set?

OK, I'll shut up now....

[Zilch]

Re: Flex Your PCD Mettle:

The earlier directivity maps were made with the red driver set and baffle without roundover, V1.1 XO; the new ones with V1.2, which is +/- 0.25 dB flatter. The Red/Grn comparison is both with XO V1.2. I suppose it would be more definitive if I were now to move these drivers onto the non-roundover baffle and remeasure the directivity, but I believe the trend is clear -- the roundover is better in the region of interest, but not hugely so. There's more baffle work to be done; the improvements are incremental at this point.

Yes, I can reprocess with less smoothing, but that may obscure what we're looking for. Here's the normalized views at 1/6 octave, which should ignore the small response differences: