Flex Your PCD Mettle:

[Greg01]

Pete

Thanks for answering my ramble from last night which is recreated in its entirety below. However, I never got around to asking you the question I really intended to ask.

I know that Zilch's cabinet has the inset baffle. I was going to build my own cabinet of approximately 60 liters tuned to 40 hertz (with a slot port). My cabinet may or may not have an inset baffle depending on how you answer the following question. I would normally flush mount the woofer. In this cash I would rear-mount the woofer .75 inches from the front of the baffle to maitain the distance pursuant to your model.

I usually do not inset a baffle. I would just have a flat baffle with a .5 inch roundover (this is the largest roundover bit I am comfortable using in my quarter inch collet router). Do you think this would make a material difference in the frequency response? Sorry for not getting to my point in my last post.

Pete

I would like to build this project. I would like to clarify a point or two. The cabinets that Zilch used appear to have the baffle inset about .75 inches or so. If I have read what you posted earlier correctly, because of the waveguide and the size of the woofer, this would not have too much of an effect on the frequency response. Is this correct?

Also... After compensating for the baffle step response it appears that this system will only have ~ 90 db efficiency. Is this correct as well?

I know watts are cheap and I don't mind putting some power to this project. I just wanted to clarify these things before I committed to a project.

[Pete Schumacher ®]

I think doing what you're planning is great. It keeps the offsets as measured by Zilch, so everything should come out the same. As for diffraction on the woofer output? Any diffraction signature will be small due to the large size of the driver area. The roundover won't change the response much, but it sure will look a lot nicer.

Having a flush mount waveguide on a flat front baffle is how I would do it too. Zilch is using those knockdown cabinets from PE.

Pete

Thanks for answering my ramble from last night which is recreated in its entirety below. However, I never got around to asking you the question I really intended to ask.

I know that Zilch's cabinet has the inset baffle. I was going to build my own cabinet of approximately 60 liters tuned to 40 hertz (with a slot port). My cabinet may or may not have an inset baffle depending on how you answer the following question. I would normally flush mount the woofer. In this cash I would rear-mount the woofer .75 inches from the front of the baffle to maitain the distance pursuant to your model.

I usually do not inset a baffle. I would just have a flat baffle with a .5 inch roundover (this is the largest roundover bit I am comfortable using in my quarter inch collet router). Do you think this would make a material difference in the frequency response? Sorry for not getting to my point in my last post.

[Zilch]

EconoWave Deluxe crossover and parts list updated for compression driver protection at high SPL. Increase R1 power rating as required; use 9.1 in parallel with 10 Ohms for 20W (shown, optional) or 4 x 4.7 Ohms in series/parallel for 40W. I left standard at a single 10W, as I detect no appreciable temperature rise under moderately high (90 dB @ 2m) listening conditions:

http://techtalk.parts-express.com/sh...9&postcount=72

Also added final vertical polar map; we're good (-6 dB) +/- 15°.

Forward axis is midpoint between drivers +2.0° and XO at 1.565 kHz.

All measurements w/L-pad @ 1:00....

[Zilch]

As previously, measurements on axis midway between woofer and waveguide, 10.5" c/c distance, 44" mic distance, HF w/L-pad @ 1:00:

Performance on the Schumacher V1.1 XO, "Deluxe":

Comparison to 3012LF, HO HF at "Max":

I'm not seeing 5 dB to grab there; 3 dB perhaps....

The files:

[donprice]

Zilch - Can you double check the files you posted? The FRD files are fine. I'm guessing the files with no extension are ZMA files. Where are the project (CSP) files?

Thanks!

[Zilch]

Everybody's gotta work up their own CSP files for now; I haven't loaded this in PCD yet, myself.

Good question re: ZMA. Don't know what file extension WT2 attaches to them, actually. Assume that's what they are for now.

[Added some measurements of potential interest, above.... ]

[donprice]

Ahhh, that was Pete that worked up the worked up the model and you did the building and measuring. Got it. Maybe Pete can hook me up with the CSP file.

Reverse engineering is much simpler when you have the plans to learn from.

[Pete Schumacher ®]

Looks like the HO has a bit of lift in the lower end compared to 600Hz. A shallower LP filter may be in order. Also, the extra sensitivity needed from the tweeter may benefit from a change in the component values. That peaking around 5KHz can be smoothed nicely by changing around the cap value and resistor L-pad.

If you have the parts laying around, try this.

Low pass:

2.5mH 15uF

High Pass:

3.9uF
2.7 Ohm series - 8 Ohm shunt.

[Pete Schumacher ®]

Ahhh, that was Pete that worked up the worked up the model and you did the building and measuring. Got it. Maybe Pete can hook me up with the CSP file.

Reverse engineering is much simpler when you have the plans to learn from.

I attached it in my reply to zilch. ZMA extensions have been added.

[donprice]

Thanks guys. That's a huge time saver. Now I just swap out the 3015 curves for the 3012HO curves and I almost done

[Pete Schumacher ®]

Thanks guys. That's a huge time saver. Now I just swap out the 3015 curves for the 3012HO curves and I almost done

You'll want to adjust the acoustic offset as well. The 15 likely is more closely aligned with the tweeter, so the z axis value should be reduced.

[Zilch]

Toed 'em in to see if the cab edges might be doing the 5 kHz thing; it doesn't appear so.

HO is Cyn, LF Grn. 4.7 uF as the core HO highpass looks a bit better, Red. XO frequency went down to 1.675 kHz, but forward axis moved up to 1° +5.5° from the midpoint between woofer and waveguide:

Vertical Polar Map doesn't make it (-6 dB) to +/- 15° at this frequency:

Here's the full +/-20°:

Compare to LF crossed 125 Hz lower:

Footnote: Lower null comes down from above the XO frequency in measurements, and upper from below it, aka the "Zilch Effect." PCD vertical driver offset up is (+) and down (-), no? Because they move opposite in the PCD sims.

[No, not because I measure them upside down.... ]

[Mayhem13]

Maybe a little non-descript foam?

[Zilch]

May be time to swap in a bunch of DE250s; we're talkin' 1 dB here....

[Zilch]

Looks like the HO has a bit of lift in the lower end compared to 600Hz.

Looking at the factory curves, I'm seeing HO having more of a rising response than LF, enough so that we have to give up some of the HO's efficiency for it to play as flat down to 450 Hz:

http://www.parts-express.com/pdf/290-590s.pdf

http://www.parts-express.com/pdf/290-589s.pdf

I believe I'm seeing that in comparing the vertical polars at the low end, as well. I can't measure with sufficient resolution in that region to be sure, but there are certainly indications. Where is baffle step kicking in with these 17" wide baffles, ~800 Hz, no...?

[Jeff B.]

Footnote: Lower null comes down from above the XO frequency in measurements, and upper from below it, aka the "Zilch Effect." PCD vertical driver offset up is (+) and down (-), no? Because they move opposite in the PCD sims.

[No, not because I measure them upside down.... ]

It was recently discovered that PCD's vertical axis panning is upside down. Once I was questioned on this, it was very easy to quickly confirm it. I am surprised that I missed this, but the new version matches CALSOD exactly off-axis (I thought I tested this before?). I have corrected it in my working revisions and the change will be in Ver. 7 (If I ever get around to finishing it. )

One thing that bewilders me in you graphs is the phase response. It appears to be minimum phase based on the phase response line. However, you have an in-phase crossover based on how they sum acoustically, which would preclude it from having a minimum phase summation. It almost looks like the phase has been extracted from the final summed response, but that wouldn't be correct. Your phase should be wrapping near the crossover point.

[Zilch]

It was recently discovered that PCD's vertical axis panning is upside down. Once I was questioned on this, it was very easy to quickly confirm it. I am surprised that I missed this, but the new version matches CALSOD exactly off-axis (I thought I tested this before?). I have corrected it in my working revisions and the change will be in Ver. 7 (If I ever get around to finishing it.)

It's critical when using the measured location of the vertical null to establish/verify z-axis acoustic center locations. The only way I could get that happening was to measure and design midway between the drivers, where the differences are small.

In the meantime, will inverting the vertical offset polarity remedy the issue? I haven't tried that yet, relying instead on the measured location in all work to date.

One thing that bewilders me in you graphs is the phase response. It appears to be minimum phase based on the phase response line. However, you have an in-phase crossover based on how they sum acoustically, which would preclude it from having a minimum phase summation. It almost looks like the phase has been extracted from the final summed response, but that wouldn't be correct. Your phase should be wrapping near the crossover point.

Yes, it's system minimum phase I'm showing, not total phase, as verification that the measurement was taken on the forward axis. I'm usually also showing the inverse null with the cursor at its minimum so it can all be correlated. I suppose I could bump up the resolution of the minimum phase, but the presentations soon become too information intensive.

It's only in the past several months that I have gone so deeply into these vertical polar determinations, and one thing I concluded immediately is that it's essential to find and design on the forward axis when forward lobe is narrow. It's an iterative process trading frequency response against forward axis steering to get it optimized, and when working with waveguides and the wide center-to-center distances they impose, it's essential to be precise when designing passive crossovers for them, as the usable vertical "window" can be quite small.

Wayne's video on vertical nulls is instructive:

http://www.pispeakers.com/misc/Vertical_Nulls.wmv

[Zilch]

Four DE250 drivers, #1 on a different QSC waveguide and also shown with foam bumpers. Note expanded vertical and horizontal scales:

[That ain't it.... ]

[Pete Schumacher ®]

Four DE250 drivers, #1 on a different QSC waveguide and also shown with foam bumpers. Note expanded vertical and horizontal scales:

[That ain't it.... ]

It's looking like there may be a benefit to mounting that waveguide on a flush baffle.

[Zilch]

It's looking like there may be a benefit to mounting that waveguide on a flush baffle.

It's easy enough to move the baffle forward to accomplish that, as Jack earlier demonstrated.

We're at +/- 2 dB, but there's certainly no harm in doing better.

I'm trying to figure out why the 3012LF system doesn't show so prominent a peak at 4.8 kHz as the HO here....