• Please note that Audiokarma will be offline briefly for updates early on Monday morning, September 28th.

dumptruck is finally building ewaves for real

It appears that the first pulse is noise and the 2nd is the speaker. Brilliant. Now I have to figure out why I can only get this ugly little blip from the speaker. To clarify what's going on here, I realized that if I start the gate at what the yellow arrow in the pic below is pointing at, I get pretty much exactly Zilch's response measurement.

attachment.php
 
Last edited:
dumptruck said:
It appears that the first pulse is noise and the 2nd is the speaker. Brilliant. Now I have to figure out why I can only get this ugly little blip from the speaker. To clarify what's going on here, I realized that if I start the gate at what the yellow arrow in the pic below is pointing at, I get pretty much exactly Zilch's response measurement.
Click to expand...

Sounds like your zero time isn't the first arrival at the mic. Not sure if I said that right... I haven't used ARTA so I don't know where you adjust that.

HOLMImpulse gives options including "detect time zero" (based on measured peaks) or "time zero locked" (time alignment).

HOLMImpulse is also free and giving it a try might help you sort it out.
 
I figured it out - I was measuring the #*(%@$ crosstalk+noise from the unconnected mic channel. Unbelievable. NOW for some real measuring.
 
Just wait until you hear a distinct 40-50 Hz LF growl from your horns :sigh:

Hint: turn off the subwoofer :yes:
 
I will try to keep garbage-y data to a minimum after this, but I wanted to get some evidence of reality posted after that nonsense. Red trace is Zilch's (dB levels are not matched/calibrated). The highest level trace is on the woofer axis, and then it is steps heading above axis from there. Not sure what the deal is with that jump down at 650Hz... not quite as attractive as Zilch's in general, in this measurement. Reducing smoothing has very little affect on this plot, if you were wondering.

Anyway, next I'm gonna get the DE250 into play. Should be less variation with those, i.e. easier to spot problems in my technique. If that goes well, I'll revisit the woofer measurement, looking for problems and seeing how well my pair matches up. If that goes well too, then off to the races with finding nulls and whatnot.

attachment.php
 

Attachments

  • 414z response test1.png
    414z response test1.png
    31.8 KB · Views: 325
Cool.
Don't feel bad. The other day I was measuring and I forgot to mute the other channel. Crazy nulls. I was like, "man, that speaker is a POS!" :D
 
Okay, so on to woofer #3. Just made a deal for TD15M. Now, before you all start pestering me about the center-to-center spacing, especially with the QSC waveguide, I diagrammed it out, did the math, and estimate 40.3° spacing between nulls. Right in the target zone. There is a bit of z-offset (red line on left), which may be cancelled out by the crossover? I guess we'll see.

edit: just doing a little back-maintenance here. The target zone is actually more like 60° for this waveguide. Oops. On the other hand, this woofer turned out to want a lower crossover point, which widens the null spacing.

attachment.php
 

Attachments

  • TD15+QSC layout.png
    TD15+QSC layout.png
    21 KB · Views: 277
Last edited:
I think I mentioned it, but those drivers are 4 ohm. I hope that doesn't throw a wrench in the X-Over. If it's a problem, let me know before Tuesday :D
 
Nah, I knew they were 4Ω. I don't mind the lower impedance and haven't purchased any crossover parts for this project :thmbsp:.
 
yer gonna love those TD's... and interesting, you overlapped the horn and the woofer mount like I had to... I had no room, but it worked out for the better getting things closer.. I back mounted the woofer, and front mounted the horn..
 
dumptruck said:
Okay, so on to woofer #3. Just made a deal for TD15M. Now, before you all start pestering me about the center-to-center spacing, especially with the QSC waveguide, I diagrammed it out, did the math, and estimate 40.3° spacing between nulls. Right in the target zone. There is a bit of z-offset (red line on left), which may be cancelled out by the crossover? I guess we'll see.

attachment.php
Click to expand...

You'll never find me criticizing anything based on center to center spacing. I'm of the opinion that it's one of the most overblown issues I've ever encountered.

Nice choice on the woofers. :)
 
skywave-rider said:
That's a nice drawing, what did you use to create that?
Click to expand...
Adobe Illustrator (I've been using it constantly for my job for so long that it's the easiest way to do practically anything graphical for me), but the woofer is cut-n-pasted from the PDF on the AE site, so I didn't actually do much drawing there, just scaling and aligning. P.S. if you ever need anything like that for a project and have the measurements, it is generally just a few minutes work for me, feel free to ask.
 
Last edited:
Question for the geeks with measurement gear: I want to make sure I understand how to arrive at a good starting point in PCD. Is this how it could/should work?

1. Select an arbitrary axis, such as straight out from the tweeter or between the driver centers.
2. Measure both drivers unfiltered and each driver unfiltered from there, maybe invert phase on one.
3. Load all three measurements and the relevant data in PCD and adjust z-offset on the simulated combination of the individual measurements, until the result matches the combined measurement.
4. Go to town designing in PCD, keeping an eye on where the primary lobe/notch is ending up, repeating 1-3 as needed if you need/want to select a different design axis.

yeah? :dunno:
 
That sounds like my attempts with HOLMImpulse. One other item is using a protection cap on the HF driver so you don't blow it.

I think you need to include smoothing or gating prior to exporting to PCD.

What I'm fuzzy on is the measured phase data and how "time zero" is determined/fixed to keep the phase correct between the LF & HF sections. I don't think I have it right since I get different z-offsets for the "normal polarity" and "reversed polarity" combinations.

I believe that if you get the phase data correct you can change your design axis without additional measurements (assumes PCD properly models the frequency response changes as you move of axis).

Good luck with your measuring and modelling efforts. I'll pay close attention since I have the QSC/DE250 combo up and running and the TD15M are, theoretically, under construction.
 
So we are going to make the same thing if your TD15 show up eh? That should be helpful to at least one of us :). I've got the gating down - just didn't mention it. If I'm remembering right, the "time zero" thing was a non-issue with ARTA when I was doing this last. I will be posting files for your examination, though. Gotta make baffle #3 first, though :sigh:.

On that protection cap, is that really required? I was planning to look into that. It seems like I'd want the data down below 1kHz, at which point a 1st order filter isn't going to do much, or is it? The driver will be on SO little power.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom