Review my Xover! Another W5-704D build

[fusseli]

I plan to make a couple sets of mild-budget monitors based on the W5-704D. I want to try and use the ND28F this time to go with them, I used the ND20FA on the last set I made. Plan on using larger cabinets for better extension on this batch, 0.5cuft tuned to 43Hz. I am also going with LR2 slopes and processed for 4dB of BSC.

The 4-6khz sag seems to be a product of the ND20's response. Thoughts on using the notch on the tweeter resonance, do I need it with these slopes? Once I get a design nailed down I plan to run the drivers through my WT3 and tweak as necessary with the actual .zmas.

EDIT: the design will now be the 704D and Vifa XT25

[johnnyrichards]

The Fs impedance spike on the tweeter is not tooooo high, will a basic L-Pad flatten it out, or do you really need a big, expensive coil on it?

Just seems an oddball choice to use an expensive notch filter to tame a 4 ohm spike.

If you want to use an LCR comp, try using the "Auto Calc LCR Comp" feature, and see what the values are.

My guess is however, you will have at least as good of a result using a simple L-Pad as you will with an expensive notch filter, though. Save a few bucks on a budget build in the process.

Edit: I realize you are trying to flatten the hump in response, but I think you will be surprised what a basic L-pad will do with that, too.

[fusseli]

Good idea, thanks for the (obvious?) suggestion. That and a couple tweaks to the HPF to get back in phase gets me here:

[fusseli]

Haha, so now I'm thinking of trying a Vifa DX25 instead of the ND28. I'm not limited to baffle real estate so a bigger tweeter faceplate is a non-issue... At the same time, the ND28 looks like a fine match for the 704D and the Xover is simple enough. Thoughts?

[greywarden]

The only issue I can see is the lump that peaks at 8kHz... if its not going to be screaming at you it will be fine.

[Wolf]

Since I already used those drivers....together....

...try a series xover, centered at 2kHz.

You'll like it!
Wolf

[fusseli]

The stuff near 8khz is a biproduct of the ND28 response, so I'm not really concerned about it.

Since I already used those drivers....together....

...try a series xover, centered at 2kHz.

You'll like it!
Wolf

I see! Well I better stick with the ND28, then. Doing something that's already been done just takes all the fun out of it

[fusseli]

Here's the simplest yet. Shifted the xover to 2700Hz to and got rid of the zobel on the woofer. Six components now, single coil on the woofer and 3rd order electrical with l-pad on the tweet.

There's a dip off-axis around 5-6Khz due to the phases starting to go out of alignment, but it's not too bad. Other than that I think I like it...

[Wolf]

The stuff near 8khz is a biproduct of the ND28 response, so I'm not really concerned about it.



I see! Well I better stick with the ND28, then. Doing something that's already been done just takes all the fun out of it

I don't think you understood me... I used the 704D and the ND28F together, and so did Dan Neubecker.
My 'Canvas':




Later,
Wolf

[fusseli]

I see, now all of the fun really has been taken out of it

[Wolf]

I see, now all of the fun really has been taken out of it

No-no, please continue! This will give another version of the same design, from another perspective.

Have fun!
Wolf

[fusseli]

Haha ya, I intend to. I like small flange tweeters, that's why the ND28 caught me in the first place. Looking at your Canvas, your low end looks a bit different, did you run PCD with your own in box measured .frd? I'm assuming that's a yes I am looking at about 14L also, tuned at 43 Hz with 2" port. And wow the impedance profile on the series Xover is nuttier than squirrel shtuff.

The power response on my 2.7khz LR2 looks more flat than yours We'll see in a couple weeks what changes when I have drivers in-hand, I'm using Zaph's data until then for this initial run. Watch me have something wrong in the model somewhere so I can put my foot in my mouth...

[Wolf]

Haha ya, I intend to. I like small flange tweeters, that's why the ND28 caught me in the first place. Looking at your Canvas, your low end looks a bit different, did you run PCD with your own in box measured .frd? I'm assuming that's a yes I am looking at about 14L also, tuned at 43 Hz with 2" port. And wow the impedance profile on the series Xover is nuttier than squirrel shtuff.

The power response on my 2.7khz LR2 looks more flat than yours We'll see in a couple weeks what changes when I have drivers in-hand, I'm using Zaph's data until then for this initial run. Watch me have something wrong in the model somewhere so I can put my foot in my mouth...

The frd's I used were measured by Dan Neubecker, so they were real measurements, as were the zma's from my WT2.

The impedance profile being nutty was due to the notches I used in the SXO.

The reason I didn't use a 2" port was the port cancellation higher than I wanted it to be and more dramatic than wanted as well. So- I used 4x 1" vents to be the same equivalent end area and took the port resonance high enough to no longer be an issue. The FR low end is a splice from RM.

Later,
Wolf

[fusseli]

Ah, gotcha! I haven't modelled port resonances at all yet and nothing from them will show up in my fr from PCD, anyways.

[Wolf]

Ah, gotcha! I haven't modelled port resonances at all yet and nothing from them will show up in my fr from PCD, anyways.

Unibox showed me that effect, and is why I use it.

Later,
Wolf

[fusseli]

I have played around with Hornresp before to see how big of an impact port resonance may be in a particular scenario, if it isn't anything drastic then I think it permissible. Anything less than a couple dB falls within what I'd think a good realistic margin of error for sim-based design, anyways. I haven't experimented and/or heard any port resonance issues before, though.

[fusseli]

Executive decision... In order to make this build a little more unique compared to my last 704D build, and after considering some reading I've done on the XT25 and ND28, I'm now going for the XT25 tweet. I made such a rash last minute decision to change to the XT25 that I removed the ND28s from my cart and added in the Vifas instead, without even modelling anything, and ordered them with the 704Ds. I'm working on a new Xover to accomodate, I plan to stick with the single coil on the woofer.

[Wolf]

You'll find the XT25SC90 likes 3kHz+, which the 704D is okay with. BUT- you'll likely need second or third order on the Sc90 to make it happy, and then maybe even require
some Fs compensation, dependng on your rolloff alignments, slope, and xover point. They tend to get nasal with material in upper male vocals if the Fs is not contended with
or xover'd too low or shallow. The reason I'm telling you this is that usually a one-order difference electrically will solve the acoustic center mismatch, and you may need to
go 12dB on the woofer to get them lined up well.

Later,
Wolf

[fusseli]

Ah, sorry I meant to specify the larger XT25TG30 (Fs near 530Hz and around 91dB @ 2.83V... it was a quick and rash change of heart to a large faceplate tweeter!!). Using traced zma and frd I get this with the same number of components, crossover point is lowered a bit from before. The Fs spike is hacked down considerably and is attentuated well over 40-45dB... thoughts?

[Wolf]

Even the TG30 has issues as you've seen. Personally, I would maybe tack a 1-2 ohm additionally prior xover in series to attenuate the topend a bit more as they look
a little hot there. You'll also notice you're approaching -60deg in phase on the system impedance This is due to where your 2 ohm resistor is placed. Make it a 4 ohm
resistor our front, and that will be better.

You'll also notice you are not at an LR2 with the TG30. Your target is way off. It's closer to 3rd or fourth order. Shift your target selections until you get to where you are
and it aligns with the tweeter plot to see what kind of rolloff alignment you have. The xover point will still remain the point the drivers cross in the middle.

If you go 12dB on the woofer, that peak at 3kHz will be minimized. Just drop a 10uF on it to see.

Later,
Wolf