Help setting crossover impedance for Dayton Audio B652 Air.

I have the Dayton Audio B652 Air bookshelf speakers along with the DA 2 way 2,500 Hz 2nd order crossover purchased fro Parts Express. Since this crossover can be set at 4 or 8 ohms I need to know which will work better. However, the stock woofer is 4 ohms and the AMT tweeter is rated at 8 ohms. Please help.
 
IIRC, there are mods for the stock crossover and/or complete redesign that will sound better than an off the shelf crossover. Or just design your own crossover. I think someone just posted somewhat recently about modding one on PETT.
 
It appears there were several revisions to this speaker so keep that in mind when hunting for mods. The box could also use some padding and reinforcement. Unless this OTS crossover is made specifically for this speaker there's a good chance it will sound worse. If you can accurately measure in box response and make .frd and .zma files it's fairly straight forward to build a proper crossover for this speaker. I'd be glad to help out in that regard as I currently have plenty of free time.
 
I measured response with the tweeter and woofer connected solo in box, 22 inches away on the plane of the bottom half of the woofer. This is roughly in a 13 foot cubic room, but I took care to reduce direct reflections. If this is acceptable I will post the REW files.

I have a boxsim file but there is so much info I could have entered incorrectly. The driver positions on baffle are probably okay.

A 2500Hz crossover will almost certainly fail. It needs the crossover somewhere in the 4500Hz range.
 
Last edited:
Started to make a big long post on how to take measurements in order to make .frd and .zma files for use in xsim, or winpcd, but thought it easier to just link to a another forum where a reputable member modded the b652air crossover. After reading a lot of posts on diyaudio, I'd venture to say members here will quickly pick up xsim or winpcd crossover design software.


Fun with the Dayton B652-AIR -

Techtalk Speaker Building, Audio, Video Discussion Forum
 
I think changes since augerpro's measurements have moved the woofer peak down to 3.26k, I ended up using 7.2uF and 80 ohms with the 300uH inductor to get rid of the resonance on the waterfall chart. 40 ohms is not enough, and you can see the residual peak on augerpro's chart.

I don't know how relevant my crossover is though since it assumes you will be listening with the speaker tilted back a bit and off axis 15 degrees.
 
Some of the posts do mention that there have been revisions, some of them fairly large, like a different woofer. And I think augerpro was intentionally keeping mods on the cheap.

Did you strengthen/brace the box or add any stuffing? Funny thing, supposedly the spring in the spring clips can buzz loud enough to be heard.

If you like the way it sounds, well, that's what counts.
 
The spring clips definitely buzz, but that is more because I play them loud due to distance. It drives me crazy. I might be able to change it soon.

I took measurement with and without extra stuffing. The main difference is a peak in the response just after 1KHz. A 1LB towel doesn't quite get rid of it, and actually increases the bass resonance. So I need to figure out a different stuffing tactic. I don't know how I could fit two towels in there. Hmm, shrink the towels with a vacuum bag...

As for bracing, a precisely cut wide piece of cardboard is enough to stop the front baffle from ringing when I tap it. You have to cut it so it slides in and wedges flush between the sides without bending, as any creases or curvature ruins the stiffness. Anyone who thinks this doesn't work, obviously hasn't tried it. This is a cheap thin box so it's not hard to brace this way. (now that I recall, this might have been the way I got rid of the 1KHz resonance before, hence why stuffing did not help)
 
Yep, flimsy boxes can have a noticeable affect on a speaker's sound.

For a more permanent fix, what about wooden dowels or scrap pieces of wood for bracing? Glue them in place and be done with it. Have any of that thin wavy foam mattress topper laying around? Attaching that to the walls helps without drastically changing tuning on a ported design. I've went as far as gluing/screwing cement board onto the inner walls. As far as stuffing, Rockwool works as well as fiberglass but w/out the itchiness. Denim insulation works even better but you will need to find some thin batts. Even heard of people throwing a few spools of yarn in there.
Order some binging posts and ditch those spring clips.
 
But I wouldn't be done with it. You can't stuff around a brace very easily in a box this small. Hence removable cardboard braces. Or perhaps, some stuffing that presses against the sides hard enough to damp them.

I wonder how many woofers there are that would actually be better than the stock woofer in every way. From what I've seen the spectral decay on the stock woofer is actually pretty clean aside from the 3.26KHz resonance which drove me insane. Hence augerpro's notch filter. And once you take care of that you can benefit from the rigidity of the cone. You can get a softer woofer but then you get more cone flex.
 
I'm an amateur, why waste money I don't have building my own mediocre speakers? I got the B652 for it's affordabilty and it's potential, and I can learn as I go along. Of course now my problem is no one will humor me now, they just tell me to spend more money. Ah, there is the explanation. They have more money than I do.
 
Amateur, ha! Dude, don't let that stop you. You are here aren't you, so you must have that desire to get your hands dirty.



Here are just a few options to get you started. Look through these links and you will find some very affordable diy builds, as in all you do is build following a plan, the leg work has already been done. You could build literally dozens of speakers for little more than the b652 cost and have something that sounds so much better than 'mediocre', even if you have only basic wood working skills.

undefinition


Speakerbuilder.net!  The art and engineering of loudspeaker design


Zaph|Audio


A Guide to HTguide.com Completed Speaker Designs.


I'm working on voicing these. Unfortunately, also playing around with a simple little tube preamp so it gets a bit frustrating when changing so many variables while trying to finalize a build.
 

Attachments

  • photo 2 (16).JPG
    photo 2 (16).JPG
    461.4 KB · Views: 107
Start on something super easy and cheap lie a PicoNeo. Two $15 drivers (sometimes on sale for $12), a pair of binding posts ($4), a piece of pvc for a vent, and whatever kind of materiel you like working with, just don't make no flimsy box. No crossover. There intended use is for nearfield listening and I think you would be totally impressed with their abilities.
 

Attachments

  • IMG_0585.jpg
    IMG_0585.jpg
    869.8 KB · Views: 131
I'm an amateur, why waste money I don't have building my own mediocre speakers? I got the B652 for it's affordabilty and it's potential, and I can learn as I go along. Of course now my problem is no one will humor me now, they just tell me to spend more money. Ah, there is the explanation. They have more money than I do.
Your approach is perfectly fine - saves you from having to build an enclosure from scratch, and there is plenty to learn this way.

More than a decade ago, when waveguides still were an absolute rarity, some folks in Germany came up came up with a revamped crossover for the Behringer B2031P as well. Arguably a much more rewarding modding target, but they've been discontinued for years now. I guess the enclosures and drivers would have been equally

You said a 2.5 kHz XO won't work and 4.5 kHz would be a better choice - I would be very worried that the apparent ~3.2 kHz breakup mode (assuming it is one) would be wreaking havoc on directivity. The likes of Boxsim are no good predicting radiation in regions of non-pistonic motion. That said, smallish ribbon tweeters are in fact notorious for their excursion-related distortion issues towards the low end of their range, so you probably wouldn't want to go below ~4 kHz either.

The Stereophile measurements honestly are less terrible than they could be. The breakup mode seems to behave itself reasonably well once EQ'd out (mind you, you aren't going to eliminate the resulting peak in 3rd harmonic at ~1.1 kHz if inductance / B*L nonlinearity is a factor), and other than that there's just the usual issue of very mismatched woofer and tweeter directivity at crossover. Some designs have used the baffle step to their advantage in the past, but apparently not this one... the XO point might just be too high altogether.

What's intriguing is that things are actually looking less disjointed in the vertical. So some experimenting with baffle width may be worth it. This may be why e.g. the usual ADAMs are looking the way they do.

You may be stuck between a rock and a hard place there.
 
Last edited:
Thanks for all the ideas guys.

The reason this speaker appealed to me is pretty much just the AMT. Any peaks in the 3-5KHz region drive me absolutely nuts. Augerpro's mod with the notch filter convinced me to give it a go.

And then I decided I wanted to listen with equidistant acoustic centers, which everyone tells me is inadvisable but the resulting response actually seems pretty even and phase is pretty flat. Maybe this is related to how the AMT works. I was ready to throw away this idea until I saw that it actually works. The question is where do I go from here?

This is the frequency response and spectral decay. For the response I used a 15 cycle window. For more extensive mods (if worth pursuing) I'm thinking a shallower woofer to move the acoustic centers closer (so you don't have to tilt the speaker back as much) and better match the dispersion of the woofer with the tweeter makes sense. At least for my amateur crossover.

The treble sounds hot, so maybe you are right about the dispersion problem. I just don't trust dome tweeters enough to pay for one I haven't heard yet. And it would be hysterical to try to mod one into the baffle anyway.
 

Attachments

  • keantoken B652 response with 15 cycle window in 13 foot cubis room.png
    keantoken B652 response with 15 cycle window in 13 foot cubis room.png
    36.1 KB · Views: 108
  • keantoken B652 spectral decay.png
    keantoken B652 spectral decay.png
    151.8 KB · Views: 129