Hi All,
I'm designing a compact Bluetooth speaker (with emphasis on low-end oomph and power) which consists of a Tang Band W3-1876S (4-ohm) woofer and a Dayton Audio RS75-4 (full range) per channel. This would be accompanied by passive radiators and I have the transfer function looking pretty good in winISD with PRs within excursion limits.
I wanted to crossover to the woofers fairly low 500-800Hz region to focus on bass and also due to enclosure size constraints they are down firing so I don't want them dealing with frequencies higher than this - hence why they are accompanied by full range drivers (which I will section off from the rest of the enclosure) so I can crossover lower.
I am using an existing amplifier PCB I designed with 50W per L/R channel (which I can limit down as DA + TB are 30WRMS combined).
Since both drivers are 4-ohms I was worried using them in parallel the combined resistance would be too low (TB doesn't offer an 8-ohm version of this woofer) so I started looking into series crossovers. I'm getting a bit muddled around phase alignment and also if I can introduce resistors to reduce the sensitivity of the full range to bring it in line with the woofer which I imagine negates my original reason for series crossovers.
I'd be grateful for any advice or if I should just design a parallel crossover and use the 8-ohm variant of the Dayton Audio driver and accept a combined resistance of 2.67ohms. I have limited understanding of the phase graph other than I imagine I'm looking for them to align as best as possible around the crossover point?
Also I appreciate this is a bit of a bonkers set up but it's just a bit of fun to try and make a speaker that creates a feeling of punching way above its size, I want it to confuse people when they hear it - open to suggestions 😆 Thanks!
I'm designing a compact Bluetooth speaker (with emphasis on low-end oomph and power) which consists of a Tang Band W3-1876S (4-ohm) woofer and a Dayton Audio RS75-4 (full range) per channel. This would be accompanied by passive radiators and I have the transfer function looking pretty good in winISD with PRs within excursion limits.
I wanted to crossover to the woofers fairly low 500-800Hz region to focus on bass and also due to enclosure size constraints they are down firing so I don't want them dealing with frequencies higher than this - hence why they are accompanied by full range drivers (which I will section off from the rest of the enclosure) so I can crossover lower.
I am using an existing amplifier PCB I designed with 50W per L/R channel (which I can limit down as DA + TB are 30WRMS combined).
Since both drivers are 4-ohms I was worried using them in parallel the combined resistance would be too low (TB doesn't offer an 8-ohm version of this woofer) so I started looking into series crossovers. I'm getting a bit muddled around phase alignment and also if I can introduce resistors to reduce the sensitivity of the full range to bring it in line with the woofer which I imagine negates my original reason for series crossovers.
I'd be grateful for any advice or if I should just design a parallel crossover and use the 8-ohm variant of the Dayton Audio driver and accept a combined resistance of 2.67ohms. I have limited understanding of the phase graph other than I imagine I'm looking for them to align as best as possible around the crossover point?
Also I appreciate this is a bit of a bonkers set up but it's just a bit of fun to try and make a speaker that creates a feeling of punching way above its size, I want it to confuse people when they hear it - open to suggestions 😆 Thanks!
Last edited:
It shouldn't. They are not strictly in parallel, since they don't play together due to being separated in frequency.
I don't know what passive radiators you are using, but it doesn't look like you're getting the most out of the W3-1876. In 0.8 liter with an ND-90PR and 6 grams of added mass, you should be able to do something like this:
If you're willing to accept some more power limiting at the bottom (passive radiator excursion), you can get even more extension out of the woofer. This may seem like just a few more dB or Hz, but the closer you can get to rolling off at 50-60 Hz, the more natural the speaker will sound with varied music. I had that model handy because I've done multiple desk speaker designs trying to squeeze bass out of the really small speakers, and a roll-off starting too high tends to sound a bit anemic on mainstream music with a decent amount of bass in it.
I'm also not clear on how your enclosure model is set up. It seems like you have two W3 woofers and three passive radiators, but the enclosure image from WinISD seems to show one W3. I'm not sure if that's what's modeled or if it's just a peculiarity of the interface. At any rate, you typically need at least 2x the volume displacement in your passive radiators vs the driven woofer displacement, and you don't seem to have it. If you're OK with the limits this creates, it may be an acceptable trade-off for you.
I'm a little concerned that your down-firing woofers might color the sound playing that far up into the midrange. I haven't tried that configuration personally, so I'm just speculating. If I were doing something new, it's something I would mock up quickly and cheaply to see if it sounds OK before doing a complicated build. The RS75 should play down to 200 Hz easily in a small sealed box. I would typically try to cross closer to that point, but how hard you are going to push them will determine whether crossing this low makes sense.
If you're willing to accept some more power limiting at the bottom (passive radiator excursion), you can get even more extension out of the woofer. This may seem like just a few more dB or Hz, but the closer you can get to rolling off at 50-60 Hz, the more natural the speaker will sound with varied music. I had that model handy because I've done multiple desk speaker designs trying to squeeze bass out of the really small speakers, and a roll-off starting too high tends to sound a bit anemic on mainstream music with a decent amount of bass in it.
I'm also not clear on how your enclosure model is set up. It seems like you have two W3 woofers and three passive radiators, but the enclosure image from WinISD seems to show one W3. I'm not sure if that's what's modeled or if it's just a peculiarity of the interface. At any rate, you typically need at least 2x the volume displacement in your passive radiators vs the driven woofer displacement, and you don't seem to have it. If you're OK with the limits this creates, it may be an acceptable trade-off for you.
I'm a little concerned that your down-firing woofers might color the sound playing that far up into the midrange. I haven't tried that configuration personally, so I'm just speculating. If I were doing something new, it's something I would mock up quickly and cheaply to see if it sounds OK before doing a complicated build. The RS75 should play down to 200 Hz easily in a small sealed box. I would typically try to cross closer to that point, but how hard you are going to push them will determine whether crossing this low makes sense.
Last edited:
Your speakers are not in parallel when using a crossover (either series or parallel crossover). PArallel is better though. A series crossover is not as protective of the high frequency driver as a parallel because inductors with higher DCR will cause less rolloff (a plateau) at low frequencies. I am not sure why the sim is not showing this with your crossover.
Some criticism:
It doesn't make sense to have a stereo pair of speaker s mounted right next to each other, I would put them at either side of thr cabinet,
The large amount of texture on the baffle does no good for frequency response
You need 2x the excursion volume from a PR as from the woofer. So if PR's are same size and excursion, you need 4. Your high PR notch is evidence that you need more mass on the PR's, so if they are tuned properly now, you need another PR so you can add more mass. THey could also be simulated wrongly in WinISD.
Some criticism:
It doesn't make sense to have a stereo pair of speaker s mounted right next to each other, I would put them at either side of thr cabinet,
The large amount of texture on the baffle does no good for frequency response
You need 2x the excursion volume from a PR as from the woofer. So if PR's are same size and excursion, you need 4. Your high PR notch is evidence that you need more mass on the PR's, so if they are tuned properly now, you need another PR so you can add more mass. THey could also be simulated wrongly in WinISD.
So even with this setup here your impendence is super low for a 4 ohm Amp.
I just finished a 3rd order series crossover with a 4 ohm woofer and 4 ohm tweeter. I wouldn't call myself experienced with them hut I do know how it all works now.
Do you have actual measurements and impendence files you can share? I can take a whack at it. If you zip the files you can upload them to this thread
So basically, there is a need to improve a commercial product (to be), and naturally it won't be
mentioned in the ads concerning it. And where is the forum cut?
mentioned in the ads concerning it. And where is the forum cut?
Ah Thanks, yeh I think another passive radiator is probably needed to bring it up to 4 as someone else suggested. Height is a limitation which is why I'll probably end up having to use so many. Planning on ND-65 from Dayton. Definitely two woofers modelled in WinISD too. Good should on crossing over lower, I'll look into that. What software are you using? Looks slightly different to WinISD
Okay, so actually I could use the 4-ohm TB woofer and the 8-ohm equivalent of the full-range and then the sensitivities would be the same too? Then I can use a parallel crossover.
Thanks for the feedback. I have a tendency to put form over function in certain areas! If you see the last speaker I did it also had the textured baffle. Also I see what you're saying about putting the drivers further apart but the speaker itself is only about 300mm wide as it is. Just a fun little speaker to take with me when I travel. Will look definitely look to add another PR though, you're right it's needed.
Hey I managed to upload the driver files I had to the post. Just zma/frd that I had found online or traced from the datasheet. Based off what others have said I could use the 4-ohm TB driver and 8-ohm Dayton so the sensitivities align and use a parallel crossover? Thanks!
Here's WinISD with a 4th PR added, weight on each taken up to 6.4g:
Cone Excursion
Response
Still not getting the tuning freq. much lower. I get this with 10g of weight but not as flat anymore so perhaps not worth it?:
Cone Excursion
Response
Still not getting the tuning freq. much lower. I get this with 10g of weight but not as flat anymore so perhaps not worth it?:
You can but lower ohm drivers will draw more power. Think of it as a lower ohm driver has a higher sensititivy. You'll need to pad it down with a resistor. It will be the XO way easier to make with a parallel though
UniBox - it's an older Excel-based enclosure modeler.
UniBox shows similar issues with the ND65-PR. The passives have their own set of parameters that influence the tuning/response, so sometimes one is a better match for a certain woofer than another.
Here ya go. 4 ohm impendence.
I managed to flatten out that nasty hump of the full ranger. You'd be better off running a small tweeter and crossing over to that. It would sound a lot better.
Also, what are you doing with that third W3? I see you're crossing one full range to one W3 but there is a 3rd tang band in your model.
Attachments
Interesting, I could look at the ND-90 but could only fit them down firing and probably only the one. Sort of limited to putting PR's on the back but there's not much height there. I could look for an elliptical one with greater surface area.
Couldn't quote your other response but thanks for taking a look at the crossover! Would you be concerned at all about the phase difference at the crossover point? I think I'll look to run them in parallel after the feedback from before but still open to series.
There's just two TBs in my model (downfiring). You can see three PRs on the back if that's what you mean.
The 1876 likes a 1-2ltr box, and will get you to 40Hz vented or PR loaded, HOWEVER, it will unload given too much power. About how 50W/pair is highest you should go, just wired in series, but you'll need a 40Hz active highpass the direction you are heading.
The best solution I found was a 4th order bandpass using either the ND105, DSA115, or Peerless 830878 PR in about a 1.5ltr box. You can find my related Goosebump design if you look for it. I could not get it to unload on about 25W/pair due to the sealed air spring. It also adds sensitivity to the design and rolls off at about 50 and 75Hz. The low end improves with boundary loading.
The best solution I found was a 4th order bandpass using either the ND105, DSA115, or Peerless 830878 PR in about a 1.5ltr box. You can find my related Goosebump design if you look for it. I could not get it to unload on about 25W/pair due to the sealed air spring. It also adds sensitivity to the design and rolls off at about 50 and 75Hz. The low end improves with boundary loading.
What is going to screw everything up is the offset of the drivers. I can simulate the difference for you. This will show you what happens and that it must be accounted for.
My very first build. I completely ignored this and my FR has phase cancelations causing massive dips. It still sounds good, just not as goof as it could have sounded.
How far, in millimeters, are the two drivers spread apart?
The positions aren't completely finalized but also not an easy answer I'm afraid with the TBs down firing and the full range drivers facing forward. I suppose the fact that it's being crossed over so low the orientation and alignment shouldn't be such an issue? I thought this was only really a problem at higher frequencies due to the shorter wavelengths?
One limiting and very important missing thing here is that the 75 and 1876 should NOT occupy the same box volume or you will modulate the 75 like PRs with the xmax of the 1876 and cause horrible IMD. If this scraps your goals, it is due for a redesign.
Very good point and one I intend to do :
That has reminded me though one of the reasons for crossing over a bit higher (500-800) was so this other enclosure can remain small and hug very tightly around the 75's otherwise size wise it simply won't work. I need to run the 75's in WinISD with a crossover filter set above say 600Hz and workout exactly how small the volume around them can be. As you can see it's super tight in there so I have to do some re-jigging anyway: