Hello everybody, I am a long time lurker and a first time poster (more like second, but nevermind...)
Back in 2019 I built a pair of two ways, with Dayton DSA135 and DC28 drivers and an 8’’ passive radiator. I thought I could get away with a pre-assembled xover (as many beginners do), and I throw one in. The resulting sound was anticlimactic to say the least…
These boxes have been sitting in my bedroom mainly for playing white noise throughout the night, and ocassionally some music (with little to no joy). Recently, I have decided to use them in the living room in a 2.1 configuration, and possibly as my fronts in a 5.1 setup in the near future. So, I thought I’d better take some measurements and see what’s really wrong with them, and boy, was I horrified…(viewer discretion advised)
tweeter alone:
Apart from the very low quality, pre-assembled xover (which is not worth talking about), the main problem appeared to be the recess I made for the tweeters. My brilliant thinking was ‘let’s align these little guys with the woofers’ but I had no idea that it was gonna cause all kinds of phase issues, as they are clearly evident with these dramatic nulls.
Instead of building new cabinets, I felt I'd much rather save these ones, but something clearly had to be done about those cavities. I figured if I can fit a waveguide in there, the problem would be solved. So I ordered a couple of Visaton wg148’s and while I am at it, I purchased a pair of SB15nrx’s, because why not? I also had a pair of Morel MDT29 tweeters from another half baked project, and I decided to use them with the waveguides. Again, why not? After a nerve racking operation, with virtually zero margin for error, I was able to fit both drivers.
And here are the new measurements (red being the tweeter and orange the woofer)
I have been fiddling with some xover designs in Xsim and I am slowly getting there, but that baffle loss I am getting around 600hz is giving me headaches. The baffle is quite narrow at around 165mm (6.5 inches) and as you can see in the graph, between 1.1k and 500hz, almost 10db is lost. Is this normal behavior or should I be worried? If I am to flatten it out, I am guessing the sensitivity of the speaker is gonna suffer a lot, am I right? (Mind you the cabinet is not properly sealed right now, there are some air leakages here and there because of all the cables and what not. But I am assuming it shouldn’t be a problem at these frequencies.)
I will be updating my progress along the way but as a beginner, I'd love to hear what you people have to say about this project and the baffle loss particularly.
Many thanks, peace.
Back in 2019 I built a pair of two ways, with Dayton DSA135 and DC28 drivers and an 8’’ passive radiator. I thought I could get away with a pre-assembled xover (as many beginners do), and I throw one in. The resulting sound was anticlimactic to say the least…
These boxes have been sitting in my bedroom mainly for playing white noise throughout the night, and ocassionally some music (with little to no joy). Recently, I have decided to use them in the living room in a 2.1 configuration, and possibly as my fronts in a 5.1 setup in the near future. So, I thought I’d better take some measurements and see what’s really wrong with them, and boy, was I horrified…(viewer discretion advised)
tweeter alone:
Apart from the very low quality, pre-assembled xover (which is not worth talking about), the main problem appeared to be the recess I made for the tweeters. My brilliant thinking was ‘let’s align these little guys with the woofers’ but I had no idea that it was gonna cause all kinds of phase issues, as they are clearly evident with these dramatic nulls.
Instead of building new cabinets, I felt I'd much rather save these ones, but something clearly had to be done about those cavities. I figured if I can fit a waveguide in there, the problem would be solved. So I ordered a couple of Visaton wg148’s and while I am at it, I purchased a pair of SB15nrx’s, because why not? I also had a pair of Morel MDT29 tweeters from another half baked project, and I decided to use them with the waveguides. Again, why not? After a nerve racking operation, with virtually zero margin for error, I was able to fit both drivers.
And here are the new measurements (red being the tweeter and orange the woofer)
I have been fiddling with some xover designs in Xsim and I am slowly getting there, but that baffle loss I am getting around 600hz is giving me headaches. The baffle is quite narrow at around 165mm (6.5 inches) and as you can see in the graph, between 1.1k and 500hz, almost 10db is lost. Is this normal behavior or should I be worried? If I am to flatten it out, I am guessing the sensitivity of the speaker is gonna suffer a lot, am I right? (Mind you the cabinet is not properly sealed right now, there are some air leakages here and there because of all the cables and what not. But I am assuming it shouldn’t be a problem at these frequencies.)
I will be updating my progress along the way but as a beginner, I'd love to hear what you people have to say about this project and the baffle loss particularly.
Many thanks, peace.
Attachments
Last edited:
Implanting baffle step compensation in the crossover will sacrifice a bit of sensitivity as you say. However, you could do the baffle step compensation at line level with a small buffer amp to mitigate the loss. This could be built on perf board and run from batteries in a small guitar effects box especially if you use a fairly low power op amp like a TL072. I used the design from Rod Elliot’s sound-au.com site.
My speakers are 15cm wide and use a pair of 5 1/4” woofers.
My speakers are 15cm wide and use a pair of 5 1/4” woofers.
This is the exact topic where I often get confused. What you are suggesting is basically a kind of single band EQ right? Therefore, if I use a graphic eq (say 31 bands) I could simply get my response flattened out. But isn't that also demanding more power from the amp? If that is the case, what is the difference between a flat but low sensitivty speaker, and a speaker that has been equalised to achieve the same flat response.
Last edited:
This thread might give you some ideas about the crossover, although it uses a different tweeter which won't cross as low as the DC28F.
https://www.diyaudio.com/community/threads/2-way-crossover-dsa135-8-5-nd25fw-4-1-soft-dome.389025/
Also, Parts Express sells the "C-note" kit which used the same woofer, that may help too. The schematic is available in the 'manuals and resources' section of that page.
https://www.parts-express.com/C-Note-MT-Bookshelf-Speaker-Kit-Pair-with-Knock-Down-Cabinets-300-7140
I did an Xsim model purely based on Dayton's data, which although considered to be reasonably accurate, is of course not measured in your cabinet. I can post it if you like but it won't be anything like you will get with accurate measurements.
Geoff
https://www.diyaudio.com/community/threads/2-way-crossover-dsa135-8-5-nd25fw-4-1-soft-dome.389025/
Also, Parts Express sells the "C-note" kit which used the same woofer, that may help too. The schematic is available in the 'manuals and resources' section of that page.
https://www.parts-express.com/C-Note-MT-Bookshelf-Speaker-Kit-Pair-with-Knock-Down-Cabinets-300-7140
I did an Xsim model purely based on Dayton's data, which although considered to be reasonably accurate, is of course not measured in your cabinet. I can post it if you like but it won't be anything like you will get with accurate measurements.
Geoff
You aren't losing 10dB due to baffle loss. that's impossible. 4 pi (360 degree) to 2 pi (180 degree) radiation has a "Gain" of 6dB maximum. If you are going by the woofer's response in the 4KHz region, that's woofer breakup (not baffle step).
The midpoint (-3dB) frequency of your baffle step losses can be calculated via F3 = 115 / baffle width (in metres). Assuming your baffle is about 170mm wide, baffle step F3 ~ 676Hz.
Baffle step "Rise" (going up in frequency) is a gradual phenomenon. It's not just a single frequency thing.
Most speakers don't need a full 6dB of baffle step compensation. Often 3 - 4dB is enough. assuming you are using the 8 ohm version of the SB15NRX - which is 86.5dB sensitive after baffle step, you'll be targeting a system sensitivity ~ 82 - 83.5dB (allowing for 3 - 4.5 dB baffle step losses). So you on axis response should be around this.
If you only want to use your gated measurements for crossover design (and not do a quasi-anechoic " full range" measurement on the woofer), you'll want to target the blue line level as your system response:
The midpoint (-3dB) frequency of your baffle step losses can be calculated via F3 = 115 / baffle width (in metres). Assuming your baffle is about 170mm wide, baffle step F3 ~ 676Hz.
Baffle step "Rise" (going up in frequency) is a gradual phenomenon. It's not just a single frequency thing.
Most speakers don't need a full 6dB of baffle step compensation. Often 3 - 4dB is enough. assuming you are using the 8 ohm version of the SB15NRX - which is 86.5dB sensitive after baffle step, you'll be targeting a system sensitivity ~ 82 - 83.5dB (allowing for 3 - 4.5 dB baffle step losses). So you on axis response should be around this.
If you only want to use your gated measurements for crossover design (and not do a quasi-anechoic " full range" measurement on the woofer), you'll want to target the blue line level as your system response:
The shape of the EQ is different for baffle step than in a graphic equalizer, but yes, either way the baffle step loss is compensated by boosting power to the bass. By doing this at line level, you are not throwing away the midrange power of the amp into a resistor, since you can amplify a little to compensate for the loss of midrange I think you would end up with slightly higher SPL at clipping with the line level solution and of course have a simpler crossover with fewer components.
Either will give you the same ultimate response with adequate components, but the components for the line level solution may be rather cheaper.
Either will give you the same ultimate response with adequate components, but the components for the line level solution may be rather cheaper.
Those driver responses look great. You are getting a 6 dB boost in the low end of the tweeter from the waveguide, just as expected and there is a 6 dB baffle step at around 600 Hz. All normal. Crossover at maybe 2000 Hz. You will want to compensate that low frequency rise of the tweeter. A series capacitor will do that, in addition to a 2nd order or 4th order crossover.
Thank you Geoff, but as I mentioned I am not using those drivers anymore. I'll see what I can do with these new ones.
I am rather going by the peaks around 1k and 2k. After carefully inspecting the factory specs of the driver, I realised it loses almost 5 db from 1k down to 700hz. So I think it makes sense that the total loss between 1.1k and 500hz is almost 10db's.
Yes, both drivers are 8ohms. I am indeed targeting the blue line that you indicated, but in that case I am a little concerned about the loss of sensitivity as you mentioned. I am listening at moderate to soft levels, from a 3.5 meters of distance (12 feet), but my amp is a poor little Marantz from the 90s which can only deliver a modest 40watts per channel (at 8ohms). Even though I am planning to get something with more juice, I still want to be able to drive the speakers comfortably until I am able to make that purchase. Maybe I am worrying too much about it, but don't you think 82-83db sensitivity would be on the low side for such an amp?
Mmmm, it kinda makes more sense now, I'll think about this, thank you!
Yes I am also happy with the reponses, especially when I compare them with the previous version's FR
The best crossovers I could put together so far were around 2-2.3 kHz as you guessed. It's not the easiest thing to deal with a waveguide's response on top of such a huge woofer breakup, but I believe I am getting there slowly. I'll try the series capacitor too, thanks!
Last edited:
If you have time to put into it, have a look at Visaton Boxsim free software. The example design in that software uses drivers very similar to yours, including the waveguide. So that example crossover architecture may work well for you with value changes. The boxsim software has an automated optimizer that will change the component values to produce the optimal crossover performance for a given schematic. To use it with non Visaton drivers you just have to load either the TS parameters or load the .frd and .zma files into a project. Unfortunately those can't be saved into the driver library, so you just copy a project file to use them in a new design. It's one of the most complete freeware speaker design programs and well worth learning. https://www.visaton.de/en/literature-software/software I believe the added capacitor after the high pass filter, 33 uF in the schematic, is used to flatten the waveguide low frequency rise in response.
Just need frequency response files and impedance curves
to work something out. FRD / ZMA
and however measurements were taken/ if on center.
then need to know XY or Y coordinates for woofer.
If kept at same position then phase should be right.
3rd or 4th order would easily help with huge baffle step.
or whatever is happening up top
all things learned skinny baffles are not that useful
and 5" speakers dont need waveguide tweeters.
But hey it will make a nice 2 way with a crossover.
The original finish seemed very nice in the original picture.
So should be good after the crossover and refinish.
passive seems well sized.
Dont have off axis for vertical, so will just guess what it is
with the tall center to center spacing.
Be nice to just get them to sum.
In first post just looked like 3K null so tweeter polarity needed
reversing. Not being flat was just baffle step and the odd recessed tweeter
still though the finish was there, looks good in the pic.
Great overall improvement and fitting so far with the waveguide
to work something out. FRD / ZMA
and however measurements were taken/ if on center.
then need to know XY or Y coordinates for woofer.
If kept at same position then phase should be right.
3rd or 4th order would easily help with huge baffle step.
or whatever is happening up top
all things learned skinny baffles are not that useful
and 5" speakers dont need waveguide tweeters.
But hey it will make a nice 2 way with a crossover.
The original finish seemed very nice in the original picture.
So should be good after the crossover and refinish.
passive seems well sized.
Dont have off axis for vertical, so will just guess what it is
with the tall center to center spacing.
Be nice to just get them to sum.
In first post just looked like 3K null so tweeter polarity needed
reversing. Not being flat was just baffle step and the odd recessed tweeter
still though the finish was there, looks good in the pic.
Great overall improvement and fitting so far with the waveguide
Audio Excite- Kensho 2
SB15NRXC30-4 & SB26STAC-C000-4 (vs Morel MDT29)
http://www.audioexcite.com/?page_id=6373
http://www.audioexcite.com/?page_id=4121
Madisound Leap Crossover
Morel MW144/8 & MDT29/4
https://www.madisound.com/library/stacks/MW1448-MDT29-4.pdf
Both crossovers 2000 hz. No waveguide.
SB15NRXC30-4 & SB26STAC-C000-4 (vs Morel MDT29)
http://www.audioexcite.com/?page_id=6373
http://www.audioexcite.com/?page_id=4121
Madisound Leap Crossover
Morel MW144/8 & MDT29/4
https://www.madisound.com/library/stacks/MW1448-MDT29-4.pdf
Both crossovers 2000 hz. No waveguide.
Last edited:
Wild guessing can help.
But 3rd or 4th order is what it will take.
If not, well designed 2nd order with notch
will help pull the baffle step.
And not have the usual annoying 30 ohm peak
at the crossover, as typically seen
just need frd and zma
it will be a hair pull to knock out the baffle step
and excessive waveguide response.
even if 2nd order
extra cap will be added to make tweeter
on 3rd order. Guarantee less distortion with higher order.
No need to fuss with phase complications.
So as mentioned just use a 3rd order for both
Why stop there. The woofer response will be more
ideal = 4th order
So bog standard 4th order linkwitz likely
be it. Or even Butterworth/Bessel
depending how that waveguide response
needs wrestling to be flat.
On axis at least, without vertical data
to see how much the center to center spacing
is a challenge.
Which would be a much much lower crossover.
Something the waveguide can help with.
Otherwise a very small 3/4" tweeter with neo
would make center to center way easier ( closer together)
But 3rd or 4th order is what it will take.
If not, well designed 2nd order with notch
will help pull the baffle step.
And not have the usual annoying 30 ohm peak
at the crossover, as typically seen
just need frd and zma
it will be a hair pull to knock out the baffle step
and excessive waveguide response.
even if 2nd order
extra cap will be added to make tweeter
on 3rd order. Guarantee less distortion with higher order.
No need to fuss with phase complications.
So as mentioned just use a 3rd order for both
Why stop there. The woofer response will be more
ideal = 4th order
So bog standard 4th order linkwitz likely
be it. Or even Butterworth/Bessel
depending how that waveguide response
needs wrestling to be flat.
On axis at least, without vertical data
to see how much the center to center spacing
is a challenge.
Which would be a much much lower crossover.
Something the waveguide can help with.
Otherwise a very small 3/4" tweeter with neo
would make center to center way easier ( closer together)
Last edited:
I don't know your cabinet dimensions nor PR Vd and mass, but I am assuming your SB15NRX will only take maybe 20 watts between 30 - 50Hz before it reaches Xmax anyway, so your Marantz amp should have enough power to drive these to their potential assuming you play them full range,
At bass frequencies, drivers become excursion not thermally limited when it comes to power handling.
At bass frequencies, drivers become excursion not thermally limited when it comes to power handling.
Thanks for the recommendation, I'll try to find some time to have a look at it, I am not having much fun with Xsim to be honest...
I have ended up at some 3rd order 4th order asymetric filters indeed, mostly around 2.3-2.5k They look nice on paper, and the reverse null was there when I inverted. However I am kinda concerned about the parts count, 'since best crossover is no crossover' as they say. Diy audio is not really a thing in my country which makes it difficult to come by audio grade components. (Whenever some friend living abroad is visiting Istanbul, I ask them to bring me some drivers and what not) Therefore, I will be using some inexpensive polycaps (those infamous yellow ones) and I am not sure if going higher on the filter order is the best idea. But judging by the FR, that might be the only option really...
I have seen this project the other day, but the woofer's measurements are interesting. The breakup is quite lower comparing to my measurements and the factory specs. I wonder if he made some modifications on the woofer...
So, according to this I don't really need much, considering those regions will be covered by an active sub anyways. It almost sounds too good to be true but what do I know, maybe I should post another thread about my amp needs
Here is a crossover design I put together:
.
I am getting quite a null in reverse polarity (should it be better?)
The overall response looks kind of satisfying, but I am concerned about the woofer's persistent break up. I was able to knock it down with a 3rd order filter but then I had trouble getting the reverse null. Do you think it's fine if I leave it like this or is it going to be audible?
I am getting quite a null in reverse polarity (should it be better?)
The overall response looks kind of satisfying, but I am concerned about the woofer's persistent break up. I was able to knock it down with a 3rd order filter but then I had trouble getting the reverse null. Do you think it's fine if I leave it like this or is it going to be audible?
