Advices on First Crossover Design (VituixCAD2)

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
VituixCAD and my quest to learn speakers design

Hi Everyone,

I hope you can enlight me with some sound advice's on my project

A bit of background
I have nearly no experience in the speaker building, long ago I only had to build a subwoofer for a car and spent time on a IASCA oriented community forum for guidance. I am not an electronician or good in math but will try to fill a bit the gap as I go.


My Goal is first to learn best/bad practice about speaker cabinet & crossover design and manufacture, then to end with an acceptable(basic) level of knowledge and skill that will allow me to realize projects with less holding hand from you.



What do I listen music on? :
Often from my computer using the UMC204HD with my equalized headphone Beyer dynamic DT 990 Edition 250 Ohm, I Like the details I get from the music

Sometimes also from the computer but linked to a integrated amplifier KODA kd-261 or an Onkyo A-7070 with a pair of Cabasse Jersey 221 (4 ohm)
So my signal chain : Computer G1.Sniper Z97 USB -> UMC204HD -> TS/TRS ouput -> RCA input on Amp

I have no immediate plan to change anything else but I'd like later to find a way to be in Balanced or optical. (Ground loop)



What music? (I don't listen to) :

It's easier to start to cite what I am NOT likely to listen to ; Metal, Trap, Jungle, Hip Hop, Rap, J & K Pop, Jazz, meditation, easy listening, Psychedelic, Yodle, and probably others I don't know the name.
I will cite some Artists that come to my mind that I like; R.L. Burnside, Art of Noise, Shu De, Mike Oldfield, Kenji Kawai, Sade, Meute, Nana Mouskouri, Son Lux, Eivor, Talvin Singh.




New project (In progress)
I have bought some tools & components to help me in this project

- Dayton Audio DATS V2
- Dayton Audio EMM-6 microphone with a stand
- Behringer UMC204HD
- Cables & Adapters
- The T.amp 40C

Due to the unforeseen failures of the Scan-Speak and the LS Plan, new drivers have been purchased

-2x Gradient Acoustics Axis AXP-08
https://www.intertechnik.com/media/1381683_88463.pdf
The Woofers will be in a (slightly) ported configuration in their own individual enclosures
Each had a run-in for 6h close to resonant frequency of the drivers without much improvement on the response the Fs is higher than specsheet

-2x Fostex FF85WK
https://www.fostexinternational.com/docs/speaker_components/pdf/ffwk.pdf
The full range drivers will have their own enclosure, with a sort of open back/dipole configuration and could be if needed reinforced in the top end with Tweeters

The actual foreseen speaker design is not engraved in stone but is likely to be the following :



Shape : Floor-standing cabinets on casters to allow proper placement
Ways : 2 or 3

Impedance : near 8 ohms
Sensitivity : Close to 90db
Frequency Range : As low as I can get the AXP-08 without compromising the sound quality, can still reinforce the low end after by adding a sub-woofer in the future if need be. And for the top end the usual 20khz
Crossover Type : Passive (for the learning experience)

Budget : Ideally around €600 (Drivers, Cabinets & Filters)





What has been done?
- Sketchup of potential candidate for the cabinets and drivers placement but most likely to be modified (no more LS-Plan)

https://i.imgur.com/IC8Fcu2.png
https://i.imgur.com/vDOII86.png
https://i.imgur.com/tf1LgQL.png
- DATSV2 measurements AXP-08 & FF85WK.
Gradient Axis AXP-08

D 165 [mm]
Re 6.7714 [Ohms]
Fs 61.235 [Hz]
Zm 81.233 [Ohms]
BL 8.7957 [N/A]
Qms 5.4791
Qes 0.49826
Qts 0.45672
Vas 29.317 [liters]
L10k 0.24316 [mH]
n0 1.2882 [%]
dBSPL 93.2 [1W/1m]
Ms 14.796 [grams]
Cms 0.457 [mm/N
Fostex FF85WK
D 60 [mm]
Re 7.2852 [Ohms]
Fs 135.26 [Hz]
Zm 40.863 [Ohms]
BL 3.7808 [N/A]
Qms 3.5005
Qes 0.75949
Qts 0.62408
Vas 0.88652 [liters]
L10k 0.23637 [mH]
n0 0.27542 [%]
dBSPL 86.5 [1W/1m]
Ms 1.7535 [grams]
Cms 0.79 [mm/N]
- A simulation has been done with VituixCAD for the ported alignment for the AXP-08
https://i.imgur.com/WpqaCtX.png
- Have familiarized myself to make SPL measurements in REW, and tested some speakers in different but not optimal conditions, not trying anymore to actually reach actual SPL of drivers just the difference and common patterns between the AXP-08 & FF85WK that may indicate that it's linked to the way it's done more than response
https://i.imgur.com/j17lZ62.png
Many things where bad with that measurement, soundcard calibration and settings in computer, ended up going better with the FF85WK
https://i.imgur.com/NT2gfvc.png
- IEC baffle is in planning and will be started soon (still waiting on tools & materials)



Initial project (Obsolete)
I have some drivers at home that I have not used in a long time and didn't want to sell them, but now I am thinking about using them to build speakers and see if it is a disaster or not.

Drivers for this first project:

Tec LS Plan (Tweeter)
In Possession : Yes a pair
Frequency range : 5-30khz
Nominal Impedance 6.00 ohm
Sensitivity, Spl (2.83V,1m) : 92 dB
Nominal Power : 40 Watt

Scanspeak 13M/4535 (Full Range)
In Possession : Yes three
Frequency range : ? (Frequency curve seem to indicate best between 150Hz and 11Khz)
Nominal Impedance 4.00 ohm
Sensitivity, Spl (2.83V,1m) : 90 dB
Nominal Power : ? Watt
Resonance Frequency : 56hz

Scanspeak 22W/8534G00 (Bass Woofer) wished for a 25w-8565-00 but it's a lot more money
In Possession : No
Frequency range : ? (Frequency curve seem to indicate best up to 800hz)
Nominal Impedance 8.00 ohm
Sensitivity, Spl (2.83V,1m) : 88.8 dB
Nominal Power : 70/120 Watt ?
Resonance Frequency : 30hz


I have tested a few softwares to help me design passive filters for the crossover schema but finally stayed with VituixCAD2

The simulation

F54zun2.png


So what do you think will this work? Am I missing something ? Can I improve ?


Thanks
 
Last edited:
..1st: show us range of off-axis results, particularly attention to the top of the mid. and the bottom of the tweeter.

2nd: the 1.8kHz to 3khz will be audible as "bright"/"aggressive" sounding, at least on-axis (as shown).

3rd: is this including the effects of the baffle (including driver placement), and if so - at what distance? Also, is it including driver placement near a boundary (like the floor)?

4th: looking at the impedance, do not use this with an amplifier that has a higher output impedance (..in other words, ONLY use this with a solid-state amp). If you use something like a typical tube amp the midrange will become elevated relative to the rest of the freq. range because of the interaction with the amp and the Impedance.


-once you've got 1 & 3 clarified, people should be able to give you better advice.
 
Last edited:
Thanks for your reply ScottG,


1- I have not yet measured the off-axis response of the drivers I am still learning the ropes of REW and at the moment the ropes are full of knots


2- How can I attenuated that in the design? It would be better I guess than having to equalize that after.



3- No it does not include the box yet, I am trying to understand how to implement a crossover, from what I have read the planar tweeter have less reflexion from floor or ceiling as long as I am having it mounted vertical (so I hope) and I have no idea how to include that in the software yet.

4- I was not planning to use a tube amp but it is good to know that additional factor


I will get back here as soon as I have at least the first point, because the third I have no idea how to work with that.


Thanks
 
1. ..no, a SIMULATION of the off-axis response in VituixCAD.

2. you wouldn't do that until finding out about the off-axis performance in that range. Most people listen off-axis, so it might not be a big deal.

3. well, this (along with #1) let's us know that you need to spend more time with the software to understand what's required for the design. Baffle effects play a HUGE roll in the resulting response, and additionally there is the particular effect called baffle-step loss. Vertical directivity isn't as important as horizontal as you start going above 6-7.5 kHz (..and horizontal is always important). It can be more important as you near the crossover-region depending on the distance and angle from the loudspeaker (..and as you get closer to the loudspeaker it becomes more important).
 
Last edited:
Thanks ;)


So here is my progress with adding directivity in the design,I believe it is correctly done but I could be wrong as it is a first for me.

Normally I should have the 4535 in a sealed compartment inside a bigger ported box, the tweeter is closed anyway

I notice that there is big bump in the low frequency due to the impedance response taken from the enclosure simulation is that linked to the port ?

I have no experience interpreting the directivity graph, I just deduce that as usual the low frequency will be heard everywhere and the highest only straight in front

I have a bit modified the crossover also to simplify


Here are the screens :


pCNdF61.png


pkFoR4P.png


u8M6klo.png


Thanks for your insights
 
-was it with the baffle you want? :confused:

It looks like the width is barely larger than the 22W/8534G00 itself with the "flare" in horizontal response near 1.5 kHz.


Right now the area between 1-2 kHz is not good, there is actually a rise/"bump" in response off-axis around 1.5 khz, and it's overall "hot"/elevated in this region when compared to the result 500 Hz to 1 kHz.

Verticals are a mess between 3 and 8 kHz.


Also, what is the distance for the plot? 2 meters, 1 meter? It makes a difference with regard to baffle step loss.
 
Last edited:
Hi ScottG,


I think I have messed the dimension of the enclosure in VituixCAD, and maybe one area that I have it wrong is the microphone placement in the diffraction, should I place the mic in front of the speaker each time or keep it in the center ?



I have to check it again


What do you think can cause that dispersion between 1 & 2 Khz ?


For the vertical I think it could be wrong, I suppose the LS Plan is simulated as a circular speaker and not as a planar, I will dig a bit that aspect


I am modeling a speaker that I can use as a reference to make the simulation


Any idea why there is that big bump between 50 & 100 hz ?


Thanks a lot for your help
 
Hi ScottG,


I think I have messed the dimension of the enclosure in VituixCAD, and maybe one area that I have it wrong is the microphone placement in the diffraction, should I place the mic in front of the speaker each time or keep it in the center ?



I have to check it again


What do you think can cause that dispersion between 1 & 2 Khz ?


For the vertical I think it could be wrong, I suppose the LS Plan is simulated as a circular speaker and not as a planar, I will dig a bit that aspect


I am modeling a speaker that I can use as a reference to make the simulation


Any idea why there is that big bump between 50 & 100 hz ?


Thanks a lot for your help


I'm not familiar with VituixCAD for saying what to do. :eek: (..I know of it's excellent capabilities generally, but that's about it.) ..mic placement should generally be centered on the tweeter's 0 degree axis at either 1 meter (for baffle-step loss shown), or 2+ meters (without baffle-step loss). At least in real life (..but it could be different in VituixCAD.) :eek: Usually you are targeting/simulating an ANECHOIC condition (in other words no effects from any reflections or other room effects).


Again, that horizontal "flare" looks like a baffle artifact to me.

My guess is the elevated bass response is an artifact of enclosure (relative to driver parameters) and crossover. It's really not as bad as it looks though because the upper-end of the driver should be crossed-over higher in freq. (ie. it's low-pass filter needs to be raised to a higher freq.), plus you'll have floor (boundary) gain here.


General rule for midrange to woofer:

Target a high-pass crossover for the midrange near 300-500 Hz depending on its height from the floor to minimize "floor-bounce" (..do a search on that phrase).

Meanwhile the woofer should be placed closer to the floor not only for some boundary gain from the floor (which can be up to 3 db), but also to raise the freq. for it's "bounce" effect - allowing you to crossover (low-pass) higher in freq..


Ex. Magico S5, while certainly not perfect - it has a reasonably good design:

SoundStage! Hi-Fi | SoundStageHiFi.com - Magico S5 Loudspeakers

In particular note that the 2 meter graphs (..plotted at 1 meter for spl-reference) ANECHOIC:

SoundStageNetwork.com | SoundStage.com - NRC Measurements: Magico S5 Loudspeakers

Note that there is about 2db of pressure loss from the average between 50-200 Hz. The floor will add-in a bit of gain here to the result, so that loss isn't bad.

Look to fig. 5 here for IN-ROOM results (for the MKII version of the S5):

Magico S5 Mk.II loudspeaker Measurements | Stereophile.com

Note that fig. 2 is at *1.27* meters ANECHOIC (as opposed to the in-room above) and has some gain because of the baffle-step compensation. The crossover there is just above 300 Hz and the woofer is near the floor with the midrange up quite a bit off of the floor. Additionally, the filter is "steeper" to further limit floor-bounce. (..and it's not that it removes floor-bounce as seen in the IN-ROOM response 280-500 Hz dip, but the result is less than it might have been.)
 
Last edited:
Hi ScottG,


Again thank for your help on the subject, the VituixCAD thread was indeed an excellent suggestion as Kimmosto is very responsive and helpful for me to understand what aspect of the program needed my attention in this initial step.


The crossover design will be refined especially to match existing and available parts in my country.



I think I could settle with the following results for the baffle design, at least it seem to me I have less potential issues than before, what do you think ?


Sketch : I will leave the woofer higher to lessen that potential 3dB increase but will put the port closer to the floor at the back, I don't know if it's true but that seem to be an OK practice from what I have read, the port size & shape is not yet totally decided, I could still raise the baffle if needed.
r4OzF63.png




VituixCAD
3mLcPK8.png


mYRvEU7.png


RsgZA3m.png




Don't hesitate if you see something that is wrong
 
..floor-standing looks pretty good assuming anechoic 3 meters up to 500 Hz.

Liked to see a bit more baffle-width and larger radius lateral cabinet edges (..that latter is more important than the former).

1.2-to-4 kHz is to elevated, particularly considering the exaggerated horizontal "flair".

Tweeter high-pass should be lowered to at least 4 kHz.. maybe 3.5 kHz.
 
Last edited:
..floor-standing looks pretty good assuming anechoic 3 meters up to 500 Hz.

It's just the diffraction simulation from VituixCAD but the listening distance set to 10000cm


Liked to see a bit more baffle-width and larger radius lateral cabinet edges (..that latter is more important than the former).

I am a bit constrained on the width of the baffles unfortunately this is already a bit more than I should have, as for the edges rounding I can't do more than close to 17mm radius with 18mm MDF, or at least i don't see how.


1.2-to-4 kHz is to elevated, particularly considering the exaggerated horizontal "flair".

Yeah that flair is moving around but I don't find what I can do to get rid of it, I just moved it a bit by playing with the mid/high placement (offset)


Tweeter high-pass should be lowered to at least 4 kHz.. maybe 3.5 kHz.


I will try to go a bit lower but this type of ribbon tweeters are not made to go low, I think 4.5Khz are the lowest crossover point


Thanks for you reply
 
-still need to go lower with the tweeter. As far as the tweeter's capability, it's a matter of distortion and power handling/excursion, and it has a lot of surface area. My guess is that that with your filter and summed response that 3.5 kHz should be OK, but it's something you can model. (..though you should have distortion measurements of the tweeter (IB) to do that properly.) More problematic would be the pressure drop of the tweeter itself, which to an extent should have at least some correlation with the volume enclosure you have for it (..and it's a dipolar design without a rear chamber - so you should be able to get a bit more extension from it with the correct rear-chamber volume for your cross-over.)

As for the horizontal "flair".. don't worry about it too much (..if you can't really do much with the box *size and edges), instead worry about it in conjunction with the on-axis freq. response. In this case you simply have to much pressure between about 1 kHz and 9 kHz. BECAUSE of the horizontal "flair" it should actually be DEPRESSED/LOWER in pressure near the trouble-spot around 2 kHz. Based on your graph: 91 db is the average; that 2 kHz region should be about 90 db, NOT the current 93 db. BTW, this will also "help out" with the lower high-pass for the tweeter, because the overall pressure is lower (..near and lower than its high-pass).

*Note: the edges are more important than the width (as far as minor changes are concerned) - if you can *narrow* the baffle width a bit but increase the round-over of those lateral edges, you'll be better-off.
 
Last edited:
-still need to go lower with the tweeter. As far as the tweeter's capability, it's a matter of distortion and power handling/excursion, and it has a lot of surface area. My guess is that that with your filter and summed response that 3.5 kHz should be OK, but it's something you can model. (..though you should have distortion measurements of the tweeter (IB) to do that properly.) More problematic would be the pressure drop of the tweeter itself, which to an extent should have at least some correlation with the volume enclosure you have for it (..and it's a dipolar design without a rear chamber - so you should be able to get a bit more extension from it with the correct rear-chamber volume for your cross-over.)

As for the horizontal "flair".. don't worry about it too much (..if you can't really do much with the box *size and edges), instead worry about it in conjunction with the on-axis freq. response. In this case you simply have to much pressure between about 1 kHz and 9 kHz. BECAUSE of the horizontal "flair" it should actually be DEPRESSED/LOWER in pressure near the trouble-spot around 2 kHz. Based on your graph: 91 db is the average; that 2 kHz region should be about 90 db, NOT the current 93 db. BTW, this will also "help out" with the lower high-pass for the tweeter, because the overall pressure is lower (..near and lower than its high-pass).

*Note: the edges are more important than the width (as far as minor changes are concerned) - if you can *narrow* the baffle width a bit but increase the round-over of those lateral edges, you'll be better-off.


Sorry but this time you lost me, most of what you write I have no clue about


I have lowered more the tweeter filter and the mid higher filter to 3.5 Khz, I have also attenuated everything closer to 90dB, but then I have a big issue around 1Khz, the impedance there is going as low as 2 ohm (My amp wont like that I think)


Here is a bunch on sentences where I am lost :




  1. When you speak about pressure drop do you say that because I am using a vent?
  2. when you say it's a dipole design, is it based only because of the event oriented toward the back?
  3. You completely lost me creating depression to lower between 1 & 9Khz, is that a design that involve changing the volume/shape of the baffle or something related to passive electronic design ? (like I do with L-Pad Attenuation)
  4. I could in theory make the edge rounder on the baffle design but then It would be difficult to match that on the the MDF, do you have an example for me to understand what you mean ?
  5. That bump now between 15Khz & 30Khz is stubborn, but it don't matter i think I could not hear that.
I thought I was getting close but as I go it gets harder, thanks that you are sticking with me while I stumble during that process.


jIIs3rD.png


eZTUmeL.png
 
I have lowered more the tweeter filter and the mid higher filter to 3.5 Khz, I have also attenuated everything closer to 90dB, but then I have a big issue around 1Khz, the impedance there is going as low as 2 ohm (My amp wont like that I think)


Yup. The Impedance is a problem. "Massage" the crossover a bit more.

HOWEVER, excepting that problem it looks a LOT better with the one exception of pressure drop in the 3 kHz range. Still, that's a narrow-band "dip" - and perceptually that won't be easily heard. So :up: with respect to linearity on and off-axis horizontally.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.