Advices on First Crossover Design (VituixCAD2)

[ScottG]

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.

1. I said "drop in pressure" regarding the tweeter. What I'm referring to is the tweeter's natural high-pass character (withOUT an electrical filter).

2. The tweeter is a "dipole" (or dipolar actually) because it has a rear opening as seen on the picture of the driver on its website. You won't actually be using it as a dipole of course, and instead will create its own "sub" enclosure within the loudspeaker - effectively making it a monopole. Any sort of volume will effect the driver's natural high-pass character - which is what I was trying to convey.

3. ..don't worry about this, your latest design with regard to linearity on and off-axis is very close to what it should be.

4. I mean shrink the baffle width to about the size of the woofer (..I know, it's already pretty close to that), and then use larger radius roundover's for the vertical edges of the cabinet. It's a US source, but this is where I source my large roundover stock for this sort of use:

QUARTER ROUNDS

5. Don't worry about added pressure from 15 - 20 khz. In a planar that tends to just sound like more "air".

 

[jplesset]

I would also be strongly tempted to loose the l-pad on the woofer. Nothing good comes from wasting power there.

 

[kimmosto]

L-pads have shorting wires over the series resistor so they don't have any effect. Although proper way to short components is using Shorted command in context menu.

I suppose none of those component values and perhaps whole circuit will not be final because responses are traced and actual acoustical centers are probably something else than in the simulation. Maybe useless to repeat anymore here that XO simulation before actual measurements in project cabinet is waste of time. It's not directly my time, but doing design against instructions sucks energy also from advisers and might give impression to other new users that this procedure is okay.

Simple choice: either you read and believe "designing loudspeaker without measurements" or not. If answer yes, VituixCAD is not for you

 

[ScottG]

-yeah, it's almost certainly going to be different than modeled.. On the other hand though, it is a good learning experience.

This is my general work-flow:

https://www.diyaudio.com/forums/mul...hoosing-particular-drivers-2.html#post5303468

With this design I'd add-in measuring and listening with different enclosure volumes for the tweeter.. and perhaps the midrange (..though I already have a good idea of how I like to load midrange drivers).

 

[MrZoltan]

L-pads have shorting wires over the series resistor so they don't have any effect. Although proper way to short components is using Shorted command in context menu.

I suppose none of those component values and perhaps whole circuit will not be final because responses are traced and actual acoustical centers are probably something else than in the simulation. Maybe useless to repeat anymore here that XO simulation before actual measurements in project cabinet is waste of time. It's not directly my time, but doing design against instructions sucks energy also from advisers and might give impression to other new users that this procedure is okay.

Simple choice: either you read and believe "designing loudspeaker without measurements" or not. If answer yes, VituixCAD is not for you

Hi,


It's a very complex subject for me to build a speaker the proper way instead of throwing everything in a box with aftermarket crossover, as I said before I don't have background that help that but I have determination, in any case I doubt that i will be able to understand it all in the end.



I don't want to give that impression that I want to bypass best practices, I just want to understand some essential things before I commit to actually build something, and I feel I am getting close but I still have some information's that are lacking


I have bought measurements tools and I intent to put them to good uses



I guess It can be painful for you guys to seemingly have a never-ending flow neophytes, so far VituixCAD has a steep learning curve but an enriching experience to me.


Thanks for your patience

 

[MrZoltan]

-yeah, it's almost certainly going to be different than modeled.. On the other hand though, it is a good learning experience.

This is my general work-flow:

https://www.diyaudio.com/forums/mul...hoosing-particular-drivers-2.html#post5303468

With this design I'd add-in measuring and listening with different enclosure volumes for the tweeter.. and perhaps the midrange (..though I already have a good idea of how I like to load midrange drivers).

I confirm it is a good experience for me, spending time in a simulation helped me on many subjects that I had no idea existed :


1- Diffraction on the front baffle, I had no idea that drivers could be impacted by the shape of the wood panel they are affixed on.


2- I understand now why the driver selection is so important, the constrain I have put myself into to match speakers I already have that does not have the same sensitivity or impedance.

3- L-Pads, Zobel usages where unknown and I can now see the result of the different crossover networks and see how it impact frequency, impedance and diffraction response

4- Simulating that crossover allowed me to see what I am getting into in concern to availability of parts in my country, adjusting the values to match real world and seeing the cost it represent to build one.

5- And much more that I learned and that I hope to learn next

Thanks I will have a look at you workflow, it may be not efficient but I will post a last simulation and ask a few more questions if you still have the heart to answer.


Thanks

 

[MrZoltan]

The reason I have to not go to the build process yet is that I am without an area to build, and once I commit to order the mdf and work on it I have to arrange myself with someone to provide the tools & space to do so.

So ScottG since the last time you asked me to massage the crossover a bit, I put myself into it and had to take a break and do some more research as I was fumbling around like a mad man with all the parameters.

Before I go on with my last iteration of that crossover, and I accept it could different once in a baffle, the few questions :

1- About that driver in an infinite baffle (a mdf plank) do the measurement of the actual drivers make it more relevant than the factory data traced? Is it only to see the difference between the one in my possession and the factory one ?

2- Near-field measurement with the microphone close from each drivers is doable in my situation, but far-field could be an issue due to the spatial area I am working with, have you any suggestion for me to read that would explain how I can do that without going in the street with all my stuff ? (I just can' imagine me on the sidewalk in that busy street pulling an extension cord and hoping it does not rain)

3- Now that I am understanding better how to place that frequency and impedance curve how do I interpret if that phase and group delay is acceptable?

4- Is the simulation representative of the phases I would have in an enclosure?

5- I see often that some drivers (tweeters and mids) are offset on the Z axis in relation to the woofer, I have read that this in relation to the magnets alignments and that it help to receive the frequency at the same time at our ears, I played a bit with time delay on a subwoofer I had in the car that made a huge difference to move the sub to the front, but now the question, I have played a bit in the simulation by displacing the tweeter, the mid and even the two, but in all case it seemed to make it worse, is it a limitation on the sim or my knowledge ?


So the last iteration of the Xover, is it going better ?

Thanks

 

[ScottG]

1- About that driver in an infinite baffle (a mdf plank) do the measurement of the actual drivers make it more relevant than the factory data traced? Is it only to see the difference between the one in my possession and the factory one ?

2- Near-field measurement with the microphone close from each drivers is doable in my situation, but far-field could be an issue due to the spatial area I am working with, have you any suggestion for me to read that would explain how I can do that without going in the street with all my stuff ? (I just can' imagine me on the sidewalk in that busy street pulling an extension cord and hoping it does not rain)

3- Now that I am understanding better how to place that frequency and impedance curve how do I interpret if that phase and group delay is acceptable?

4- Is the simulation representative of the phases I would have in an enclosure?

5- I see often that some drivers (tweeters and mids) are offset on the Z axis in relation to the woofer, I have read that this in relation to the magnets alignments and that it help to receive the frequency at the same time at our ears, I played a bit with time delay on a subwoofer I had in the car that made a huge difference to move the sub to the front, but now the question, I have played a bit in the simulation by displacing the tweeter, the mid and even the two, but in all case it seemed to make it worse, is it a limitation on the sim or my knowledge ?


6- So the last iteration of the Xover, is it going better ?

1. When it comes to upper freq. response you don't do it without measurements. This is particularly true for a tweeter where the final rear enclosure/volume and resulting high-pass roll-off is in question. The midrange's upper response is also crucial. NEVER rely on manufacturer data for this. ..Now a woofer in a 3-way? Ehh.. You should also have T/S param's measured for your Mid. and Woofer.. (it's not AS crucial, but it's still very important - because manufacturer's spec.s can be off quite a bit.)

2. If you can't find at least 4 feet of clear space, consider another hobby. You do far-field with gating for higher freq.s (usually 300 Hz up), and you try to setting the gating at the lowest freq. that's practical (for the room). Near-field for your lower freq.s. and then use the measuring program's feature for combining the two responses for a full-range measurement (..this is for your Mid. and Woofer).

YouTube
Gating:
REW gate settings, help please. -

Techtalk Speaker Building, Audio, Video Discussion Forum

I've done measurements before where the mic. was hanging from the ceiling and the measurement baffle was on the floor (..well, about 1.5 feet up from the floor actually), pretty much centered in the room away from all points of reflection and "pointed"-upward at the ceiling (..like a table). Just a 5-by-5 foot "table"/baffle should do it. (..and I've even used (vibration) damped cardboard before (with tape) to increase baffle dimensions.)

3. That's a "whole 'nuther topic". What I will say is that you don't want any impedance "swings" where it dips low into Impedance while (effectively) intersecting a substantial dip in phase ("steep" negative phase angle). That sort of thing tends to suck amperage from the partnering amplifier, particularly if it happens lower (in freq.) than the midrange. Anyway, something for you to ask about from others and to learn on your own.

4. Depends on the driver's actual T/S and the enclosure.. If they are close to your model then it should be fairly close. "Leaky" enclosures (like from a poorly mounted driver) can screw this up some. (..it really messes up bass-reflexes when that happens.)

5. While this can be done for time alignment (usually with a reference listening axis that is OFF of the tweeter's 0 degree axis), it is most often done to improve the resulting response in relation to BAFFLE DIFFRACTION. Try moving around the mid. and tweeter relative to the baffle to see changes in freq. response.

6. The latest is a bit worse as far as freq. is concerned. 1.2 kHz to 2.8 kHz is at least 1db to high. 3.8 kHz to 5 kHz is also at least 1 db to high. The woofer needs a bit more pressure overall. I'd start by moving the low-pass for the woofer a bit higher in freq. to see what the net effect is in the midrange. Tweeter high-pass still needs some work as well..

 

[faridj]

Hello ScottG I read the details on l-pad and need some expert advice. I am working on a speaker project (my first) where my tweeters and mid range is high sensitivity compared to woofer. I have the TS parameters, impedance measurement for all drivers. Do you think you can help me with the required ideal crossover?

 

[faridj]

I am using the following components:

Tweeter & Midrange
Cerwin Vega V465C
Sensitivity: 94 dB
4 Ohms
100W RMS
Frequency response 55Hz to 20,000Hz

Woofer
Cerwin Vega V84D
Sensitivity: 82 dB
4 Ohms
250W RMS
Frequency response 20Hz to 300Hz

I have all the data just need some advice prior to doing a frequency measurements.

 

[ScottG]

Hello ScottG I read the details on l-pad and need some expert advice. I am working on a speaker project (my first) where my tweeters and mid range is high sensitivity compared to woofer. I have the TS parameters, impedance measurement for all drivers. Do you think you can help me with the required ideal crossover?

That's not my area of expertise/interest. (..see my previous response at #3 above.)

-really in this thread I'm just trying to get Zoltan up to the point where he can better interact with others for advice on this topic (..along with some other general advice).

A forum member like "wolf_teeth" is quite a bit better for the more technical details on crossover design.

 

[kimmosto]

...have you any suggestion for me to read that would explain how I can do that without going in the street with all my stuff ? (I just can' imagine me on the sidewalk in that busy street pulling an extension cord and hoping it does not rain)

Few times already linked and mentioned how far field measurements can and should be done at home and merged to near field which is corrected with baffle effect response by diffraction simulation. No larger than 10m2 room needed
VituixCAD Measurement Preparations.pdf.

 

[MrZoltan]

Hi ScottG,


Thanks for the answers

I have between to 2 and 3 meter (6.5 to 10 feet) length and 2 meters (6.5) width to work with, and my ceilings are also around 10 feet high so it ok

Well I think it's time I stop fiddling with the simulation and prepare myself with REW and the gated measurement procedure, at least that way I will see what the actual responses are on the drivers, excluding the Scanspeak 8535 that I don't have.


I will have a look at the links you posted and see if that is enough for me to understand what I have to do, if not I'll search for more.


I will practice with my actual baffles (Cabasse MT221) and once I have it right I'll test the individual drivers



That will take me some time I guess

 

[MrZoltan]

Few times already linked and mentioned how far field measurements can and should be done at home and merged to near field which is corrected with baffle effect response by diffraction simulation. No larger than 10m2 room needed
VituixCAD Measurement Preparations.pdf.

Thanks I have saved that one already before from the VituixCAD thread, I read it but some parts of the procedure are still a bit obscure for me


And I will have to familiarize myself with REW instead than ARTA or CLIO, but I suppose they are pretty similar


I like the idea of a turning table

 

[kimmosto]

I will have to familiarize myself with REW instead than ARTA

Some differences exist and therefore REW is not recommended and supported tool with VituixCAD.

 

[MrZoltan]

Some differences exist and therefore REW is not recommended and supported tool with VituixCAD.

Hi Kimmosto,


Do you mind to elaborate on this? I have tried to dig a bit in the forum and didn't find much about those differences


As in your documentation you mentions this :


Supported frequency and impedance response file types
VituixCAD supports tab, space or semicolon delimited .txt or .frd or .zma (for impedance). Following software exports are supported:


AudioTools
ARTA, LIMP
Clio
Edge
FRD tools
HOLM Impulse
justMLS
Klippel
LMS short ascii
LspLAB
REW
SoundEasy

Is there something that would justify that I purchase a ARTA personal license ?


Thanks

 

[MrZoltan]

FYI, I had some time to setup & experiment with REW


Had some issues with the setup of the loopback and adjust the levels but now I seem to have some readable results (most of it)


I hadn't time to play with the delay for the gated measurements but I will get there when I have some alone time, at the moment I have reached the limit of the waf (new thing in my vocabulary)


Have a nice weekend!

 

[MrZoltan]

Some differences exist and therefore REW is not recommended and supported tool with VituixCAD.

Nevermind, I just found this post from you in another thread :

I've wrote that to quite a few threads maybe dozen times. So, Umik-1 etc. or any other random latency gear or software what forces user to "single channel measurement mode" is not recommended for speaker designing because timing/phase differences between different measurements will be lost and timing (IR response) normalized by the software. Capability to capture timing difference to output signal of power amp is very basic requirement with advanced multi-way simulation because it enables stable and reliable off-axis simulation.
Another very handy feature is polar response measurement and export sequences. That is included e.g. in ARTA and CLIO.
I suppose REW is still using date+time file naming which is worthless with off-axis measurements.

 

[kimmosto]

Do you mind to elaborate on this? I have tried to dig a bit in the forum and didn't find much about those differences

REW:
* No measurement sequence with automatic file naming: root filename + off-axis angle in degrees
* No measurement sequence with pause timer
* No support for motorized turntable
* VituixCAD does not read mdat so time windowing, smoothing and exporting to frd/txt files must be done with REW. This is not technical problem or missing feature. Just that separate instructions are not written for REW, and I don't give technical support to REW users for measurement gear, program settings, measurements in dual channel mode, time windowing and exporting due to previous items. This means that changes to get valid support, perform measurements and post processing - how it should be done with VituixCAD - are low(er) with REW despite of few years experience. So far I have not seen valid VCAD simulation measured with REW though it's technically possible.

 

[ScottG]

I suppose REW is still using date+time file naming which is worthless with off-axis measurements.[/I]

REW:
..smoothing and exporting to frd/txt files must be done with REW..

I don't know if this helps (or confirms problems) but:

referencing pg. 230:

https://www.roomeqwizard.com/REWhelp.pdf


"Export -> All measurements as text

Export the measured data for all loaded measurements as text files in a chosen directory. The file names use the measurement titles. The same settings dialog appears as for a single measurement, the settings are applied to all the files. Any note entered will appear in all the files, followed by any
notes from each individual measurement."