Do I look at the FR curves?
This is for a 2-way and the good folks at minidsp suggested an active crossover. They told me their OpenDRC-DI is made for 2 channel stereo. The nano is multi-way.
I just want to use their OpenDRC-DI for parametric EQ of the woofer running full range. I have the tweeter on an LRC circuit now.
This is for a 2-way and the good folks at minidsp suggested an active crossover. They told me their OpenDRC-DI is made for 2 channel stereo. The nano is multi-way.
I just want to use their OpenDRC-DI for parametric EQ of the woofer running full range. I have the tweeter on an LRC circuit now.
Starting out with 2nd order is reasonable. You may find for eg, your design needs you to cross at one point and the woofer has an issue at another point and you need it down, so this tells you how steep to cross.
At minimum, FR, distortion v freq, polars and what is the driver meant to be doing in the design? Is it a midbass in a 2 way, or a midbass in a 4 way for example? Is the driver well behaved above passband or does it have a particularly bad break up peak/resonance? These will all factor into determining what xover F I choose as well as the order.
Even with well behaved drivers operating well within passband, I prefer steep to shallow as a general rule, but I only do active designs these days.
Even with well behaved drivers operating well within passband, I prefer steep to shallow as a general rule, but I only do active designs these days.
Yes, I see Seas free air resonance and then recommended frequency range. So I need to do some in-room measurements of each driver separately and then decide?
I can see a resonant metal driver would either need a steep x-over or else crossed over will below the breakup. But Seas nextel is pretty forgiving.
Brett, does active x-over require a separate amp for each driver?
To complicate things further, Ascend Acoustics said they use 2nd order electrical and natural roll off drivers which was 4th order acoustic on their CBM-170. So it seems no matter how you design the x-over, the driver can do something else. I have measured similar with my "hi-pass" filter but still getting output below what is supposed to be the cut-off.
This gets so complicated and with it, the chance to mess things up than fix them. Maye a cap on the tweeter like Epos is the best solution.
Maybe a topic for another thread but I am wondering where is the best place to put the driver on the front baffle. Using a single driver it can go anywhere. Except there is a horizontal cabinet brace across the middle, Would this affect reflections off the back of the driver?
I can see a resonant metal driver would either need a steep x-over or else crossed over will below the breakup. But Seas nextel is pretty forgiving.
Brett, does active x-over require a separate amp for each driver?
To complicate things further, Ascend Acoustics said they use 2nd order electrical and natural roll off drivers which was 4th order acoustic on their CBM-170. So it seems no matter how you design the x-over, the driver can do something else. I have measured similar with my "hi-pass" filter but still getting output below what is supposed to be the cut-off.
This gets so complicated and with it, the chance to mess things up than fix them. Maye a cap on the tweeter like Epos is the best solution.
Maybe a topic for another thread but I am wondering where is the best place to put the driver on the front baffle. Using a single driver it can go anywhere. Except there is a horizontal cabinet brace across the middle, Would this affect reflections off the back of the driver?
Hi, this is the nugget, only thing that matters is the acoustic slope. You just tailor the electronics so that certain acoustic slope/response is achieved with the drivers and the enclosure response included.
This is why you absolutely need to build and measure the speaker first, and then design the crossover. Purpose of the loudspeaker is to produce good acoustic output/response and purpose of the crossover is to help achieve this. Crossover should split the signal in electric domain so that it can combine back in acoustic domain and the drivers + enclosure is in the middle of all this so their effect absolutely has to be included in crossover design.
If you are new to this you can do crossover how ever you like but to really get results, in other words have no problems with crossovers, you should learn how to do two channel measurements and see some VituixCAD documentation if this is something you want get involved. VituixCAD and measurement software are free, only thing left to purhcase is a sound interface for your computer and a measurement mic with calibration data. A few hundred dollar investment + loads of time to learn all of it but once you get it the crossovers are a breeze.
Basically the box + driver configuration decides the crossover to use. You can manipulate it some for your listening environment and preference and this needs experience from your part. But very good crossover to a particular speaker construct can be made with good measurements and tweaking it in the simulation software, without knowing much at all. Now that the crossover part is no problem, making better speakers is just matter of better acoustic design, the drivers and the structure/enclosure.
Of course the work can be too much for now, so you can very well pencil in some xo on paper and be happy with it. Many speakers have been made like so and are still made. Most important thing is enjoying the music after all.
You could ask the forum if there is some errors after you come up with a crossover, there can be issues with passive crossover arrangements like too low impedance.
Have fun!🙂
Last edited:
Firstly you are after a slope after you filter which means you'll be compensating for peaks either way. What should matter is that you get breakup down regardless of how it looks to begin with, metal or not.
Secondly, an axial peak usually doesn't represent the average, it can be exaggerated.
Most computers already have a sound card.
They do have integrated soundcards, but there are two things an external interface provides: phantom power for measurement mic and another IO for loopback. Loopback is required with repeated measurements to keep track with timing. Integrated sound card or a USB mic would be fine for listening position measurements, but preserving time information with repeated measurements is key here, to have the easy xo simulations going that show the acoustic output of the loudspeaker as well as possible in home conditions.
Poor measurement data is not very useful and to get proper data it requires some cheap extra equipment and a bit of work to figure out how it all works out. Proper accurate work could be done in anechoic chamber in the nearest acoustic lab but that is too expensive for hobbyists. Few hundred as investment is not too bad in comparison and gives very good approximation of the speaker output to all directions and possibility to approximate sound in the listening position.
Anyway, a crossover can be also calculated on a paper or copied from existing designs. I've never done one on paper, I used fullrange driver and was scared of crossovers before looked into the modern way of doing things. Now it feels trivial with the software and measurement data, all mysteries lay plain in sight.
Hope it helps to give an idea how to think over crossovers, everyone should do it how ever they like. Peace!
Poor measurement data is not very useful and to get proper data it requires some cheap extra equipment and a bit of work to figure out how it all works out. Proper accurate work could be done in anechoic chamber in the nearest acoustic lab but that is too expensive for hobbyists. Few hundred as investment is not too bad in comparison and gives very good approximation of the speaker output to all directions and possibility to approximate sound in the listening position.
Anyway, a crossover can be also calculated on a paper or copied from existing designs. I've never done one on paper, I used fullrange driver and was scared of crossovers before looked into the modern way of doing things. Now it feels trivial with the software and measurement data, all mysteries lay plain in sight.
Hope it helps to give an idea how to think over crossovers, everyone should do it how ever they like. Peace!
Thanks for the help.
I am going to buy OpenDRC-DI and their cheaper mic. I can let you know more once I get the in-room driver measurements.
I am going to buy OpenDRC-DI and their cheaper mic. I can let you know more once I get the in-room driver measurements.
DSP, way to go. See post #32 for an explanation of the process:
Dayton AMTPRO-4: how low to cross when used open back?
You don't want to incorporate the failing-end of a driver's passband in your sound. You want to ensure that isn't being used. The electronics should form the XO region, not the drop-off of the driver.
Steep slope (which of course requires DSP) works really nicely. Rather falsely anthropomorphic to think drivers blend sweetly if the slope is mild. An instance of music is added from the pieces, not blended from the pieces.
B.
Dayton AMTPRO-4: how low to cross when used open back?
You don't want to incorporate the failing-end of a driver's passband in your sound. You want to ensure that isn't being used. The electronics should form the XO region, not the drop-off of the driver.
Steep slope (which of course requires DSP) works really nicely. Rather falsely anthropomorphic to think drivers blend sweetly if the slope is mild. An instance of music is added from the pieces, not blended from the pieces.
B.
Last edited:
Do they? I remember x.1 outputs back in the days but only one input with miniplug for the headset. If there are more inputs on todays computers one would need a phantom power source for the mic and some cable adapters so no external sound interface is required then.
I've been into laptops for a long time, they don't even have enough USB ports, everything minimized and external peripheral is required to do anything other than write on the forums through TV 😀
Buying less stuff is always good for the planet, bad for the economy, which is why the planet or the economy has to go. I hope for the latter and support using existing gear if at all possible. Getting of topic..
Last edited:
True, and it truly is! here is an example, I've got no idea what the slopes and frequencies are and I don't even have to, the devices are chosen to play together and they seem to. The info is readable from the graphs so I know what they ended up but this is what the system wanted, not what I wanted or think they should be.
Took maybe half an hour to tweak the xo but of course the measurements took maybe half a day with the setup and all. The point is not to show of, I don't know if my measurements are error free for example. The point is the crossovers are no mystery, a breeze, no problemo, a mere task to tune but not an obstacle. It takes another half hour to tweak for other slopes or "styles" of crossovers if one wants to compare how they sound like. Only obstacle is to get the acoustic design work together, and it is seen in the graphs if it does or not and it can be reasoned what needs to be changed if one wishes to improve on something.
Am I enthusiastic? yes I am but it is only because this is something very powerful for any speaker designer, especially us hobbyists, and is almost free (the mic and possibly an audio interface). And sounds better than ever, but I don't have anything expensive to compare to so absolute performance could still be poop. I just don't count it on the crossover anymore, it is the drivers or wrong system for the application if the sound is not nice. The knowledge and attention should be put on the stuff affecting in the acoustic domain and it is now possible since more brain power is freed from the crossover stuff since it is trivial with the free software tools.
Anyway, have fun!🙂
Took maybe half an hour to tweak the xo but of course the measurements took maybe half a day with the setup and all. The point is not to show of, I don't know if my measurements are error free for example. The point is the crossovers are no mystery, a breeze, no problemo, a mere task to tune but not an obstacle. It takes another half hour to tweak for other slopes or "styles" of crossovers if one wants to compare how they sound like. Only obstacle is to get the acoustic design work together, and it is seen in the graphs if it does or not and it can be reasoned what needs to be changed if one wishes to improve on something.
Am I enthusiastic? yes I am but it is only because this is something very powerful for any speaker designer, especially us hobbyists, and is almost free (the mic and possibly an audio interface). And sounds better than ever, but I don't have anything expensive to compare to so absolute performance could still be poop. I just don't count it on the crossover anymore, it is the drivers or wrong system for the application if the sound is not nice. The knowledge and attention should be put on the stuff affecting in the acoustic domain and it is now possible since more brain power is freed from the crossover stuff since it is trivial with the free software tools.
Anyway, have fun!🙂
Attachments
Last edited:
Sound card?
I thought it was all included with the minidsp and microphone. So you say I now need to buy something else?
I thought I just run the mic into minidsp and then upload the measurements to my computer.
I downloaded VituixCAD2_setup but no luck download REW:
2a) I can not download REW (Windows 64-bit installer (42.1MB, includes private 64-bit Java 8 runtime) for windows XP.
- ERROR [windowsx64_5_20_2.exe is not a valid win 32 application]
2b) Windows 32-bit installer (42.6MB, includes private 32-bit Java 8 runtime) and also no good
ERROR [the procedure entry point RegGetValueA could not be located in the dynamic link library]
I thought it was all included with the minidsp and microphone. So you say I now need to buy something else?
I thought I just run the mic into minidsp and then upload the measurements to my computer.
I downloaded VituixCAD2_setup but no luck download REW:
2a) I can not download REW (Windows 64-bit installer (42.1MB, includes private 64-bit Java 8 runtime) for windows XP.
- ERROR [windowsx64_5_20_2.exe is not a valid win 32 application]
2b) Windows 32-bit installer (42.6MB, includes private 32-bit Java 8 runtime) and also no good
ERROR [the procedure entry point RegGetValueA could not be located in the dynamic link library]
What is a "plugin used for?
- The plugin cost an extra $10?
How do I plug the mic into my computer's USB port?
- My computer is 10m away.
What if I put the mic 1 meter from the driver for in-room measurement?
- I can hang thick towel next speaker to eliminate room reflections?
Is VituxCAD for passive x-over?
- The plugin cost an extra $10?
How do I plug the mic into my computer's USB port?
- My computer is 10m away.
What if I put the mic 1 meter from the driver for in-room measurement?
- I can hang thick towel next speaker to eliminate room reflections?
Is VituxCAD for passive x-over?
Hi, see VituixCAD documentation, especially the measurement docs, they are right on top of the author page here
VituixCAD Loudspeaker simulator
USB mic can't do reliable repeated measurements because operating system can change delays at will and they vary between every measurement. It is fine for room correction, where you use one measurement at a time, but for crossover work multiple drivers are measured in multiple axis. Cable length problem is another issue 🙂
VituixCAD can do passive and active, it even has options to choose which DSP you have since they seem to implement even basic filters a bit differently... anyway, tere reall is no competitio. at least in free software, i have no idea what professionals use, viuixCAD maybe 😀
I haven't checked how to measure with REW other than RTA. Using ARTA for "spinorama" measurements sice there was extra steps in REW workflow few years back to do the filenaming or ssometging. VituixCAD rew measurement manual should reveal how to work with that software.
VituixCAD Loudspeaker simulator
USB mic can't do reliable repeated measurements because operating system can change delays at will and they vary between every measurement. It is fine for room correction, where you use one measurement at a time, but for crossover work multiple drivers are measured in multiple axis. Cable length problem is another issue 🙂
VituixCAD can do passive and active, it even has options to choose which DSP you have since they seem to implement even basic filters a bit differently... anyway, tere reall is no competitio. at least in free software, i have no idea what professionals use, viuixCAD maybe 😀
I haven't checked how to measure with REW other than RTA. Using ARTA for "spinorama" measurements sice there was extra steps in REW workflow few years back to do the filenaming or ssometging. VituixCAD rew measurement manual should reveal how to work with that software.
Thanks tmuikku . Can it do notch filters? I'm thinking that is the easier way to start.
I found an interesting article about x-over design.
Myths & Facts about Loudspeaker Crossovers: Identifying Legitimately High Fidelity Designs | Audioholics
I found an interesting article about x-over design.
Myths & Facts about Loudspeaker Crossovers: Identifying Legitimately High Fidelity Designs | Audioholics