I'm working on a project with multiple smaller drivers in various baffle arrangements and wanted to know if there was a reliable way to sim or calculate total system voltage sensitivity and in which areas of FR any gain would be seen.
As an example, if I were to use 6 x Peerless 830870 bass-mids (3 paralleled groups of 2 series connected drivers in each group) in a tight circular grouping around a centrally positioned tweeter. Driver specs attached below >>>
- In which FR range would I be seeing sensitivity gains?
- At which frequency will the drivers acoustically uncouple from each other?
I understand the CTC spacing of all the drivers matters most here, but also the geometry of how they're arranged ie. if we're dealing with an oval arrangement, not just a simple circle or line array configuration.
- Which software is best at figuring this out? I understand Basta does this to some extent, but it won't run on my PC for some odd conflict reason.
As an example, if I were to use 6 x Peerless 830870 bass-mids (3 paralleled groups of 2 series connected drivers in each group) in a tight circular grouping around a centrally positioned tweeter. Driver specs attached below >>>
- In which FR range would I be seeing sensitivity gains?
- At which frequency will the drivers acoustically uncouple from each other?
I understand the CTC spacing of all the drivers matters most here, but also the geometry of how they're arranged ie. if we're dealing with an oval arrangement, not just a simple circle or line array configuration.
- Which software is best at figuring this out? I understand Basta does this to some extent, but it won't run on my PC for some odd conflict reason.
Attachments
My first thought in a similar way you might use Basta, could you take the the individual driver responses and impedance and input them into VituixCad and model response for baffle size and position and save as driver 1, driver 2 etc. . Possibly somewhat time consuming I know.
Then just add the drivers to a design as driver 1, 2,3,4 and so on?
I haven't tried this particular idea, but I think it should work in the same way as doing this with a typical three or four way with a bass/mid/tweeter /super tweet setup, so I may be pointing you in the wrong direction.
Then just add the drivers to a design as driver 1, 2,3,4 and so on?
I haven't tried this particular idea, but I think it should work in the same way as doing this with a typical three or four way with a bass/mid/tweeter /super tweet setup, so I may be pointing you in the wrong direction.
It might be very helpful if HifiJim reminded us about doing this and the correct use of the Vituixcad positional measurement system. The answer is in one of his earlier design threads from maybe a year or two years back.
Need to get this understood so that the phase aspects are correct.
Need to get this understood so that the phase aspects are correct.
Yes. See one way to get the total SPL:
https://www.diyaudio.com/community/...0-rebalance-system-power.394278/#post-7229463
Determining when interference between drivers becomes a problem is more of a diffraction type modeling problem. I suggest you use the Edge (by Tolvan Data) to figure that out.
They don't per se. If you put them at opposite sides of your room, you'll still see the gains at all frequencies if you measure from the middle of the room. You'll have to analyse the interference patterns.
Thanks guys. I heard Vituix is rather powerful SW. I don't do many SW sims aside for crossover design - hence the questions. I'm new to Edge. Looks to be very useful for my needs.
@AllenB I understand the coupling concept. My question is where the transition point is where the WL of frequency reproduced is smaller than the distance between drivers. Some say its half WL and others say its closer to 1/3rd. I get the concept that lower frequencies couple more effectively in a room due very long WLs (longer than the room dimensions). I realize at HF the coupling gain goes away and you only end up with the driver's base efficiency alone despite being in an array.
@AllenB I understand the coupling concept. My question is where the transition point is where the WL of frequency reproduced is smaller than the distance between drivers. Some say its half WL and others say its closer to 1/3rd. I get the concept that lower frequencies couple more effectively in a room due very long WLs (longer than the room dimensions). I realize at HF the coupling gain goes away and you only end up with the driver's base efficiency alone despite being in an array.
Hi,
VituixCAD is great tool to make experiments, and get answers like this.
Add two ideal sources and delay the other somehow, just move it further. You could use the diffraction tool to put transducers as you like, for example x drivers in oval shape, and get "spinorama" out of it.
Ok here is about interference. When two sources have their phase less than 90deg apart, 1/4wl, they sum constructively so that the sum is louder than any of the two sources alone. At 1/2wl, with 180deg phase difference, there is destructive interference as they are out of phase and in ideal case could be inaudible. Between, about at 120deg, or 1/3wl, SPL goes below either, so destructive interference it is.
Here, two sources, other one moved back 341mm which accounts for one wavelength at 1kHz. First null is at 500Hz, as 341mm is half wavelength of that. 341mm would be third of about 335Hz so the blackline crosses the blue one. Below about 136,4cm, or 251Hz, 1/4wl, sum of the two is about as good as it gets.
So, the other sound source is behind the other for the observation point we have the graph this is worst case, longest path length differrence there is in this system. If you imagine now having the sources at same physical locations but changed observation angle path lenght drifference from each to you would be less than this. Hence we can assume it's roughly a point source as long as transducers are about 1/4wl apart because it's all constructive interference toward any observation point. And as you see it's not o strict, bit above or below is fine, it's just an approximation.
You can have pretty good idea of response with diffraction tool, real time adjustment. Great way to study things, just use imagination to figure out how to make a test for what is it you want to get more info about
It's all about wavelength, spend some time with vituixCAD or any other software you like that has realtime adjustment (mousewheel) and you'll get a ton of intuition how sound interacts with objects, transducers, all interacting with each other.
Have fun!
VituixCAD is great tool to make experiments, and get answers like this.
Add two ideal sources and delay the other somehow, just move it further. You could use the diffraction tool to put transducers as you like, for example x drivers in oval shape, and get "spinorama" out of it.
Ok here is about interference. When two sources have their phase less than 90deg apart, 1/4wl, they sum constructively so that the sum is louder than any of the two sources alone. At 1/2wl, with 180deg phase difference, there is destructive interference as they are out of phase and in ideal case could be inaudible. Between, about at 120deg, or 1/3wl, SPL goes below either, so destructive interference it is.
Here, two sources, other one moved back 341mm which accounts for one wavelength at 1kHz. First null is at 500Hz, as 341mm is half wavelength of that. 341mm would be third of about 335Hz so the blackline crosses the blue one. Below about 136,4cm, or 251Hz, 1/4wl, sum of the two is about as good as it gets.
So, the other sound source is behind the other for the observation point we have the graph this is worst case, longest path length differrence there is in this system. If you imagine now having the sources at same physical locations but changed observation angle path lenght drifference from each to you would be less than this. Hence we can assume it's roughly a point source as long as transducers are about 1/4wl apart because it's all constructive interference toward any observation point. And as you see it's not o strict, bit above or below is fine, it's just an approximation.
You can have pretty good idea of response with diffraction tool, real time adjustment. Great way to study things, just use imagination to figure out how to make a test for what is it you want to get more info about
It's all about wavelength, spend some time with vituixCAD or any other software you like that has realtime adjustment (mousewheel) and you'll get a ton of intuition how sound interacts with objects, transducers, all interacting with each other.
Have fun!
Last edited:
Use dual channels measurement with mic on axis on each driver and then input the coordinates x,y,z in the main program.
See also here https://kimmosaunisto.net/Software/VituixCAD/VituixCAD_Measurement_REW.pdf
I didn't realize you could do all that in Vituix. Pretty impressive. Looks like I have some homework.
In trying to figure this out, I started cutting out a test baffle to see how I close i can get everything with a WGed tweeter like the seas T35C002. I don't know if its better to use the tweeter with or without WG as a compromise to CTC spacing. It will do 94 - 95 dB/2.83V on its own, but having it set further back with WG probably is better for acoustic center alignment with mids
In trying to figure this out, I started cutting out a test baffle to see how I close i can get everything with a WGed tweeter like the seas T35C002. I don't know if its better to use the tweeter with or without WG as a compromise to CTC spacing. It will do 94 - 95 dB/2.83V on its own, but having it set further back with WG probably is better for acoustic center alignment with mids
I would imagine a small 3/4" neo magnet tweeter would work well
for Center to Center spacing.
Neo would reduce the magnet size and reduce or remove
the common large bezels with larger tweeters.
Moving slightly down to 3" drivers and the small tweeter
would allow a tighter mounting pattern.
Which is what would be working against you.
Overall center to center spacing.
I tend to draw things to scale in Google sketchup.
To see how things mount. Then also can get exact
driver center mounting positions.
Which can then be placed in exact position in Virtuix
cad.
Listening position could be considered the center
or on the tweeter In sim.
If listening/microphone position is kept same for simulation export.
Instead of each driver center.
No need to do mind numbing calculations to get X/Y coordinates
correct.
Dont move the microphone and then phase will be correct.
I sim each driver and phase/ individually
for Center to Center spacing.
Neo would reduce the magnet size and reduce or remove
the common large bezels with larger tweeters.
Moving slightly down to 3" drivers and the small tweeter
would allow a tighter mounting pattern.
Which is what would be working against you.
Overall center to center spacing.
I tend to draw things to scale in Google sketchup.
To see how things mount. Then also can get exact
driver center mounting positions.
Which can then be placed in exact position in Virtuix
cad.
Listening position could be considered the center
or on the tweeter In sim.
If listening/microphone position is kept same for simulation export.
Instead of each driver center.
No need to do mind numbing calculations to get X/Y coordinates
correct.
Dont move the microphone and then phase will be correct.
I sim each driver and phase/ individually
Still curious.
But having modeled multiple size mids/tweets
for basic D' Appolito arrangements.
Its the wonderful thing about models to
be able to observe the behavior before
building.
The center to center spacing really plays
into getting anything that is successful.
And even when pushing physical boundaries
of getting center to center spacing very tight.
By using smaller and smaller drivers.
I abandoned the arrangement.
So I see this proposed concept behaving
similar but multiplied in every direction.
Not to sound negative
it is actually extremely positive because
of what models can show you.
And rather fun if you enjoy the process.
But also makes relative basic audio design
and theory extremely visual and more understood.
Still curious. likely wont trace and get data
for proposed drivers. Just use drivers
in my library to observe the behavior.
But having modeled multiple size mids/tweets
for basic D' Appolito arrangements.
Its the wonderful thing about models to
be able to observe the behavior before
building.
The center to center spacing really plays
into getting anything that is successful.
And even when pushing physical boundaries
of getting center to center spacing very tight.
By using smaller and smaller drivers.
I abandoned the arrangement.
So I see this proposed concept behaving
similar but multiplied in every direction.
Not to sound negative
it is actually extremely positive because
of what models can show you.
And rather fun if you enjoy the process.
But also makes relative basic audio design
and theory extremely visual and more understood.
Still curious. likely wont trace and get data
for proposed drivers. Just use drivers
in my library to observe the behavior.
profiguy,
No one answered your question on the sensitivity gain before we got off onto waveguides and CTC considerations. From my line array studies the arrayed sensitivity increase depends on the number of drivers (in your case 6) and how they are connected (your case 3 parallel connected groups of 2 series connected 8 ohms drivers). The impedance for the 6 drivers connected in the array is 5.33 ohms. Thus the total arrayed sensitivity increase is 7.78 dB (which is 10log6) plus 10log8/5.33 = 1.69 dB for the impedance change. This totals to 7.78 + 1.69 = 9.47 dB. This is as good as it gets for direct radiating drivers with sound fields overlapping.
If you separate the drivers so that their sound fields don't overlap the sensitivity increases goes down.
See page 17 of my NFLAWP for discussion.
https://audioroundtable.com/misc/nflawp.pdf
Jim
No one answered your question on the sensitivity gain before we got off onto waveguides and CTC considerations. From my line array studies the arrayed sensitivity increase depends on the number of drivers (in your case 6) and how they are connected (your case 3 parallel connected groups of 2 series connected 8 ohms drivers). The impedance for the 6 drivers connected in the array is 5.33 ohms. Thus the total arrayed sensitivity increase is 7.78 dB (which is 10log6) plus 10log8/5.33 = 1.69 dB for the impedance change. This totals to 7.78 + 1.69 = 9.47 dB. This is as good as it gets for direct radiating drivers with sound fields overlapping.
If you separate the drivers so that their sound fields don't overlap the sensitivity increases goes down.
See page 17 of my NFLAWP for discussion.
https://audioroundtable.com/misc/nflawp.pdf
Jim
@Jim Griffin I greatly appreciate your answer. This was the other part of info I was looking for.
Yes, I figured the actual arrangement of drivers will govern how tightly they couple. In the case of smaller fullrange drivers this can be a considerable amount of gain from acoustical efficiency alone.
The PDF link won't load for me for some reason.
My questions would be - Is this 1/3 rd WL point the -3dB uncoupling frequency you're referring to?
If not, how far down will the response be at this point?
Does this affect acoustic phase compared to a single driver?
Last edited: