I'm starting to design a subwoofer that is attempting to do what the Genelec W371A claims to do. I thought it was really interesting concept. It seems like a similar concept to using multiple subwoofers and DSP to get a flat bass response, except in this case the subwoofers are in the same cabinet.
Here is a preliminary drawing just to get things started. The cabinet volumes and port design are from VituixCAD. I'm a little surprised that the cabinet ended up being as small as this, but I'll take it!
Of course the design work isn't done. I want to make the cabinet a little wider so the 15" woofer fits better, I want to include a double thickness baffle, and I haven't accounted for the volume that will be taken up by the bracing that I haven't drawn yet. Or the volume the drivers themselves take up. So it'll be a little bigger than this, but atleast I know it will be a manageable size
The plan is for the 15" woofer to play from ~300hz down to ~60hz. (Genelec claims 500hz, but I don't if I need to go that high.) The 12" subwoofer is intended to play from ~200 to 20hz.
I already did some playing around with my subwoofers, and found that woofers oriented like this create peaks and nulls in different places in the 60-200hz range. Genelec says they use this effect to get a flatter bass response. That's what I'm hoping to achieve.
The 12" subwoofer will extend down to 20hz, and the left and right subwoofers will get some DSP treatment to try to flatten that range as well. This cabinet design is tuned to 17hz, but it only actually goes down to 20hz before dropping off the cliff.
Since this is a bit ambitious, I'm expecting to start with just 1 of these and do some testing. If it works out, I'll make it pretty and build a second one. I'll try to post here along the way, and I'd appreciate comments and suggestions.
Here is a preliminary drawing just to get things started. The cabinet volumes and port design are from VituixCAD. I'm a little surprised that the cabinet ended up being as small as this, but I'll take it!
Of course the design work isn't done. I want to make the cabinet a little wider so the 15" woofer fits better, I want to include a double thickness baffle, and I haven't accounted for the volume that will be taken up by the bracing that I haven't drawn yet. Or the volume the drivers themselves take up. So it'll be a little bigger than this, but atleast I know it will be a manageable size
The plan is for the 15" woofer to play from ~300hz down to ~60hz. (Genelec claims 500hz, but I don't if I need to go that high.) The 12" subwoofer is intended to play from ~200 to 20hz.
I already did some playing around with my subwoofers, and found that woofers oriented like this create peaks and nulls in different places in the 60-200hz range. Genelec says they use this effect to get a flatter bass response. That's what I'm hoping to achieve.
The 12" subwoofer will extend down to 20hz, and the left and right subwoofers will get some DSP treatment to try to flatten that range as well. This cabinet design is tuned to 17hz, but it only actually goes down to 20hz before dropping off the cliff.
Since this is a bit ambitious, I'm expecting to start with just 1 of these and do some testing. If it works out, I'll make it pretty and build a second one. I'll try to post here along the way, and I'd appreciate comments and suggestions.
You have your port drawn at 0.28" high? I'd suggest that is not a great idea, and to make your port at least an inch in height, more if possible. What is the port airspeed at tuning? In a cabinet of that size, a double baffle will waste volume for little benefit. Use bracing intelligently instead.
Also, while the drivers are technically opposed, by offsetting them in the vertical plane, you are creating a 'rocking couple', plus they will be fed different signals.
Also, while the drivers are technically opposed, by offsetting them in the vertical plane, you are creating a 'rocking couple', plus they will be fed different signals.
Genelec uses accordion edge pro sound driver.
likely doesnt go lower than 70/60 Hz
in a sealed box
Mainly used for upper response up to 500 Hz
Rear woofer is sub for " flat mode", will extend frequency down to 30 Hz
with a rear ported subwoofer
You would run a pro sound driver down to 60 Hz
then cross over the rear speaker for 60 to 30 Hz like usual subwoofer.
Otherwise the " directivity" mode
is a normal cardioid subwoofer.
The rear woofer is then used to cancel
like any other cardioid subwoofer with a DSP delay line
" Null steering" is another canceling mode
but needs to use a microphone and DSP
with subtraction / feedback
Similar art to noise cancellation for headphones
or hearing aids. This case it trys to remove
the dips created by the room and reflections.
Not create intentional dips.
Reflection dips are easy to create.
put a sub next to a wall or floor...bingo
otherwise you can use the built in mic, to measure room response
and manually adjust the filters for a flat response.
Like a DSP with filters
basically a sealed pro sound driver.
then rear ported sub woofer
with DSP for Crossover, EQ, Nulling and Delay line
likely doesnt go lower than 70/60 Hz
in a sealed box
Mainly used for upper response up to 500 Hz
Rear woofer is sub for " flat mode", will extend frequency down to 30 Hz
with a rear ported subwoofer
You would run a pro sound driver down to 60 Hz
then cross over the rear speaker for 60 to 30 Hz like usual subwoofer.
Otherwise the " directivity" mode
is a normal cardioid subwoofer.
The rear woofer is then used to cancel
like any other cardioid subwoofer with a DSP delay line
" Null steering" is another canceling mode
but needs to use a microphone and DSP
with subtraction / feedback
Similar art to noise cancellation for headphones
or hearing aids. This case it trys to remove
the dips created by the room and reflections.
Not create intentional dips.
Reflection dips are easy to create.
put a sub next to a wall or floor...bingo
otherwise you can use the built in mic, to measure room response
and manually adjust the filters for a flat response.
Like a DSP with filters
basically a sealed pro sound driver.
then rear ported sub woofer
with DSP for Crossover, EQ, Nulling and Delay line
I was using a calculation for a port area of ~25cm^2/ 4 in^2, resulting in a velocity of ~10m/s at 17hz. 7.8m/s@20hz. (Sorry for mixed units. VituixCAD seems to only do metric units.)
I like how the slot integrates into the design. But I don't know how or if a narrow slot port changes the calculation. I looked into it briefly, but I don't find anything yet. I mean, right now I am looking at the AudioEngine speakers on my desk that have a slot that is ~1/8" tall, so a narrow slot can work.
Genelec's promotional materials show a full width slot port with a large internal flare. The slot is far too big to work with the drivers I have been looking at. So right now, I'm trying to keep close to that design and adapt as needed.
I'm not worried about the speakers wobbling. I'm imitating a speaker cabinet that costs ~$9k each, and Genelec isn't going to overlook something like that! (The entire system is ~$25k all in.)
The reason this shouldn't be a problem is because the design uses 2 completely different drivers. The top, front facing driver (sealed cabinet) only plays down to ~60hz, so it won't do much rocking. The bottom, rear facing driver is the high excursion subwoofer. Just the weight of the two drivers should hold the thing down! Genelec's cabinet is 43" tall, and mine is 35". So I don't think that will be a problem.
But to quote a youtuber I follow, "You don't know until you know!"
I have been looking for a pro woofer that matches what Genelec uses. But then I balked because those woofers typically have a much higher sensitivity. I told myself I'm using DSP so sensitivity doesn't really matter, but I really don't know anything about pro woofers.
The 15" dayton woofer I used here is flat from 500 to 60hz in a sealed box. So I figured it was practically the same as a pro woofer?
I expect to measure and manually adjust filters, delay, etc. It will be a project!
To the question of port dimensions. I searched for information on slot ports related to speaker builds, but I'm not finding anything more than rules of thumb getting passed down.
So I turned to engineering sources. It turns out there is a "Hydraulic Diameter", which is D(h)= 4(volume)/(wetted area)
Video:
Air is considered a fluid, so this hydraulic diameter should directly apply to my port design. That means it is indeed far too small. I'll update when I work out the slot port dimensions that results in the correct hydraulic diameter!
So I turned to engineering sources. It turns out there is a "Hydraulic Diameter", which is D(h)= 4(volume)/(wetted area)
Video:
Air is considered a fluid, so this hydraulic diameter should directly apply to my port design. That means it is indeed far too small. I'll update when I work out the slot port dimensions that results in the correct hydraulic diameter!
General rule of thumb is around max 1:6 ratio for slot port dimensions.
but many go as high as 1:8
Slot port is basically a square or rectangle, is less efficient and has more friction than
a circular port.
They are common and ok to use. They share 3 walls
and have higher friction.
Just means when determining the length for a giving area to calculate
Fb or tune frequency. The constant for the correction factor is higher.
Since the port appears to be longer.
Usually end factor correction is around .732 to .850
sharing 3 walls and having higher friction.
Correction Factor is 2.227 for a slot port sharing 3 walls
So the correction factor will shorten the calculated length.
Far as overall port area needed. Apply power to bring the speaker up to Xmax
in the model. Then look at port velocity.
Basically how fast the air moves through a port.
A smaller area will cause higher velocity.
So your looking for enough area to keep velocity
low enough.
In general your looking for around 18 to 22 m/s
So it can be rather simple and straight forward to determine
the cross sectional area needed for a port area.
Just need 18 to 22 m/s (meters per second)
and try to make the opening to around 1:6 to 1:8 ratio
but many go as high as 1:8
Slot port is basically a square or rectangle, is less efficient and has more friction than
a circular port.
They are common and ok to use. They share 3 walls
and have higher friction.
Just means when determining the length for a giving area to calculate
Fb or tune frequency. The constant for the correction factor is higher.
Since the port appears to be longer.
Usually end factor correction is around .732 to .850
sharing 3 walls and having higher friction.
Correction Factor is 2.227 for a slot port sharing 3 walls
So the correction factor will shorten the calculated length.
Far as overall port area needed. Apply power to bring the speaker up to Xmax
in the model. Then look at port velocity.
Basically how fast the air moves through a port.
A smaller area will cause higher velocity.
So your looking for enough area to keep velocity
low enough.
In general your looking for around 18 to 22 m/s
So it can be rather simple and straight forward to determine
the cross sectional area needed for a port area.
Just need 18 to 22 m/s (meters per second)
and try to make the opening to around 1:6 to 1:8 ratio
Surprised you didn't find the late bjorno's vent 'cheat sheet'. Anyway, curious how it fares to what you found.
Using his rough cabinet width
I used a 1" 53/64 x 14.5" port or 1.8125" x 14.5" or 170 cm2 port area
seems to be under 18 m/s all the way to thermal rating.
driver does well in 95 liters tuned around 25 Hz
I used a 1" 53/64 x 14.5" port or 1.8125" x 14.5" or 170 cm2 port area
seems to be under 18 m/s all the way to thermal rating.
driver does well in 95 liters tuned around 25 Hz
I did a little math and came up with 2.32" x 14.5" port = 4" diameter. That's alot more like what Genelec showed in their drawings.
Should I want a low frequency spike at the tuning frequency? I tried these dimensions, and got a 10db peak at 26hz. But 58 liter and 17hz tuning results in a nearly flat curve down to 20hz. There is a small dip, but that seems like a non-issue for eq-ing.
My goal is flat, clean sound. I'm sure I"ll have to boost the lows with EQ some , but yeah, I was trying to avoid having major peaks.
Hmm, the pioneers limited large rectangular horn terminus transitions to a 1:1.273 ratio to acoustically approximate a square one, so where practical have done this with vent ducts. Is this what you mean?
Re slot vents, way-way back when I was involved in designing industrial heat/AC systems the ROT was pretty much 'brick walled' at 9:1 due to max acceptable air, mechanical, noise through the bends and FWIW worked well for designing folded TLs, etc., so seems reasonable to assume that your near 52:1 ratio is going to act as an acoustic choke.
I missed the "apply power to bring speaker to xmax"..... I do expect to play this at a more reasonable volume, but I see that I missed!
Power is at 220w. 108 db. Peak driver displacement is 14.4mm @24hz. Same box volume as before.
That's resulting in a port of 9.4cm diameter x 130cm long. Using the Hydraulic Diameter calculation I'm getting port dimensions of 2.09" x 14.5.
I hear everyone's concern about port ratio, but I want to try the full width slot. And it just got bigger too! It should be "just" physics. I plan to build just the vented enclosure to test. If it doesn't work, I'll report back.
To GM, my AudioEngine A2+ speakers use a 24:1 slot. They play pretty low and accurately for a tiny speaker. In contrast, the numbers I put up yesterday were 13.55:1. And optimizing the port for Xmax that I did just now brings it closer to 7:1.
Obviously I was referring to your original 0.28" slit slot. 😉
Based on the A2's ~65 Hz low end the vent is likely being used to make the alignment somewhat aperiodic, i.e. damped vent, which slit slots are popular for plus with sound power expanding exponentially (1/f) combined with modest power output a very high aspect ratio is the 'hot ticket' 😉.
Drop it down an octave or two combined with higher power handling and it's a whole new ballgame as simming shows, real world measurement proves.
All that said, being 'trained' by the pioneers, my vent designs were large and 'critically' damped using a simple impulse 'click' test to get high SQ BR, TL alignments, so never had any problems with excessive vent mach, though still plenty of vent output/driver damping.
Based on the A2's ~65 Hz low end the vent is likely being used to make the alignment somewhat aperiodic, i.e. damped vent, which slit slots are popular for plus with sound power expanding exponentially (1/f) combined with modest power output a very high aspect ratio is the 'hot ticket' 😉.
Drop it down an octave or two combined with higher power handling and it's a whole new ballgame as simming shows, real world measurement proves.
All that said, being 'trained' by the pioneers, my vent designs were large and 'critically' damped using a simple impulse 'click' test to get high SQ BR, TL alignments, so never had any problems with excessive vent mach, though still plenty of vent output/driver damping.
I'm getting tripped up on port size and length because the numbers I am getting seem a bit silly. I was trying to get numbers that made more sense. It feels like the cabinet is becoming a transmission line instead of a box with a port. I'm pretty sure a manufacturer isn't going to build a sub with a 60" port, so there has to be a better solution. (Genelec's subwoofers that use their "laminar spiral enclosure" may be an exception, but I'm not sure how long those ports are. I had a 7060B once, but never took it apart to see!)
I took a couple hours to play with different drivers, and I decided that I started with the wrong drivers. If I use higher sensitivity drivers, I save money on the drivers, I save a lot of money on amps, and I can get the same frequency response without crazy port designs. The only down side is the vented cabinet needs to be a little bigger, but in my case, increasing the size is a complete non-issue.
The driver I looked at last is the Dayton DCS385-4 15". It looks good in a 3.5 ft^3 vented cabinet, and it looks good in a 1.6 ft^3 sealed cabinet. The model says it hits Xmax at 107watts vented and 90 watts sealed, producing 105 and 107db. Right now I'm thinking I'll use 4 of them. They aren't "cheap", but they are a lot less than the other drivers I started looking at.
Reflecting on the Genelec model, they use a 12" driver in a vented enclosure, and claim down to 23hz. The more I play with different drivers, it seems that a 12" driver doesn't want to go much lower than that. I want to go down to atleast 20hz, and the cabinet is already wide enough for a 15", so why not use a 15"?
This is almost certainly at 1w input. Stick 100w in there and see what happens.
Chris
From what I remember the Genelec claimed 23 Hz was at -10 dB
so a rather unrealistic number.
So the real world -3 dB would be no different than many 12" subs
in the volume they are using.
Assume around 35 to 40 Hz
with a decent driver maybe 30 Hz
Your working with a 14.5" width for the port.
which is workable.
I modeled a 1.8 x 14.5 port and got a low tuning.
For the most part, for 12" the cabinet is wide enough
Far as the front mounted 15" The design seemed tight
and workable on paper. Real world, mounting the 15"
would be difficult.
Just make the baffle wider to 18"
And for a nice looking cabinet, go golden ratio.
18" wide x 29" tall.
add the needed depth to get the
right volumes.
You mentioned 60" port which sounds really to long.
Sounds like maybe you were trying to get
very low tunings around 15 Hz
and likely not using 2.227 end factor correction.
your sharing 3 walls the port will be significantly shorter.
Also tuning will be near driver Fs. So 20 to 25 Hz port
wont be 60" long by a long shot. more like 17 to 20"
so a rather unrealistic number.
So the real world -3 dB would be no different than many 12" subs
in the volume they are using.
Assume around 35 to 40 Hz
with a decent driver maybe 30 Hz
Your working with a 14.5" width for the port.
which is workable.
I modeled a 1.8 x 14.5 port and got a low tuning.
For the most part, for 12" the cabinet is wide enough
Far as the front mounted 15" The design seemed tight
and workable on paper. Real world, mounting the 15"
would be difficult.
Just make the baffle wider to 18"
And for a nice looking cabinet, go golden ratio.
18" wide x 29" tall.
add the needed depth to get the
right volumes.
You mentioned 60" port which sounds really to long.
Sounds like maybe you were trying to get
very low tunings around 15 Hz
and likely not using 2.227 end factor correction.
your sharing 3 walls the port will be significantly shorter.
Also tuning will be near driver Fs. So 20 to 25 Hz port
wont be 60" long by a long shot. more like 17 to 20"
Last edited:
Right, as power increased over time, driver specs went from high Vas, low Qts with an Fs = Fb that allowed large vents just baffle thickness to small Vas medium-high Qts with an Fs based on using an Fb > Fs for high power apps and otherwise an Fb whatever T/S dictated, which required ever longer vents to the point where I just calc'd an alignment based on the pioneer's Av = Sd, then built it as an inverse tapered TL (TQWT) of the same summed volume and fine tuned it if need be with a terminus restriction and/or the aforementioned 'click' test and in recent years the desire for ever smaller, yet higher power speakers has resulted in all manner of other type design compromises.
That said, with the recent price drops of high SQ, all encompassing DSP systems such as dbx or similar, the goal is a sealed cab sub of whatever dimensions one wants with the largest driver that will fit, etc., and DSP its in room response and if one doesn't have enough power handling (Xmax) to meet the needs of the app, then use multiples spread around the room ala Dr. Geddes.
Makes sense, but as mentioned the easy quick reference
for rectangle port is a 1:6 ratio and people do extend it to 1:8
So 14.5 wide
1:6 = 2.41" tall
1:8 = 1.81" tall
As mentioned in post #9
using the worst case 1:8 ratio
Port velocity didnt exceed 18 m/s
even with power up to the thermal rating.
this was with a standard BB4 alignment
95 liters 25 Hz Fb
I didn't realize the Genelec was 23hz @-10db. My 10" M-Audio BX subwoofers extend down to 28hz in room. They were advertised down to 20hz. They might hit 20hz at -20db!
The 18" you mention seems appropriate. I only went down to the 14.5" internal width because the woofers I started with didn't need as much internal volume as I was expecting. I did some tests that indicated the 34" height as about optimal for getting peaks and nulls to diverge, which I need if I want to correct them with DSP. I know taller was worse; I didn't try shorter. So I will taking feet off the subs, turn them sideways, etc to see if going shorter will change the result.
I was still using the .732 end correction factor. Using the 2.227 factor does help!
I'm skimming through Troels Gravenson's article on port calculations as I'm responding. And I'm also looking at an AVSForum thread showing the 2.227 end correction factor you mentioned for a full width slot port. I hadn't realized the type of construction of the port made that much of a difference.
Using the 2.227 end correction factor, in a 100L box, tuned to 18hz results in a port that is 10cm/ 4" diameter and 50cm/ 20" length. The driver hits it's Xmax @ 25hz and 110watts. And at the same 110watts, the max vent velocity is ~29m/s @ 18hz. The SPL is 108db. (It will never get turned up that high!)
Troels Gravenson's article demonstrates that stuffing a cabinet lowers the tuning frequency. So is it advisable to increase the tuning frequency in the design knowing it will end up being lower? Or is it better to do it the old fashioned way of building to the calculation and then adjusting based on measurements?