Are multiple ports better ?
- Thread starter goldyrathore
- Start date
Hi,
I would imagine one round port being the best alternative, it gives you the largest area/circumference. Port walls always add friction to the flow, and that can't be good?
Of course if you choose multiple ports you can place their ends evenly on the box. On a large box, location of port ends might have significant effect when placed in a room.
I would imagine one round port being the best alternative, it gives you the largest area/circumference. Port walls always add friction to the flow, and that can't be good?
Of course if you choose multiple ports you can place their ends evenly on the box. On a large box, location of port ends might have significant effect when placed in a room.
my limited expereince in ported boxes tells me that a single well placed port is best, and I tend to cut the ends of my ports with angle cuts, or plate them out for better loading of the port as a whole. Ie, if the port is 3" in diameter, then cut a square plate out of MDF or the like, that is 4-5" across on it's sides, and place that flush with the end of the port, inside the box. Equalizes the air mass loading when the air is in motion, or during transition to motion.
The plate on the end of the port on the inside kills diode like chuffing/farting of the port. It tends to increase fidelity in ported systems by a very noticable amount. Sometimes this effect can be so severe it blows a woofer via this diode like issue of the port air mass. The plate equalizes the air mass loading on either side of the port : CRITICAL.
Just my bit of expereince designing sub boxes for cars when I was doing such things for a local store. back when I thought ported boxes were ok.
The plate on the end of the port on the inside kills diode like chuffing/farting of the port. It tends to increase fidelity in ported systems by a very noticable amount. Sometimes this effect can be so severe it blows a woofer via this diode like issue of the port air mass. The plate equalizes the air mass loading on either side of the port : CRITICAL.
Just my bit of expereince designing sub boxes for cars when I was doing such things for a local store. back when I thought ported boxes were ok.
One port is easier but you get more flexibility with your sizes if you use two or more.
I use PVC pipe for ports and the common sizes in Australia are 50mm - way too small, 86mm (internal diameter), 102mm and 152mm.
To keep the airspeed down, the 100mm is too small and the 152mm is too large - meaning a longer port and problems with "pipe mode" resonances.
I have generally found that a pair of 86mm ports are just right!
regards
Collo
I use PVC pipe for ports and the common sizes in Australia are 50mm - way too small, 86mm (internal diameter), 102mm and 152mm.
To keep the airspeed down, the 100mm is too small and the 152mm is too large - meaning a longer port and problems with "pipe mode" resonances.
I have generally found that a pair of 86mm ports are just right!
regards
Collo
Goldy, the short answer is NO there is no real sonic difference if the multiple ports are of a reasonable size.
The critical thing to watch is vent velocity, it should not exceed 34m/s where the driver is moving at the max required excursion.
However, all vents should be flared as this has a MASSIVE impact on turbulence, which is the one thing you want to avoid.
KBK,
I have no idea what you are talking about regarding a diode!
As Collo indicated, multiple ports are more a matter of pragmatics - different ways to get a decent vent area. You may find if a 100mm vent is not big enough and a 150mm vent is too big that you can find a pair of 90mm vents are just right.
*off topic* Collo, I think you should call your sub something along the lines of "lost in space" it reminds me of that robot, the vents like its arms, remember that old show with the robot which always rolls around saying "warning! warning!"
The critical thing to watch is vent velocity, it should not exceed 34m/s where the driver is moving at the max required excursion.
However, all vents should be flared as this has a MASSIVE impact on turbulence, which is the one thing you want to avoid.
KBK,
I have no idea what you are talking about regarding a diode!
As Collo indicated, multiple ports are more a matter of pragmatics - different ways to get a decent vent area. You may find if a 100mm vent is not big enough and a 150mm vent is too big that you can find a pair of 90mm vents are just right.
*off topic* Collo, I think you should call your sub something along the lines of "lost in space" it reminds me of that robot, the vents like its arms, remember that old show with the robot which always rolls around saying "warning! warning!"
How about several ports with different CSA and same lenght same lenght for easier calculation).
For Example one large and two smaller tubes. Plugging a smaller port changes the tuning frequency about 2-4 Hz. If you use two or three ports, plugging one of them easily changes the tuning frequency too much.
Any comments on this idea?
For Example one large and two smaller tubes. Plugging a smaller port changes the tuning frequency about 2-4 Hz. If you use two or three ports, plugging one of them easily changes the tuning frequency too much.
Any comments on this idea?
Calculation is not difficult (i guess...). instead of using 5 small tubes you could use 1 large and 2 small tubes. Isn't that the same thing?
All tubes should be equally long (for easier calculation) so that only total CSA changes when you plug the tubes. Calculation is easy with WinISD, just replace one large tube with eg three smaller tubes. That should give reasonably accurate results.
I would like my speakers or subwoofers as flexible as possible. Many people plug their speakers to make them act like sealed boxes. Why not make the speakers more adjustable.
All tubes should be equally long (for easier calculation) so that only total CSA changes when you plug the tubes. Calculation is easy with WinISD, just replace one large tube with eg three smaller tubes. That should give reasonably accurate results.
I would like my speakers or subwoofers as flexible as possible. Many people plug their speakers to make them act like sealed boxes. Why not make the speakers more adjustable.
Thanks
DIA:
Would ports of different CSA be used at the same time or would the different ones be plugged?
When using ports with different CSA simultaneously, they all should be tuned the same. So a big port would be longer and a small one shorter. Otherwise you would end up with multiple tuning frequencies, and WinISD can't calculate this.
If the idea was to use only one type of ports of equal lenght at a time, it would work. But in that case it would probably be easier to use multiple smaller ports of same size and lenght...
Would ports of different CSA be used at the same time or would the different ones be plugged?
When using ports with different CSA simultaneously, they all should be tuned the same. So a big port would be longer and a small one shorter. Otherwise you would end up with multiple tuning frequencies, and WinISD can't calculate this.
If the idea was to use only one type of ports of equal lenght at a time, it would work. But in that case it would probably be easier to use multiple smaller ports of same size and lenght...
Goodguy, that's what I'm worried about. But If I simulate a single tube with eg CSA=60 cm2 and Lenght=50 cm, I get almost the same tuning frequency as using three CSA=20 cm2, Lenght=50 cm tubes. I think.
Aren't the total CSA and the lenght of the port/ports the fundamental parameters that determine the tuning frequency?
Edit:
Just to make myself clear, I attached a drawing. I believe all of these configurations have the same tuning frequency. But as I said, this is just a gut feeling, not a fact
Aren't the total CSA and the lenght of the port/ports the fundamental parameters that determine the tuning frequency?
Edit:
Just to make myself clear, I attached a drawing. I believe all of these configurations have the same tuning frequency. But as I said, this is just a gut feeling, not a fact
Attachments
DIA:
Yes.
I was only pointing out that you can't use ports with same lenght and different CSA at the same time, because their individual tuning frequencies would be different.
Of course I could be wrong, but I can imagine the calculations for mixed port systems being challenging(you can forget about WinISD).
Yes.
I was only pointing out that you can't use ports with same lenght and different CSA at the same time, because their individual tuning frequencies would be different.
Of course I could be wrong, but I can imagine the calculations for mixed port systems being challenging(you can forget about WinISD).
There is nothing wrong with multiple ports for different tuning frequencies. I had a go here:
http://www.users.bigpond.com/bcolliso/blast-one.htm
You can use different sizes, but you would need to work out the total load presented by all the ports - this would be analogous to solving a parallel arrangement of resistors of different values.
Far easier is to choose your minimum size and then model in WinISD multiple ports of the same length. You are probably only going to come up with 2 combinations of frquencies that are practical.
Remember when modelling, to factor in the change in box volume as ports are "taken off line" and their volume effectively returned to the box.
I think either SVS or REL use the same system, but they also have a switch to change the equalisation to suit the port arrangement you are using.
I agree with Paul that you really must use port flares, however I have found the 34 m/sec he quotes to be too high. Where possible I aim for 10 m/sec. It does mean larger ports, but there is no port noise at all - no shooting ping-pong balls out of my subs!
And yes Paul, the sub does look like the "lost in space" robot. In fact one of my friends calls it the "robot". Thanks for your comments
http://www.users.bigpond.com/bcolliso/blast-one.htm
You can use different sizes, but you would need to work out the total load presented by all the ports - this would be analogous to solving a parallel arrangement of resistors of different values.
Far easier is to choose your minimum size and then model in WinISD multiple ports of the same length. You are probably only going to come up with 2 combinations of frquencies that are practical.
Remember when modelling, to factor in the change in box volume as ports are "taken off line" and their volume effectively returned to the box.
I think either SVS or REL use the same system, but they also have a switch to change the equalisation to suit the port arrangement you are using.
I agree with Paul that you really must use port flares, however I have found the 34 m/sec he quotes to be too high. Where possible I aim for 10 m/sec. It does mean larger ports, but there is no port noise at all - no shooting ping-pong balls out of my subs!
And yes Paul, the sub does look like the "lost in space" robot. In fact one of my friends calls it the "robot". Thanks for your comments
Diar's idea is really quite simple to calculate. A vent is tuned simply based on the vent diameter, length and box volume. Hence with multiple ports of the same length, merely calculate for a single port with cross sectional area equal to all of them put together. If not taken to extremes with tiny ports, this will work fine. To calculate the tuning with any of them blocked, simply calculate for a single port of a smaller CSA. No need for gut feelings here, this is simple physics.
Collo, I'm puzzled. Designing for 10m/s sounds like massive overkill (which I'm normally in favour of, but in this case it turns into a huge pest). Under what circumstances do you find this necessary? Are you designing to avoid any audible turbulence with test tones played at the vent air velocity peak where there is no background noise to mask?
I have had my AV12 subs in a 70L vented test box with a 4" 1m long vent tuned at 18 Hz. In actual use I've had no turbulence problems either for music or home theatre and this is a 12" driver which has almost 2" p-p xmax. There is a rumble filter which helps, however the vent was not flared at all. For my vent flaring experiments I used a 15 Hz sine wave with the rumble filter off. I chose this frequency as it simulates as the peak in vent velocity, the "worst case scenario." At xmax the vent velocity is 42m/s which is quite high. Without a flare the chuffing seems quite annoying when there is no other sound to mask it. I then used 3 vent flares 18/36/72 mm in radius and found each doubling of radius had a significant improvement. The chuffing noise would have been perhaps 50db, however the output of the sub within it's bandwidth is over 110 db! Is that enough masking effect for you? In very rare moments in movies you might have a very narrow bandwidth LFE transient during a moment where this is not much fulllrange output to mask the chuffing. This is where problems are possible.
however ...
if I turn on the rumble filter (3rd order HPF @ 20 Hz), the velocity now peaks at 20m/s @ 20 Hz. This is now well within the acceptable range and with a generous flare of 72mm radius, there should be no problems at all.
Collo, I'm puzzled. Designing for 10m/s sounds like massive overkill (which I'm normally in favour of, but in this case it turns into a huge pest). Under what circumstances do you find this necessary? Are you designing to avoid any audible turbulence with test tones played at the vent air velocity peak where there is no background noise to mask?
I have had my AV12 subs in a 70L vented test box with a 4" 1m long vent tuned at 18 Hz. In actual use I've had no turbulence problems either for music or home theatre and this is a 12" driver which has almost 2" p-p xmax. There is a rumble filter which helps, however the vent was not flared at all. For my vent flaring experiments I used a 15 Hz sine wave with the rumble filter off. I chose this frequency as it simulates as the peak in vent velocity, the "worst case scenario." At xmax the vent velocity is 42m/s which is quite high. Without a flare the chuffing seems quite annoying when there is no other sound to mask it. I then used 3 vent flares 18/36/72 mm in radius and found each doubling of radius had a significant improvement. The chuffing noise would have been perhaps 50db, however the output of the sub within it's bandwidth is over 110 db! Is that enough masking effect for you? In very rare moments in movies you might have a very narrow bandwidth LFE transient during a moment where this is not much fulllrange output to mask the chuffing. This is where problems are possible.
however ...
if I turn on the rumble filter (3rd order HPF @ 20 Hz), the velocity now peaks at 20m/s @ 20 Hz. This is now well within the acceptable range and with a generous flare of 72mm radius, there should be no problems at all.
Excellent. From now on I will definitely use several (3 ?) ports with 2 or three different diameters in my projects.
Collo's project was a good example how plugging a port changes tuning frequency very much when using only two similar ports. Usually there is also a problem with air velocity in ports when plugging one. Collo however uses huge ports
I'll try to keep the air velocity under 17 m/s (5% of speed of sound). That should be a reasonable compromise between chuffing noise and length and volume of the ports.
Collo's project was a good example how plugging a port changes tuning frequency very much when using only two similar ports. Usually there is also a problem with air velocity in ports when plugging one. Collo however uses huge ports
I'll try to keep the air velocity under 17 m/s (5% of speed of sound). That should be a reasonable compromise between chuffing noise and length and volume of the ports.
Hi Paul,
When I'm testing a sub for maximum output, I'm wary of using pure tones as I think an indiscretion could easily burn out a driver. I prefer to use a DVD with a blast of short duration.
I play the section repeatedly, slowly bringing up the volume. What I listen for is the quite noticable change in sound when you run out of excursion. In other words I'm taking the driver right to the limit for a short time. If there is any chuffing at all from the ports under this test, I consider them to be too small.
My decision to target 10 m/sec has its genesis in experiments with the "Sidewinder" sub I built a while back. 70 litre 10inch 125watts 26hz tune.
When I first designed it, I was using an earlier version of winisd that recommended 34m/sec, so I chose 2 * 50 mm ports for about 27 m/sec
This chuffed like an asthmatic smoking a cigar!
The current version of winisd recommends 17m/sec, so I modified the sub win 2 * 68mm ports for a peak of 13.5 m/sec
I could still hear them, so I went up to 3 * 68mm - peak 9 m/sec. Finally peaceful bliss. Thus my decision to aim for around 10.
One of the problems I face is that my router will only do a 10mm rollover for my flares. I had considered this sufficient, but from what you say this is way too small.
I followed your posts on dimples back in december with interest and always wondered how they turned out.
When you were doing that testing and developing your flares, were you able to quantify the benefit of larger flares.
A table such as the following would be a tremendous asset for the DIY community
A flare radius 10mm requires a peak speed of 10 m/sec
.......................x mm..........................17..........
.......................y mm..........................34..........
.......................z mm..........................40..........
It would require choosing an acceptable definition for "quiet" and a standardised testing method
Obviously a lot of work machining different flares to suit a range of port diameters.
Has anyone seen any theoretical work that might fill in such a table. The big speaker manufacturers must have been down this road already!
regards
Collo
When I'm testing a sub for maximum output, I'm wary of using pure tones as I think an indiscretion could easily burn out a driver. I prefer to use a DVD with a blast of short duration.
I play the section repeatedly, slowly bringing up the volume. What I listen for is the quite noticable change in sound when you run out of excursion. In other words I'm taking the driver right to the limit for a short time. If there is any chuffing at all from the ports under this test, I consider them to be too small.
My decision to target 10 m/sec has its genesis in experiments with the "Sidewinder" sub I built a while back. 70 litre 10inch 125watts 26hz tune.
When I first designed it, I was using an earlier version of winisd that recommended 34m/sec, so I chose 2 * 50 mm ports for about 27 m/sec
This chuffed like an asthmatic smoking a cigar!
The current version of winisd recommends 17m/sec, so I modified the sub win 2 * 68mm ports for a peak of 13.5 m/sec
I could still hear them, so I went up to 3 * 68mm - peak 9 m/sec. Finally peaceful bliss. Thus my decision to aim for around 10.
One of the problems I face is that my router will only do a 10mm rollover for my flares. I had considered this sufficient, but from what you say this is way too small.
I followed your posts on dimples back in december with interest and always wondered how they turned out.
When you were doing that testing and developing your flares, were you able to quantify the benefit of larger flares.
A table such as the following would be a tremendous asset for the DIY community
A flare radius 10mm requires a peak speed of 10 m/sec
.......................x mm..........................17..........
.......................y mm..........................34..........
.......................z mm..........................40..........
It would require choosing an acceptable definition for "quiet" and a standardised testing method
Obviously a lot of work machining different flares to suit a range of port diameters.
Has anyone seen any theoretical work that might fill in such a table. The big speaker manufacturers must have been down this road already!
regards
Collo
lol ... I'm having a hard time grasping how it could be anything but simple and easy!
Say you have a 4" vent which is 1m long. It is the same thing as having 4 x 2" vents of the same length. You could use 3 vents which add up to the total CSA of the single vent, and the tuning will be the same if the length is the same. Tuning is determined by vent area, length and the box volume. Simple!
Collo, I can just see your dream system now. A Tumult with a spare room dedicated to the custom made concrete vent! *chuckles*
I like your idea, I'm all for sharing ideas and results with the diy community. I feel that vent design is something that could be much better done overall. I also feel that passive radiators are a waste of money, and if vented designs were done better, we'd be spending more on better drivers, and work on really getting our vents right. To me this is a big part of diy.
You can make vents by heat bending PVC onto a template and get a flare that is bigger than your router bit.
The table idea is a good one, however it would take a LOT of work to actually determine all that properly. What I may in fact do is a bit simpler.
Use only a 4" vent with:
1. no flare
2. 18mm radius
3. 36mm
4. 72mm
5. a dimpled version
I could actually measure the SPL level of the chuffing @ 1m and at my listening position. I could then experiment with the required level of output to actually mask the chuffing, and find out where the threshold is - ie at what point it is just noticeable.
I might be able to say things like:
* doubling the flare radius decreases SPL of chuffing by 8db
* this has the same effect as half the vent velocity, hence a flare of x mm radius means the vent can be x % smaller in diameter ...
Paul,
I like the idea of heat bending PVC over a template of some sort. I'll have to try and come up with a way of making the template without a lathe. I'm going to have a go and will post some photos if I have any luck.
Your idea of using the 4 inch ports (which I assume you still have from your previous tests) sounds pretty good.
There would need to be a measure for acceptable chuffing. I would suggest listening quite near the speaker to be able to say "yes it is chuffing" or "no it's" not. I think this is more accurate for testing than doing it from the seating position. Later a figure could be worked out for how much extra could be tolerated from the seating position
It might be best to build a test box which is deliberately under-ported so that pure tones can be used without risking burning out a driver. A larger box / shorter port does the trick.
Winisd could be used to estimate port velocity from the SPL measured.
This approach might produce something like...
*** EXAMPLE ONLY **
Test environment - a 4 inch port tested at 25hz
Chuffing becomes noticable at:
No flare - 80dB output - ....estimated port velocity 10 m/sec
18mm flare - 85 db output - ....estimated port velocity 15 m/sec
36mm flare - 95 db output - ....estimated port velocity 20 m/sec
etc
...*** EXAMPLE ONLY **
I like the idea of heat bending PVC over a template of some sort. I'll have to try and come up with a way of making the template without a lathe. I'm going to have a go and will post some photos if I have any luck.
Your idea of using the 4 inch ports (which I assume you still have from your previous tests) sounds pretty good.
There would need to be a measure for acceptable chuffing. I would suggest listening quite near the speaker to be able to say "yes it is chuffing" or "no it's" not. I think this is more accurate for testing than doing it from the seating position. Later a figure could be worked out for how much extra could be tolerated from the seating position
It might be best to build a test box which is deliberately under-ported so that pure tones can be used without risking burning out a driver. A larger box / shorter port does the trick.
Winisd could be used to estimate port velocity from the SPL measured.
This approach might produce something like...
*** EXAMPLE ONLY **
Test environment - a 4 inch port tested at 25hz
Chuffing becomes noticable at:
No flare - 80dB output - ....estimated port velocity 10 m/sec
18mm flare - 85 db output - ....estimated port velocity 15 m/sec
36mm flare - 95 db output - ....estimated port velocity 20 m/sec
etc
...*** EXAMPLE ONLY **