What is Reynolds number?
What do you mean by trip?
and flow separation?
FWIW I'm an Architect with a fair appreciation of engineering. My dad was a senior lecturer in Mech Eng, my sister has a PhD in the same field! You did well in describing, although not having studied things like fluid mechanics, there's a few gaps to fill in with terminology.
The smoke idea is interesting, however I'm not sure it would be easy to see what is going on with an actual 3D vent. Perhaps it could be shown visually with one of those hmmmm what's it called? That tank where water flows past obstructions with a dye in the water to show flow, and you observe disturbances.
In my case, nothing beats an actual prototype.
In a bigger box I'd use a 150mm vent, but there's no way I can fit such a vent in a box I find acceptable in terms of size. I could fit in a 120mm vent. I did comparisons with sims of 100mm vs 120mm vents. Given the same vent length, and a rumble filter, both seem ok, less than 34m/s velocity. The larger vent is tuned higher but has a larger area - the two factors cancel each other out so they are equally effective. However, 100mm diam is easier to work with as the larger one becomes cramped inside.
Hey, I just lost my post with the answers to your questions. 
Thank you for your kinds words.
But to keep it simple for you and me, see these links which do have all the answers to your questions about the fluid mechanics.
http://www.princeton.edu/~asmits/Bicycle_web/blunt.html
http://www.see.ed.ac.uk/~johnc/teaching/fluidmechanics4/2003-04/fluids14/separation.html
http://www.engr.usask.ca/~drs694/bluff-body_aerodynamics.htm
But the most important concept with the dimples is the tripping of the boundary layer, i.e. forcing the boundary layer to become turbulent and thus stay close to the wall. You can achieve also do this by roughening the surface of the wall. I've seen suggestions of using e.g. sand or other small roughening objects to trip the boundary layer.
Thank you for your kinds words.
But to keep it simple for you and me, see these links which do have all the answers to your questions about the fluid mechanics.
http://www.princeton.edu/~asmits/Bicycle_web/blunt.html
http://www.see.ed.ac.uk/~johnc/teaching/fluidmechanics4/2003-04/fluids14/separation.html
http://www.engr.usask.ca/~drs694/bluff-body_aerodynamics.htm
But the most important concept with the dimples is the tripping of the boundary layer, i.e. forcing the boundary layer to become turbulent and thus stay close to the wall. You can achieve also do this by roughening the surface of the wall. I've seen suggestions of using e.g. sand or other small roughening objects to trip the boundary layer.
Interesting, you know I had thought it was best to make it like a golf ball, very smooth but dimpled. I find the MDF rougher than I like, but maybe it is a good thing. For my prototypes I have been using MDF, but considered real timber for the real thing to give a smooth finish. Hmmmm .... your thoughts? ... (it might come down to cosmetic, though)
Very interesting looking links. I'll have to have a better look at them later, looks like something I'll have to get into a bit more. Thanks for the window into a new area of learning 😀
Very interesting looking links. I'll have to have a better look at them later, looks like something I'll have to get into a bit more. Thanks for the window into a new area of learning 😀
paulspencer said:
I agree. However, my point is that keeping the Reynolds number low is key to avoiding turbulence (large dimensions in relation to a certain flow velocity). It is only when we want to make the vents too small for laminar flow that aerodynamic tricks such as dimples are needed: it postpones full separation of the flow by "pre-disturbing" the boundary layer. Remember that we only can ameliorate the situation: it is not optimal. So, while B&W are certainly correct in the general case, it depends on the particular dimensions. For the amateur who do not want to experiment with dimples, going for the flared 100 mm port is playing it safe. However, if getting first hand empirical knowledge is your thing: go ahead. Get dimpled! 🙂
/Kranis
The effectiveness of the dimples or anything else to cause the boundary layer to trip will be dependent on its size/height relative to the boundary layer thickness. So, I would doubt if the roughness of the MDF itself would be all that significant. If you were good at pottery or using clay, you might be able to form ridges in the clay that are spaced apart and then fire the clay to make it hard. If the ridges are sufficiently deep (in) or stick out enough (out), you should be able to get a similar effect as dimples. As for the sand idea, you could get glue and glue the sand onto the ports. If you shape the sand a little as you apply it to forms hills and valleys, this should also work. This is the quick and dirty way to test how much difference you would get.
They should be great resources for you. I would ask them if they had a copy of the textbook by Schlichting -- its the classic work on boundary layer theory and it might have information on what other things can be done to trip the boundary. Incidentally, its called "trip or tripping" because the original experiments used a "trip wire."
You might have already stated the answer, but how were you planning on making the dimples?
I did some experiments today on the vent, very interesting results. Without flares there is a LOT of turbulence in my test box. I chose a sine wave where the velocity is at its maximum and got the driver moving to about 40mm p-p excursion. Worst case scenario basically. Without flares, the air seems to project a long way past the vent at high velocity.
Just a small 18mm radius flare makes a big difference! The air flow is still not much changed regarding projection of high velocity air.
Next I tried 36mm dimpled radius (can't find non dimpled version). Suddenly things change. It is as if the velocity is 20% now!!! Putting your hand in front it has changed from a violent movement which chuffs to a more gently moving mass of air. There is still some turbulence.
Next 72mm non dimpled radius. A step better again. Very little turbulence. In fact it appears that with a rumble filter used, in the listening position, that there would be no problem at all with turbulence.
I notice before turbulence gets objectionable (chuffing), it has a more subtle effect on the sound. I noticed this when comparing the sound of the vented vs sealed version with the same tone. Putting the flares on made the vented version sound again like the sealed box.
I have two punches machined up on the lathe. Quick and easy way to get dimples. I can get fairly close to a gold ball look, but I tend to make the dimples deeper. I suspect B&W got it right with their research team.
After some of your comments, I have spoken to them both. My sister may in fact be able to model it for me. Now I think it comes down mostly to trying things out.
My initial test of a dimpled vent doesn't really tell me if it works in my case. I couldn't find the identical non dimpled flare. I suspect that I might not notice a significant difference, or that I won't need it anyway. I think I will probably do it anyway, makes for a distinctive look and an interesting talking point!
Also, the fit with the flare and the PVC isn't quite right, this is more significant than the dimples at the moment. While the 72mm flare is better than the dimpled 36mm flare, this may be due to better fit, the former being done a little better, the latter has a bit of a lip.
I'd like to do a fairly extensive write up on my site of the design process of this sub, and include all this stuff ... as well as photos.
I have also had to start de-rattling the room - wedged cardboard between the window pane and frame where it was loose, and have put blu tak behind all the picture frames. The de-rattling challenge is not small. The rattle is far worse than vent turbulence could ever be. Hmmm, maybe I might write an article "what they didn't tell you about subwoofer design..."
Another thing I have to deal with is fan noise in the amp, again, worse than vent turbulence! ... I plan to build a box around it, like a muffler. Construction similar to a TL actually ... speaking of which ....
First do a search for "martin King" "mathcad" "transmission line" on this forum ... you will get stacks ... read up first
Then post a new thread if you need to ...
Just a small 18mm radius flare makes a big difference! The air flow is still not much changed regarding projection of high velocity air.
Next I tried 36mm dimpled radius (can't find non dimpled version). Suddenly things change. It is as if the velocity is 20% now!!! Putting your hand in front it has changed from a violent movement which chuffs to a more gently moving mass of air. There is still some turbulence.
Next 72mm non dimpled radius. A step better again. Very little turbulence. In fact it appears that with a rumble filter used, in the listening position, that there would be no problem at all with turbulence.
I notice before turbulence gets objectionable (chuffing), it has a more subtle effect on the sound. I noticed this when comparing the sound of the vented vs sealed version with the same tone. Putting the flares on made the vented version sound again like the sealed box.
I have two punches machined up on the lathe. Quick and easy way to get dimples. I can get fairly close to a gold ball look, but I tend to make the dimples deeper. I suspect B&W got it right with their research team.
After some of your comments, I have spoken to them both. My sister may in fact be able to model it for me. Now I think it comes down mostly to trying things out.
My initial test of a dimpled vent doesn't really tell me if it works in my case. I couldn't find the identical non dimpled flare. I suspect that I might not notice a significant difference, or that I won't need it anyway. I think I will probably do it anyway, makes for a distinctive look and an interesting talking point!
Also, the fit with the flare and the PVC isn't quite right, this is more significant than the dimples at the moment. While the 72mm flare is better than the dimpled 36mm flare, this may be due to better fit, the former being done a little better, the latter has a bit of a lip.
I'd like to do a fairly extensive write up on my site of the design process of this sub, and include all this stuff ... as well as photos.
I have also had to start de-rattling the room - wedged cardboard between the window pane and frame where it was loose, and have put blu tak behind all the picture frames. The de-rattling challenge is not small. The rattle is far worse than vent turbulence could ever be. Hmmm, maybe I might write an article "what they didn't tell you about subwoofer design..."
Another thing I have to deal with is fan noise in the amp, again, worse than vent turbulence! ... I plan to build a box around it, like a muffler. Construction similar to a TL actually ... speaking of which ....
First do a search for "martin King" "mathcad" "transmission line" on this forum ... you will get stacks ... read up first
Then post a new thread if you need to ...
Excellent Paul!
I'm glad to see you are able to get some things tested. The results don't surprise me. The velocity should go wiith 1 / (exit area of the tube) or 1/(diameter squared) where the diameter is the actual diameter of the tube -- not the radius of the flare. So if the flare ends up making the exit diameter twice as big, the velocity should be 1/4.
It seems as if you started with the resonance frequency of the non-flared port as you used the sine wave to hit the maximum velocity. It would be interesting to test what happens if you started out with the resonance frequency with a flared configuration and removed the flare. I would assume that the resonance frequency for the sine wave to have maximum velocity should change whether it was flared or not.
I'm glad to see you are able to get some things tested. The results don't surprise me. The velocity should go wiith 1 / (exit area of the tube) or 1/(diameter squared) where the diameter is the actual diameter of the tube -- not the radius of the flare. So if the flare ends up making the exit diameter twice as big, the velocity should be 1/4.
It seems as if you started with the resonance frequency of the non-flared port as you used the sine wave to hit the maximum velocity. It would be interesting to test what happens if you started out with the resonance frequency with a flared configuration and removed the flare. I would assume that the resonance frequency for the sine wave to have maximum velocity should change whether it was flared or not.
Bilbo,
I must admit, I usually discuss audio more than the time I spend building / tweaking! ... hands on does change some things. Although, at the same time, my understanding of theory helps me make sense of things. eg. I knew that vent velocity reached a maximum somewhere near tuning and rolled off rapidly either side. This box is tuned @ 18 Hz, but max velocity is at about 15 Hz. At 20 Hz the turbulence at first seemed far worse, however when I put on the flare, I found that there was in fact more audible output in the vent, so when combined with a little turbulence, sounded louder ... I really had not anticipated the change that took place at the end of the flare, it's a very strong endorsement of large flares. The flare with the larger radius was quite a bit better than the smaller version, so I will make the flare as large as I think looks good.
Precision ports consider that the effective length of their flares extends half the length beyond the point where the flares start, for each flare. This means, the flares would lower tuning just slightly, but the difference would be so small it would not be audible.
Hopefully today I can zero in on dimpled vs non dimpled. If I can get my mic working again, then I can also measure with the RTA on ultracurve! ... well not so much measure, but see. RTA confirmed my suspicion that those sine wave generators on PC put out a lot of output either side of the freq you select. Choose a 10 Hz tone and you still get output a few octaves either side!!! Also if you put them on a loop in speaker workshop, you get a 40 Hz thump each time it repeats - extremely annoying if you turn up a 15 Hz tone in an attempt not to disturb others in another room!!!
I must admit, I usually discuss audio more than the time I spend building / tweaking! ... hands on does change some things. Although, at the same time, my understanding of theory helps me make sense of things. eg. I knew that vent velocity reached a maximum somewhere near tuning and rolled off rapidly either side. This box is tuned @ 18 Hz, but max velocity is at about 15 Hz. At 20 Hz the turbulence at first seemed far worse, however when I put on the flare, I found that there was in fact more audible output in the vent, so when combined with a little turbulence, sounded louder ... I really had not anticipated the change that took place at the end of the flare, it's a very strong endorsement of large flares. The flare with the larger radius was quite a bit better than the smaller version, so I will make the flare as large as I think looks good.
Precision ports consider that the effective length of their flares extends half the length beyond the point where the flares start, for each flare. This means, the flares would lower tuning just slightly, but the difference would be so small it would not be audible.
Hopefully today I can zero in on dimpled vs non dimpled. If I can get my mic working again, then I can also measure with the RTA on ultracurve! ... well not so much measure, but see. RTA confirmed my suspicion that those sine wave generators on PC put out a lot of output either side of the freq you select. Choose a 10 Hz tone and you still get output a few octaves either side!!! Also if you put them on a loop in speaker workshop, you get a 40 Hz thump each time it repeats - extremely annoying if you turn up a 15 Hz tone in an attempt not to disturb others in another room!!!
paulspencer said:
Paul
Here's some info about that locating the vent opening directly behind the driver can be a problem. Not a great deal of information but does say it's a no no.
http://www.speakerbuilder.net/web_files/Projects/RT3/rt3.htm
Good to see you're right into it.
I wonder if small concentric grooves forming hills and valleys would have the same result as dimples........ much easier to manufacture as they can be turned.
Rabbitz,
He only really mentions port noise. I don't see how his cross brace would change it at all ... hmmmm. I would expect problems with an unflared/unflanged internal vent.
I will have to see how it goes. I might have to try a vent setup that is similar to my intended design with the prototypes. Nothing better than an actual prototype to tell you if you are on track.
I was more concerned about how it might change the tuning in some way, that the vent might not couple correctly to the entire box, or that I might get out of band noises. Hmmmmm
....
in other news ... I have just shifted the sub next to the mains - one per side and I like it better. I calibrated using the autoeq function of ultracurve fullrange. It seems to give a more natural balance.
Got a pile of diy stuff in the works right now 😀
Ho ho ho, I love christmas!!!
He only really mentions port noise. I don't see how his cross brace would change it at all ... hmmmm. I would expect problems with an unflared/unflanged internal vent.
I will have to see how it goes. I might have to try a vent setup that is similar to my intended design with the prototypes. Nothing better than an actual prototype to tell you if you are on track.
I was more concerned about how it might change the tuning in some way, that the vent might not couple correctly to the entire box, or that I might get out of band noises. Hmmmmm
....
in other news ... I have just shifted the sub next to the mains - one per side and I like it better. I calibrated using the autoeq function of ultracurve fullrange. It seems to give a more natural balance.
Got a pile of diy stuff in the works right now 😀
Ho ho ho, I love christmas!!!
There's nothing like building the thing and testing it to see what works...... lot the theories and BS goes out the window then as you can hear it.
I think his x brace was to redirect the flow of air so it couldn't go directly to the port opening. In all my designs I've always placed to port opening so it's not behind the woofer........ elswhere in the box or at right angles to it with my subs as they have downfiring ports. Could be just a mental block but it just seemed logical to me not having the air going directly into the port but from the pressure build up in the box. Is the pressure distrubuted evenly........ dunno, fluid mechanics was not my strong point during my engineering days.
I'm sure you'll come up with some answers and the conclusions will be interesting.
I think his x brace was to redirect the flow of air so it couldn't go directly to the port opening. In all my designs I've always placed to port opening so it's not behind the woofer........ elswhere in the box or at right angles to it with my subs as they have downfiring ports. Could be just a mental block but it just seemed logical to me not having the air going directly into the port but from the pressure build up in the box. Is the pressure distrubuted evenly........ dunno, fluid mechanics was not my strong point during my engineering days.
I'm sure you'll come up with some answers and the conclusions will be interesting.
Rabbitz, this is what I've been grappling with. Given my observations so far on the behaviour of large flares, I think a large flare on the inside of the vent is a good idea. I might just make the internal flare 72mm radius as well - why not, even if not needed it will reduce port compression.
Intuitively I have thought that the sound from the driver might end up in the vent before it first pressurises the box. However, logically this sounds to me as nonsense. It should instantly pressurise the box. As long as there are no turbulent issues, I suspect that the location won't matter, as long as obstructions are not too close so as to alter tuning by altering the effective length.
In actual use I've never got anywhere near turbulence, even with an unflared vent! In all this I'm catering to the most extreme use.
Today I measured typical levels of 75 - 90 db for movies. The highest peak at listening position was around 105 db. What is interesting is that 105 db is quite a bit louder than I had thought. When I was in high school I did my first speaker project using an SPL meter, signal generator and some el cheapo speakers - I measured a max SPL at 1m of 103 db and I recall at 3 khz that was actually extremely irritating and a bit painful on the ears!
Intuitively I have thought that the sound from the driver might end up in the vent before it first pressurises the box. However, logically this sounds to me as nonsense. It should instantly pressurise the box. As long as there are no turbulent issues, I suspect that the location won't matter, as long as obstructions are not too close so as to alter tuning by altering the effective length.
In actual use I've never got anywhere near turbulence, even with an unflared vent! In all this I'm catering to the most extreme use.
Today I measured typical levels of 75 - 90 db for movies. The highest peak at listening position was around 105 db. What is interesting is that 105 db is quite a bit louder than I had thought. When I was in high school I did my first speaker project using an SPL meter, signal generator and some el cheapo speakers - I measured a max SPL at 1m of 103 db and I recall at 3 khz that was actually extremely irritating and a bit painful on the ears!
It's just a mental block with me with the port opening being behind the driver and has no foundation whatsoever and I see it like filling a sink with the plug left out which is BS.... silly hey.
I know that the pressure inside the box should be equal in all directions so should not be a problem. The only thing I can think of is some higher frequencies from the driver getting through the port before they can be absorbed or whatever but a good steep xo should stop that. Most plate amps use a 2nd order xo which does let some higher info through (depends on where you xo) which you are not using anyway. It's one of the things I like about downfiring subs even though I can't have one so I've gone side firing.
It is surprising when you find out the average levels for normal movie viewing is low. Been through this at home when someone, who remains unnamed, has told me that it's too loud and I've brought out the sound level meter. 105dB at the listening position is very, very, very loud. Anything over 100dB at home is in my books and the higher the frequency the worse it is. Most of the guys I know in the engineering game that have hearing loss, have lost it due to loud higher frequencies. Give me loud thump over loud hiss any day.
I know that the pressure inside the box should be equal in all directions so should not be a problem. The only thing I can think of is some higher frequencies from the driver getting through the port before they can be absorbed or whatever but a good steep xo should stop that. Most plate amps use a 2nd order xo which does let some higher info through (depends on where you xo) which you are not using anyway. It's one of the things I like about downfiring subs even though I can't have one so I've gone side firing.
It is surprising when you find out the average levels for normal movie viewing is low. Been through this at home when someone, who remains unnamed, has told me that it's too loud and I've brought out the sound level meter. 105dB at the listening position is very, very, very loud. Anything over 100dB at home is in my books and the higher the frequency the worse it is. Most of the guys I know in the engineering game that have hearing loss, have lost it due to loud higher frequencies. Give me loud thump over loud hiss any day.
While doing some surfing on your topic, I came across this website which deals with flow in dcts. Interestingly it gives some techniques such that you can create a low pass acoustic filter and a high pass acoustic filter, so you can deal with the noise from a ported flow.
http://www.kettering.edu/~drussell/GMI-Acoustics/Filters.html
http://www.kettering.edu/~drussell/GMI-Acoustics/Filters.html
I wouldnt fear high frequencys coming through the port that much, you should put the port opening in line with the pole peice so not only does the magnet block the possibilty of leaked high frequency but the port flow cools the moter and enhances your power handling slightly. I love my bass loud, but what I really love is infrasound low at extremely high levels. Earthquake reproduction is what its like
Do you ever notice when someone walks in your house and closes the door quickly it flexes the windows in the other rooms. Well dont you wish you could have a subwoofer that could do that, because im sure some movie out there has got to have at least one proper recording of a door slamming and the infrasonic wave that it creates. I know its not Lord of the Rings
Do you ever notice when someone walks in your house and closes the door quickly it flexes the windows in the other rooms. Well dont you wish you could have a subwoofer that could do that, because im sure some movie out there has got to have at least one proper recording of a door slamming and the infrasonic wave that it creates. I know its not Lord of the Rings
Rabbitz,
based on listening and measuring yesterday, I have a new target:
105 db clean peak SPL at listening position for the midrange
103 db treble
Doesn't sound that high does it?
Until you realise that I lose 10 db from 1m to my listening position
When I get back to active and get more power to my mids, I should be able to pull it off, although I think I'll need a horn dome tweeter.
Bilbo,
thanks for the link - it's actually very helpful for another little project of mine - a muffler box for my Behringer amp!
BA,
I've found that movies that have exaggerated bass (LOTR, matrix) and a lot of action / sci fi movies, sound best with bass that is reasonably flat. I expected that I'd need more boost, but when I calibrated the bass flat in room, I was surprised. I found with even what I considered gentle bass boost made the bass overpowering and overdone.
Looking at such movies with the RTA of Ultracurve going, I've noticed how they have built in bass boost that gives those effects like an exaggerated door slam, or those sounds in the matrix when they go into slow motion spinning around one of those fly in the air kicks. Typically you see boost centred around 30 Hz which gently rolls off either side - hence it is boosted between 20 - 40 Hz and this gives a sensation of weight and power. I'm not sure how much boost they use, probably about 5 db. But if you have flat in room response down there and you put in your own boost, and add it to theirs, it gets a bit much. And I will add that I've always liked more and deeper bass than most people - my idea of just right is many other people's idea of "too much." My most common criticism of most sound systems has always been "not enough bass."
based on listening and measuring yesterday, I have a new target:
105 db clean peak SPL at listening position for the midrange
103 db treble
Doesn't sound that high does it?
Until you realise that I lose 10 db from 1m to my listening position
When I get back to active and get more power to my mids, I should be able to pull it off, although I think I'll need a horn dome tweeter.
Bilbo,
thanks for the link - it's actually very helpful for another little project of mine - a muffler box for my Behringer amp!
BA,
I've found that movies that have exaggerated bass (LOTR, matrix) and a lot of action / sci fi movies, sound best with bass that is reasonably flat. I expected that I'd need more boost, but when I calibrated the bass flat in room, I was surprised. I found with even what I considered gentle bass boost made the bass overpowering and overdone.
Looking at such movies with the RTA of Ultracurve going, I've noticed how they have built in bass boost that gives those effects like an exaggerated door slam, or those sounds in the matrix when they go into slow motion spinning around one of those fly in the air kicks. Typically you see boost centred around 30 Hz which gently rolls off either side - hence it is boosted between 20 - 40 Hz and this gives a sensation of weight and power. I'm not sure how much boost they use, probably about 5 db. But if you have flat in room response down there and you put in your own boost, and add it to theirs, it gets a bit much. And I will add that I've always liked more and deeper bass than most people - my idea of just right is many other people's idea of "too much." My most common criticism of most sound systems has always been "not enough bass."
Paul
Phew, at first I thought it was 105dB average and not peak. I don't think 105dB peak too high as the average would roughly be about 95dB which is still loud at the listening position.
Peerless make a horn dome tweeter, 811647, with a sensitivity of 99.5dBm to 15º off axis but I don't know how good it is. With a 100W max, that's a lot of dB's, roughly 118. Have used their 810653 and 811827 which are quite good so Peerless can make good tweeters as well as exceptional mid woofers and woofers.
Phew, at first I thought it was 105dB average and not peak. I don't think 105dB peak too high as the average would roughly be about 95dB which is still loud at the listening position.
Peerless make a horn dome tweeter, 811647, with a sensitivity of 99.5dBm to 15º off axis but I don't know how good it is. With a 100W max, that's a lot of dB's, roughly 118. Have used their 810653 and 811827 which are quite good so Peerless can make good tweeters as well as exceptional mid woofers and woofers.
- Status
- Not open for further replies.