Jump to content


Photo

Is it possible to model port compression?


  • Please log in to reply
36 replies to this topic

#21 Ricci

Ricci

    Super Bass Overlord

  • Moderators
  • 1,067 posts
  • LocationLouisville, KY

Posted 31 October 2016 - 07:47 PM

I was curious so I did some quick back of the napkin calculations.

 

Compression is maximum near 13.5Hz at about 3.8dB during the 103.8 volt sweep, which is exactly where we would expect the velocity to be the highest based on the models. By simply using the voltage applied in a simulation and allowing for 3 to 3.8dB less output from 12-16Hz after compression, it looks like the MAUL was producing about 30m/s or so during the 103.8volt sweep.The 185 volts sweep had the amplifier current limiting severely so it is hard to tell how much of the extra compression was partially the amplifier. Regardless a quick velocity estimate is a maximum of a bit above 40m/s. We don't have burst data for 13.5Hz which is the worst case scenario for velocity for the MAUL but a quick look at the maximum levels at 12.5Hz and 16Hz compared to the 103.8volt sweep results in an estimated 45m/s at 16Hz and 61m/s at 12.5Hz. I did consider the flare-it program when designing the MAUL and it suggested chuffing at 37m/s and core limit of 80m/s for a 12" pipe.

 

This is all very rough work but based on this I think it safe to say that the vent has not completely brick walled at 50-60m/s but is heavily compressing the output by 7-8dB at that point and for all practical purposes may be "there". Huge increases in input power to the drivers will continue to produce less and less output increase from the vent. Of course this is also at quite low frequencies down near 12-16Hz too so that should be considered. Yes this is accompanied by significant air noise as well. Maximum velocities in the short horn section are only an estimated 20-30m/s.

 

It might be worth a look at some other vented system tests to examine the compression behavior and guesstimated vent velocities.


  • Infrasonic, Kvalsvoll and SME like this

#22 Kvalsvoll

Kvalsvoll

    Ultra Member

  • Members
  • PipPipPipPipPipPip
  • 361 posts
  • LocationNorway

Posted 01 November 2016 - 06:48 PM

@Ricci, I find it generous of you to share this data, very useful information.


My audio blog - kvalsvoll.blogspot.com


#23 deepthoughts

deepthoughts

    Member

  • Members
  • PipPip
  • 25 posts

Posted 02 November 2016 - 04:36 PM

I was curious so I did some quick back of the napkin calculations.

 

Compression is maximum near 13.5Hz at about 3.8dB during the 103.8 volt sweep, which is exactly where we would expect the velocity to be the highest based on the models. By simply using the voltage applied in a simulation and allowing for 3 to 3.8dB less output from 12-16Hz after compression, it looks like the MAUL was producing about 30m/s or so during the 103.8volt sweep.The 185 volts sweep had the amplifier current limiting severely so it is hard to tell how much of the extra compression was partially the amplifier. Regardless a quick velocity estimate is a maximum of a bit above 40m/s. We don't have burst data for 13.5Hz which is the worst case scenario for velocity for the MAUL but a quick look at the maximum levels at 12.5Hz and 16Hz compared to the 103.8volt sweep results in an estimated 45m/s at 16Hz and 61m/s at 12.5Hz. I did consider the flare-it program when designing the MAUL and it suggested chuffing at 37m/s and core limit of 80m/s for a 12" pipe.

 

This is all very rough work but based on this I think it safe to say that the vent has not completely brick walled at 50-60m/s but is heavily compressing the output by 7-8dB at that point and for all practical purposes may be "there". Huge increases in input power to the drivers will continue to produce less and less output increase from the vent. Of course this is also at quite low frequencies down near 12-16Hz too so that should be considered. Yes this is accompanied by significant air noise as well. Maximum velocities in the short horn section are only an estimated 20-30m/s.

 

It might be worth a look at some other vented system tests to examine the compression behavior and guesstimated vent velocities.

 

LspCAD 5 includes an optional attempt at modeling port compression (can check a box to see with and without model).  Unfortunately it doesn't really allow a simple model of a commonly flared port.  Instead they have a model more akin to a flared vent that defines more of an hourglass shape with an effective, assumed radius greater than 1/2 the length of the port. Comparing straight ports and more significantly flared ports gives some interesting points of consideration.

 

One thing overlooked in your assumption above is that as the port compresses, the driver excursion increases.  This is seen very easily in the LspCAD models and can be estimated/correlated by comparing high level impedance sweeps around tuning, and comparing driver vs port near field measurements.  With high excursion woofers you will often see the driver picking up a good bit of the output load as excursion increases in an exponential manner due to port compression.  This actually reduces the observed output compression, so remember the port itself is even more non-linear than the total SPL compression suggests.  As with most things, the models are overly conservative and compression isn't quite as severe as suggested, but it's a lot more than none.  Long ago I recall Deon Bearden relaying on that in his testing and research most port designs start compressing anywhere past 10m/s.  That's not to say the ports aren't useful past that point, but we should understand that behavior isn't linear, just as with real drivers well before the rated Xmax.

 

The complicating factor is always the wide bandwidth, complex signal consideration.  If a component of a complex signal pushes a port into severe compression, how does this impact the rest of the complex signal being produced at the same time?  


  • Ricci, SME and andy497 like this

#24 Ricci

Ricci

    Super Bass Overlord

  • Moderators
  • 1,067 posts
  • LocationLouisville, KY

Posted 02 November 2016 - 06:16 PM

One thing overlooked in your assumption above is that as the port compresses, the driver excursion increases.  This is seen very easily in the LspCAD models and can be estimated/correlated by comparing high level impedance sweeps around tuning, and comparing driver vs port near field measurements.  With high excursion woofers you will often see the driver picking up a good bit of the output load as excursion increases in an exponential manner due to port compression.  This actually reduces the observed output compression, so remember the port itself is even more non-linear than the total SPL compression suggests.  As with most things, the models are overly conservative and compression isn't quite as severe as suggested, but it's a lot more than none.  Long ago I recall Deon Bearden relaying on that in his testing and research most port designs start compressing anywhere past 10m/s.  That's not to say the ports aren't useful past that point, but we should understand that behavior isn't linear, just as with real drivers well before the rated Xmax.

 

Great points. I have observed exactly that with many vented subs while testing them. I have no idea how big of a contribution it would have made in that specific case but we can probably assume that airspeeds were a bit lower than posted above. 40m/s through a 12" pipe is quite the breeze let me say. :D

 

You've got experience with passive radiators. Have you ever seen a pair of PR's appear to go out of phase with each other, when being driven very hard? I watched this happen a number of times with both the 15" and 12" TC VMP's during compression sweeps.

 

Ports are a tough compromise. It'd be easy to say simply design to stay below 10-20m/s at maximum output of the system, by using a ridiculously huge vent area and a giant air volume but that is not in the cards most of the time. Large flares become difficult to use sometimes as well. Driver design keeps moving towards higher displacement and power handling while huge amplifier power becomes cheap. Everyone wants a compact sub that also goes deep and offers high output. There's no getting around the physics behind Helmholtz resonators. Trying to juggle vent length, shape, area and pipe resonance, against a overall size that is manageable, it is often vent area that gets cut back. That's why you see 15 and 18" drivers with 3 and 4" ports or thin slot ports. PR's have advantages but also some shortcomings themselves. Cost being one and a lot of real estate on the baffle/s for them being another. Also the advantages of running them in opposed pairs make that arrangement almost a necessity. Mechanical displacement limits is another one.


  • Infrasonic likes this

#25 SME

SME

    Super Bass Overlord

  • Members
  • PipPipPipPipPipPipPipPip
  • 962 posts

Posted 03 November 2016 - 09:49 AM

It's almost as if there is some factor they are missing or forgetting to include in the Reynolds number or it is not describing what we think it is. It seems to indicate that larger vent areas start to have major issues at lower velocities than smaller ones. However measurements seem to suggest that the useful core velocity limit (an output wall so to speak) is actually higher for larger vent areas not lower.

 

There's nothing wrong with the Reynolds number (Re).  The issue is only in how it's interpreted, which is problem specific.  The Moody Diagram was developed from empirical data of steady fluid flow in pipes.  It is most valid for fully developed flows.  My fluids textbook suggests that fully developed flow is only realized at on the order of 50 pipe diameters from the entrance or a major disturbance.  So there's the first issue with trying to relate information from this diagram to subwoofer ports:  Most subs don't have port tubes that long.  The second issue is that the flow in subs is unsteady; it oscillates.

 

With that said, we can still look at the trends in pipes and make educated guesses about what the implies for ports.  In pipes, the transition from laminar to turbulent flow happens in the 2-5k range of Re.  That's a lot less than the 50-100k "wall" reported in that article.  It suggests that subwoofer ports routinely exhibit turbulent flows at velocities that would be considered "well within normal operating range".  I think the conclusion that the sudden rise in compression is caused by laminar to turbulent flow transition is wrong.  Instead, I believe flow may be hitting a saturation point well into the turbulent flow regime, typically at some bottleneck point in the port system.  Often this will be around the entrance/exit.  The particular Re where this saturation effect occurs probably depends a lot on the design.  At the same time, there is still probably some compression that sets in after transition to turbulent flow even at much lower Re.  It's very interesting that increased driver excursion may have the effect of hiding that compression.

 

From the data in the paper and ensuing discussion, I have to say there's a lot to not like about ported systems.  :)  As a compromise, it is often an ugly one.  I know Bossobass Dave used to rail about how the frequency response of a ported system changes with output level.  It's true, and you can't really do anything to fix it like you can to reshape a sealed sub with high Fs and/or low Qtc.  The insight that port compression causes driver excursion to increase is especially concerning, being that excursion reduction is one of the biggest advantages of a ported design.  The data in the paper also illustrates that tuning frequency can shift upwards rather dramatically at high output levels as well.  These two changes (non-linear increase in excursion and upward shift in tuning frequency) made coincide with one another.



#26 andy497

andy497

    Member

  • Members
  • PipPip
  • 18 posts

Posted 04 November 2016 - 04:48 PM

Has any work been done on controlling the boundary layer in ports to delay separation? (e.g. targeted roughness, turbulators, etc.)  I know you need to have a precise idea of the Re range, and any given solution will be very specific to that port geometry.  Also, I imagine everything may be different in an acoustic resonator; not only is flow not fully developed, it's constantly changing direction.  Still, it sounds like an interesting area of study.



#27 mwmkravchenko

mwmkravchenko

    Newbie

  • Members
  • Pip
  • 6 posts

Posted 13 February 2017 - 03:41 AM

Calculating this kind of thing is a job for a FEA program.  

 

One thing I noticed in the comments above by SME.  You are applying fluid flow calculations to air flow calculations.  There are some similarities between the two mediums.  Where the differences begin is that a fluid is pretty much non-compressible.  Whereas a gas is.  And that is where we run into problems when we want to do some useful calculations in the junctions of the interior of a container with a port and the the end of the port.  Much has to be taken into account in order to make any useful calculations.  And that makes the calculations a real pain in the but.

 

From lurking around on some of the threads I see that some of you guys are pretty handy with Hornresp.  A few years back I asked David McBean to include particle velocity measurements in Hornresp to use in some very high SPL type of cabinets.  It is not a full around every corner, nook and so on calculation.  But it does get you somewhere in a hurry.

 

The keeping it below 10 meters/second rule of thumb goes a long way into making a serviceable port.  



#28 Ricci

Ricci

    Super Bass Overlord

  • Moderators
  • 1,067 posts
  • LocationLouisville, KY

Posted 13 February 2017 - 03:45 PM

Calculating this kind of thing is a job for a FEA program.  

 

One thing I noticed in the comments above by SME.  You are applying fluid flow calculations to air flow calculations.  There are some similarities between the two mediums.  Where the differences begin is that a fluid is pretty much non-compressible.  Whereas a gas is.  And that is where we run into problems when we want to do some useful calculations in the junctions of the interior of a container with a port and the the end of the port.  Much has to be taken into account in order to make any useful calculations.  And that makes the calculations a real pain in the but.

 

From lurking around on some of the threads I see that some of you guys are pretty handy with Hornresp.  A few years back I asked David McBean to include particle velocity measurements in Hornresp to use in some very high SPL type of cabinets.  It is not a full around every corner, nook and so on calculation.  But it does get you somewhere in a hurry.

 

The keeping it below 10 meters/second rule of thumb goes a long way into making a serviceable port.  

 

 

Hey Mark,

 

 Welcome.

 

I'd agree with your last sentence but the caveat is that it is basically impossible with any type of serviceable design using modern high power drivers. By the time the vents are made that large, the overall size is completely out of control, or the pipe resonances are bad, or you are faced with a large system with little useable output compared with the size of the device, or a combination of all of the above.



#29 mwmkravchenko

mwmkravchenko

    Newbie

  • Members
  • Pip
  • 6 posts

Posted 13 February 2017 - 07:09 PM

Hey Mark,

 

 Welcome.

 

I'd agree with your last sentence but the caveat is that it is basically impossible with any type of serviceable design using modern high power drivers. By the time the vents are made that large, the overall size is completely out of control, or the pipe resonances are bad, or you are faced with a large system with little useable output compared with the size of the device, or a combination of all of the above.

 

Thanks Josh.

 

I found this forum by accident.  And most of the posts I have read are by very thoughtful people.  I'll hang around from time to time and see what's interesting.

 

You are very right in that statement.  And I have no answer other than design to the goals that you are seeking.  A clean sounding bass from say a properly damped sealed enclosure with a low distortion woofer, or a well designed front loaded horn set the standards as to what is clean sounding.  The job of a proper design is to get you there or as close as you can.  Designing a port has always been a set of compromises.  And there are ways to get a decent sound out of a vented enclosure that have not been discussed such as the power port type of vented enclosure.  It has pretty much the lowest turbulence for a given diameter of enclosure.



#30 Kyle

Kyle

    Super Power Member

  • Moderators
  • 193 posts

Posted 13 February 2017 - 11:26 PM

One could write a phd thesis on modeling flow of a fluid -- this simply comes down to FEA, and any other modeling is just an approximation. But lets face it, this is not rocket science, its just a speaker :)
 
I will say there was only one subwoofer I have seen that did not have port issues and it was a twin 18" box with a single 18" port - no curves. Very low velocity, extremely capable. When you see ports with high air velocity then that's normally an indication of being undersized. There are reasons to do this (extension, size etc) but it does mean the port it exhibiting its limits.

  • Infrasonic likes this

#31 mwmkravchenko

mwmkravchenko

    Newbie

  • Members
  • Pip
  • 6 posts

Posted 14 February 2017 - 12:30 AM

 

One could write a phd thesis on modeling flow of a fluid -- this simply comes down to FEA, and any other modeling is just an approximation. But lets face it, this is not rocket science, its just a speaker :)
 
I will say there was only one subwoofer I have seen that did not have port issues and it was a twin 18" box with a single 18" port - no curves. Very low velocity, extremely capable. When you see ports with high air velocity then that's normally an indication of being undersized. There are reasons to do this (extension, size etc) but it does mean the port it exhibiting its limits.

 

Hello Kyle.  I have heard about you from Josh but never had the pleasure of talking with you.  I guess this is a close second.

 

I think quite a few people have written their PHD on that subject!

 

Just a speaker   :o

 

Don't get me started on that one.  The truth of the matter is at very few times are we going to tax the port compression on a properly designed subwoofer.  And the times when that does happen we usually have program material that masks the effect in the first place.

 

And the difference between no problems from a port at any time is the difference between a realistic enclosure size and a not so realistic enclosure size.

 

Twin 18 with an 18 inch diameter port.  Wow.  A genuine phallic idol to the vent gods if there ever was one!



#32 Kyle

Kyle

    Super Power Member

  • Moderators
  • 193 posts

Posted 14 February 2017 - 12:42 AM

Hello Kyle.  I have heard about you from Josh but never had the pleasure of talking with you.  I guess this is a close second.

 

I think quite a few people have written their PHD on that subject!

 

Just a speaker   :o

 

Don't get me started on that one.  The truth of the matter is at very few times are we going to tax the port compression on a properly designed subwoofer.  And the times when that does happen we usually have program material that masks the effect in the first place.

 

And the difference between no problems from a port at any time is the difference between a realistic enclosure size and a not so realistic enclosure size.

 

Twin 18 with an 18 inch diameter port.  Wow.  A genuine phallic idol to vent gods if there ever was one!

 

Hello! I'm the lurker of this forum, lol.

 

Indeed, very good points. I'm often amazed how bad a subwoofer can sound with a sinwave at full volume (capability) around port tuning, but put music into it and it just seems to work and the distortion seems to subside :)

 

The real question is how many more db can you get out of a driver + amp if you increase the box and or port port and the art of the whole process becomes the design trade offs. The double 18 box I spoke of is very huge and very impracticable. A few dB loss here and there for a subwoofer half the size might add a lot of value to most people.



#33 mwmkravchenko

mwmkravchenko

    Newbie

  • Members
  • Pip
  • 6 posts

Posted 14 February 2017 - 03:48 AM

 

Some of the best work done in this field is by people wanting to simulate musical instruments.  And this video is of a mitered organ pipe.  Or a rectangular organ pipe that has a 90 degree bend.  Sometimes they even have a 180 degree bend.  I have a boat load of thesis papers and other studies that get to the number crunching but it's nice to see something rather than read about it.


  • Infrasonic, Ricci and andy497 like this

#34 Ricci

Ricci

    Super Bass Overlord

  • Moderators
  • 1,067 posts
  • LocationLouisville, KY

Posted 14 February 2017 - 05:04 PM

Kyle makes a good point. We often simulate airspeeds under a worst case scenario with a full power sine wave at the airspeed maximum, but that's very rare. most of the time we aren't running the speaker wide open for all it's got and the content is almost always much more transient, wide bandwidth and complex. In a way it's sort of similar to the reason that I now prefer very low qts drivers for their higher efficiency. With complex, wide bandwidth material the power requirements are lower and it leaves more effective headroom in the amplifier on those types of signals. Also less thermal demands on the voice coils.


  • Infrasonic likes this

#35 Kyle

Kyle

    Super Power Member

  • Moderators
  • 193 posts

Posted 14 February 2017 - 11:08 PM

Ya maybe the saying should be: there is no replacement for [sensitivity] 

 

:)



#36 mwmkravchenko

mwmkravchenko

    Newbie

  • Members
  • Pip
  • 6 posts

Posted 14 February 2017 - 11:17 PM

Kyle makes a good point. We often simulate airspeeds under a worst case scenario with a full power sine wave at the airspeed maximum, but that's very rare. most of the time we aren't running the speaker wide open for all it's got and the content is almost always much more transient, wide bandwidth and complex. In a way it's sort of similar to the reason that I now prefer very low qts drivers for their higher efficiency. With complex, wide bandwidth material the power requirements are lower and it leaves more effective headroom in the amplifier on those types of signals. Also less thermal demands on the voice coils.


Probably the best selling point on a high efficiency driver is the correlation between where they are efficient and where the bulk of what is considered "bass" is centered. 60 hertz is where the money is. And many pro drivers will get you higher spl for a given size versus a conventional subwoofer with much less power input. Last point. Amp power never makes up for the nearly 10 dB of greater efficiency. Do the math and ask a very simple power compression question. You will find the answer!

#37 3ll3d00d

3ll3d00d

    Ultra Member

  • Members
  • PipPipPipPipPipPip
  • 413 posts

Posted 25 February 2017 - 12:46 PM

is it possible to approach this from a different angle? i.e. assume that port compression exists and model that effect (on driver excursion and output)

 

this would be analogous to the way you can model the effect of power or excursion so you'd set a port velocity limit of, e.g., 10m/s and then see what happens next.






0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users