X,
Oh yeah , im sure it would work alright as sub, low passed ...... It is a funny thing , while sitting here running sine sweeps into it there is definitely an obnoxious (perhaps ringing horribly) resonance in that 180-300hz range (5th and 7th pipe harmonics) which doesn't really look too terribly obvious in the amplitude/frequency response measurements but you can definitely hear it, so that is why i am thinking it is phase/time domain related to some extent .............. If only I had known about the Fredditech cavity concept (and how well it works) before building this thing i could have integrated it and i bet it would have cleaned this midbass up perfectly
According to what i have read about these devices like the stub/cavity (and this also applies to a parasitic resonator) have this inherent ability to add a 90 degree phase shift in their effective range which looks like it would really help to align the phase of the different outputs on this box within that midbass range (this is according to the Holmimpulse phase chart , like the one i attached to post #1302) ........
So in retrospect i think this box would have worked much better with an added Fredditech Cavity device .... Hmmmmmm.... Maybe i will try to glue one on ... LOL .... but seriously i think i actually can, it would just make the box a few inches taller ..
Last edited:
XRK , TB46 , No problem on sharing that script, you guys have given me a lot so i am happy that i can give back in some way
You have come a long way. Well done.
Thank you Sir! and yeah , i am excited about this , i think i am very close to having some good design nailed down with the Karlflex
...... I am sure some of that script can be rescaled and re-purposed for the Keystone .
Fredditech Karflex Script
Hey All ,
I just thought i should go ahead and make the 99L Fredditech Karflex script available as well ....
This is the cabinet featured in post #1324 and #1331
The outer dimensions are about 27" H x 20" W x 18.25" D
Not the most compact cabinet for a 12" driver but it definitely squeezes a lot of performance out of a Dayton Pa310.
NOTE: Reducing the width to 19" loses about one half of one decibel at 40hz , height can also be reduced slightly if needed but similar losses can be expected.... Fredditech stub height (internal) can be reduced from the default 2.5" down to 1.5" but will be somewhat less effective.... It is probably best just to stick with suggested outer dimensions for best performance at this tuning (according to simulations)
Hey All ,
I just thought i should go ahead and make the 99L Fredditech Karflex script available as well ....
This is the cabinet featured in post #1324 and #1331
The outer dimensions are about 27" H x 20" W x 18.25" D
Not the most compact cabinet for a 12" driver but it definitely squeezes a lot of performance out of a Dayton Pa310.
NOTE: Reducing the width to 19" loses about one half of one decibel at 40hz , height can also be reduced slightly if needed but similar losses can be expected.... Fredditech stub height (internal) can be reduced from the default 2.5" down to 1.5" but will be somewhat less effective.... It is probably best just to stick with suggested outer dimensions for best performance at this tuning (according to simulations)
Code:
|System FREDDITECH 5000 99LITER KARLFLEX FOR THE DAYTON PA-310
|The outer dimensions are 27" H x 20" W x 18.25" D with half-inch ply material
Def_Driver 'PA310'
dD=30cm |Piston
fs=44.4Hz Vas=81.8L Qms=8.63
Qes=0.34 Re=5.7ohm Le=1mH ExpoLe=0.618
System '99L-KARLFLEX'
| vt=99L, (net)
|Rear Chamber=74L, fB=40Hz,
|Front chamber=25L-ish (roughly 17L+8-ish from the Fredditech-Cavity)
|---------------------------------------------------------
| The speaker or driver :
|---------------------------------------------------------
Driver 'Driver' Def='PA310' Node=1=0=18=2
|--------------------------------------------
| The first section (S1, S2 , S3 in Hornresponse lingo):
|--------------------------------------------
Waveguide 'F0'
Node=50=51
wth=48.3cm hth=6.3cm
wmo=48.3cm hmo=6.31cm
Len=43cm
Conical
|OFF
AcouResistance 'Stuffing1' Node=50=51 Ra=30e3Pas/m3 | Simulate moderate stuffing
Duct 'P0' Node=51=2
WD=48.3cm
HD=28.6cm
Len=15cm
VISC=1
Waveguide 'F1'
Node=2=4
wth=48.3cm hth=28.6cm
wmo=48.3cm hmo=28.7cm
Len=34cm
Conical
Duct 'P2' Node=4=5
WD=48.3cm
HD=6.3cm
Len=43cm
Visc=1
| EXPERIMENTAL CAVITY/DUCT/STUB
| Suggested by Freddi from DIYaudio.com
| This is a very effective method of suppressing the
| upper pipe harmonics and therefore tames and smooths response
Duct 'P3' Node=5=0
WD=48.3cm
HD=6.3cm
Len=35cm
VISC=50
|---------------------------------------------
| The Aperture:
|(in this case the classic K-aperture,type #1)
|The laid back orientation of the baffle and particular slot contour was
|suggested by Freddi of DIYaudio.com, who has a keen feel for Karlson technology.
| =)
|---------------------------------------------
| This first set of code defines the front chamber
|(behind the aperture and in front of the baffle)
| from top to bottom
|---------------------------------------------------
Duct 'P6' Node=5=12
WD=48.3cm
HD=9cm
Len=6.66cm
Visc=1
Duct 'P7' Node=12=13
WD=48.3cm
HD=8.25cm
Len=6.66cm
Visc=1
Duct 'P8' Node=13=14
WD=48.3cm
HD=7.5cm
Len=6.66cm
Visc=1
Duct 'P9' Node=14=15
WD=48.3cm
HD=6.75cm
Len=6.66cm
Visc=1
Duct 'P10' Node=15=16
WD=48.3cm
HD=6cm
Len=6.66cm
Visc=1
Duct 'P11' Node=16=17
WD=48.3cm
HD=5.25cm
Len=6.66cm
Visc=1
Duct 'P12' Node=17=18
WD=48.3cm
HD=4.5cm
Len=6.66cm
Visc=1
Duct 'P13' Node=18=19
WD=48.3cm
HD=3.75cm
Len=6.66cm
Visc=1
Duct 'P14' Node=19=20
WD=48.3cm
HD=3cm
Len=6.66cm
Visc=1
|---------------------------------------------------------------------
|The following code defines the gap and the depth of the K-aperture
|itself (front panel) split into 9 parts. XRK suggests 9 parts minimum =)
|---------------------------------------------------------------------
Duct 'P15' Node=12=22
WD=6cm
HD=6.66cm
Len=1.19cm
Visc=1
Duct 'P16' Node=13=23
WD=6cm
HD=6.66cm
Len=1.19cm
Visc=1
Duct 'P17' Node=14=24
WD=6cm
HD=6.66cm
Len=1.19cm
Visc=1
Duct 'P18' Node=15=25
WD=7cm
HD=6.66cm
Len=1.19cm
Visc=1
Duct 'P19' Node=16=26
WD=8cm
HD=6.66cm
Len=1.19cm
Visc=1
Duct 'P20' Node=17=27
WD=10cm
HD=6.66cm
Len=1.19cm
Visc=1
Duct 'P21' Node=18=28
WD=25cm
HD=6.66cm
Len=1.19cm
Visc=1
Duct 'P22' Node=19=29
WD=35cm
HD=6.66cm
Len=1.19cm
Visc=1
Duct 'P23' Node=20=30
WD=47.85cm
HD=6.66cm
Len=1.19cm
Visc=1
|--------------------------------------------------------------------
|(K-aperture in multiple parts)series tuned.
|The following code defines the radiator's positions and also refers
|to the ducts above for their dimensions..
|Add a pipe character in front of radiator's code line in order to
|close off that part of the aperture in simulation
|--------------------------------------------------------------------
Radiator 'Rad_P15' DEF='P15' Node=22
x=0 y=28cm z=0 HAngle=0 VAngle=0
Radiator 'Rad_P16' DEF='P16' Node=23
x=0 y=21.33cm z=0 HAngle=0 VAngle=0
Radiator 'Rad_P17' DEF='P17' Node=24
x=0 y=15cm z=0 HAngle=0 VAngle=0
Radiator 'Rad_P18' DEF='P18' Node=25
x=0 y=8.33cm z=0 HAngle=0 VAngle=0
Radiator 'Rad_P19' DEF='P19' Node=26
x=0 y=0 z=0 HAngle=0 VAngle=0
Radiator 'Rad_P20' DEF='P20' Node=27
x=0 y=-8.33cm z=0 HAngle=0 VAngle=0
Radiator 'Rad_P21' DEF='P21' Node=28
x=0 y=-15cm z=0 HAngle=0 VAngle=0
Radiator 'Rad_P22' DEF='P22' Node=29
x=0 y=-21.33cm z=0 HAngle=0 VAngle=0
Radiator 'Rad_P23' DEF='P23' Node=30
x=0 y=-28cm z=0 HAngle=0 VAngle=0
Last edited:
I will have the #1 please, with the crab puffs, to go
Freddi ,
I plan on using the Aperture contour that you had suggested a while back when we were first coming up with this idea of using a K-slot on this box and calling this thing a Karflex ...
I call it option #1
I also plan on stuffing both the S1 section and the stub/cavity section, fully, but not packed too tight (to avoid compressing the fibers excessively) ...
Freddi ,
I plan on using the Aperture contour that you had suggested a while back when we were first coming up with this idea of using a K-slot on this box and calling this thing a Karflex ...
I call it option #1
An externally hosted image should be here but it was not working when we last tested it.
I also plan on stuffing both the S1 section and the stub/cavity section, fully, but not packed too tight (to avoid compressing the fibers excessively) ...
Last edited:
Hi Matthew and xrk971,
I second the motion on building this.
Question:
With the radiators located @ x=0 y="some value" and z=0 shouldn't the driver location also be defined beyond the Node # w/ x=0 y="some value" and z="some negative value (reflecting the location of the center of the cone?)"? It should only affect the higher frequencies, but that's what Matthew is shooting for here?
Regards,
I second the motion on building this.
Question:
With the radiators located @ x=0 y="some value" and z=0 shouldn't the driver location also be defined beyond the Node # w/ x=0 y="some value" and z="some negative value (reflecting the location of the center of the cone?)"? It should only affect the higher frequencies, but that's what Matthew is shooting for here?
Regards,
Tinkering with K-slot , front chamber, radiators and aperture configuration
Tb46,
I think thats is a good question, and worth looking into ....
Hmmm ......
.... Ok , I am examining the scripts (both the Freddi and Tb influenced Karlflex scripts that i posted recently use the same K-slot aperture) .... Node 18 is the front of the cone but i didn't assign any radiator directly to the front of the cone, instead I chose to assign the radiator to the K-slot opening that is placed out in front of the cone which is node 28 (which is also connected to all of the other parts of the K-slot through node 18) ....
Since you asked about them i did some experimenting with the Z coordinates .... When i move the K-slot out in front of the baffle (using Z= negative centimeters) the changes in response are very minor , barely noticeable ........... If i then create an additional radiator and assign it directly to the front of the driver's cone (node 18) then the upper response does change drastically (above 500hz) , still useable i think , but definitely different ........ I am just not sure if this would be a proper model, because it is essentially one radiator (the K-slot) placed a few inches in front of another radiator (the driver) which must cause higher frequency phase cancellations (destructive interference or comb filtering, or at least a null or a few?) due to the relative delay of one radiator causing two different arrival times for the listener in this simulation ....
How does XRK reconcile this in his K-slotted creations? Does he assign a radiator to the driver or just assign them to the Aperture/Slot parts to act as the radiators?
Mr XRK , WHAT SAY YE?
Hi Matthew and xrk971,
I second the motion on building this.
Question:
With the radiators located @ x=0 y="some value" and z=0 shouldn't the driver location also be defined beyond the Node # w/ x=0 y="some value" and z="some negative value (reflecting the location of the center of the cone?)"? It should only affect the higher frequencies, but that's what Matthew is shooting for here?
Regards,
Tb46,
I think thats is a good question, and worth looking into ....
Hmmm ......
.... Ok , I am examining the scripts (both the Freddi and Tb influenced Karlflex scripts that i posted recently use the same K-slot aperture) .... Node 18 is the front of the cone but i didn't assign any radiator directly to the front of the cone, instead I chose to assign the radiator to the K-slot opening that is placed out in front of the cone which is node 28 (which is also connected to all of the other parts of the K-slot through node 18) ....Since you asked about them i did some experimenting with the Z coordinates .... When i move the K-slot out in front of the baffle (using Z= negative centimeters) the changes in response are very minor , barely noticeable ........... If i then create an additional radiator and assign it directly to the front of the driver's cone (node 18) then the upper response does change drastically (above 500hz) , still useable i think , but definitely different ........ I am just not sure if this would be a proper model, because it is essentially one radiator (the K-slot) placed a few inches in front of another radiator (the driver) which must cause higher frequency phase cancellations (destructive interference or comb filtering, or at least a null or a few?) due to the relative delay of one radiator causing two different arrival times for the listener in this simulation ....
How does XRK reconcile this in his K-slotted creations? Does he assign a radiator to the driver or just assign them to the Aperture/Slot parts to act as the radiators?
Mr XRK , WHAT SAY YE?
Intrigue!
Freddi , Moray ,
I have not seen this design before ....Cool stuff!.... It looks multi-resonant, like a back loaded horn with driver offset and a split path, the second path being through those two chambers with constrictions and a port/vent (making that second path low pass filtered to avoid any phase coherence issues up higher in frequency)... In some aspects it is similar to the Karflex with labyrinthine parasitic chamber that i built, maybe i will try to adapt that method in Moray's Lowther sketch over to a Karlflex to see if it is a good way to go ... Does Moray know how far apart the main tunings are spaced on that box?
I have seen a Fostex back loaded horn design that had two different tunings .... The 100hz horn section was tuned almost one octave above the 55hz reflex port ... So the tunings are spaced almost exactly like they would be in a classic Augsperger-Weems DCR/ABC design , just using different "elements" (exchanging one helmholtz chamber/port for a horn).... I am wondering if the resonance spacing is the same with the Lowther design?
Here is the Fostex box:
hi MMJ - Moray wanted you to see this Lowther design
Freddi , Moray ,
I have not seen this design before ....Cool stuff!.... It looks multi-resonant, like a back loaded horn with driver offset and a split path, the second path being through those two chambers with constrictions and a port/vent (making that second path low pass filtered to avoid any phase coherence issues up higher in frequency)... In some aspects it is similar to the Karflex with labyrinthine parasitic chamber that i built, maybe i will try to adapt that method in Moray's Lowther sketch over to a Karlflex to see if it is a good way to go ... Does Moray know how far apart the main tunings are spaced on that box?
I have seen a Fostex back loaded horn design that had two different tunings .... The 100hz horn section was tuned almost one octave above the 55hz reflex port ... So the tunings are spaced almost exactly like they would be in a classic Augsperger-Weems DCR/ABC design , just using different "elements" (exchanging one helmholtz chamber/port for a horn).... I am wondering if the resonance spacing is the same with the Lowther design?
Here is the Fostex box:
Last edited:
Moray James,
Ok , no problem ...
Considering all of the great updates that David has been making with Hornresponse lately it would not surprise me at all if we could model that Lowther cabinet using HR in the near future (with some new release), but in the meantime Akabak software is the solution for simulating these esoteric designs ..
I may try to adapt that Lowther concept (split path with dual helmholtz chambers in series sort of arrangement) into one of my Akabak scripts to see if i can determine what the Lowther engineer was trying to accomplish with that box...
Some notes about parasitic chambers , software and excursion
By the way guys,
Speaking of chambers with only one vent (like in the Clements patent) , "AJHorn" software has this chamber built in as a feature, and they call it an "absorber chamber" ....
Also, regarding the H-pas/Clements stuff i tried simulating it as shown in the patent sketches (not much driver offset to speak of) and it honestly did not look that great with the PA-310 until i added a significant amount of driver offset then it seemed to come together for me in Akabak , although as it turns out the (virtual) reduction in excursion (in simulation) wasn't very significant, only about one-half of one millimeter at most when using their special chamber (roughly the same small reduction as i was getting from the Augsperger-Weems configuration) which makes me wonder about Akabak's methods for calculating excursion (however , compared to Hornresponse sims of the same tuning but without the parasitic chambers there was a much larger excursion discrepancy at the same voltage drive .... HR was reading multiple millimeters higher)........hmmm..... *raised eyebrow*....... I cannot help but feel that something isn't quite adding up here .....
The H-pas guys (Solus/Clements/Atlantic?) chose to use a modified version of Martin J King's spreadsheets for their modeling needs instead of Akabak, and is it possible that Akabak simply wasn't giving them the proper results? *shrug* ............
Some DIYaudio guys recently compared their real world excursion measurements taken from a Dayton PA-385S-8 to the simulated HR excursion figures for the same driver, and it really calls into question software's ability to accurately predict excursion, perhaps there are just more factors to take into account that the software does not have at it's disposal (like BL curve or dynamic changes in suspension or something else........ Things we probably wouldn't want to bother inputting, too much hassle, so maybe it is just better to keep it simple and deal with some inaccuracy right?)..
Here is the comparison that they came up with:
By the way guys,
Speaking of chambers with only one vent (like in the Clements patent) , "AJHorn" software has this chamber built in as a feature, and they call it an "absorber chamber" ....
Also, regarding the H-pas/Clements stuff i tried simulating it as shown in the patent sketches (not much driver offset to speak of) and it honestly did not look that great with the PA-310 until i added a significant amount of driver offset then it seemed to come together for me in Akabak , although as it turns out the (virtual) reduction in excursion (in simulation) wasn't very significant, only about one-half of one millimeter at most when using their special chamber (roughly the same small reduction as i was getting from the Augsperger-Weems configuration) which makes me wonder about Akabak's methods for calculating excursion (however , compared to Hornresponse sims of the same tuning but without the parasitic chambers there was a much larger excursion discrepancy at the same voltage drive .... HR was reading multiple millimeters higher)........hmmm..... *raised eyebrow*....... I cannot help but feel that something isn't quite adding up here .....
The H-pas guys (Solus/Clements/Atlantic?) chose to use a modified version of Martin J King's spreadsheets for their modeling needs instead of Akabak, and is it possible that Akabak simply wasn't giving them the proper results? *shrug* ............
Some DIYaudio guys recently compared their real world excursion measurements taken from a Dayton PA-385S-8 to the simulated HR excursion figures for the same driver, and it really calls into question software's ability to accurately predict excursion, perhaps there are just more factors to take into account that the software does not have at it's disposal (like BL curve or dynamic changes in suspension or something else........ Things we probably wouldn't want to bother inputting, too much hassle, so maybe it is just better to keep it simple and deal with some inaccuracy right?
)..Here is the comparison that they came up with:
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
I feel like I must have modeled that Lowther hybrid horn - looks very familiar but can't seem to find the sim results. I think that dual reflex chamber helps to tame the first resonant peak from the horn and maybe also fill in the floor bounce dip typically that hits at 80Hz and 150Hz. Can't be sure though.
Technically called an acoustic band stop filter after its electrical equivalent and defined as a Helmholtz resonator, so it's of course easily converted to a closed pipe [1/4 WL TL] of the same alignment: Acoustics/Filter Design and Implementation - Wikibooks, open books for an open world
GM
Moray , Freddi , Xrk .......
XRK ,
I think you are totally right about what it does ...... I stripped down one of the Karlflex scripts (eliminated the aperture/slot and the Driver offset, basically turning it into an ML-TL) and then i added that series Helmholtz Lowther hybrid network (fashioned out of duct elements) and i can see how it fully tames the first peak/ripple (3rd harmonic region) pipe resonances that would normally be located around 180hz in this simulation because the mass loading pushes that third harmonic up a little higher than a plain straight pipe or pipe with expansion (This sim was tuned to a 40hz fundamental) ....... The higher harmonics (what i am guessing are the 5th and 7th etc) are still there causing peaks & ripple but the 3rd looks effectively suppressed with a very small hiccup in response around 125hz which must be some resonant effect created by the network ............ This Hybrid Lowther arrangement seems to be as effective as using a 1/3rd driver offset in a TL or QWP.
Interesting alternative
XRK ,
I think you are totally right about what it does ...... I stripped down one of the Karlflex scripts (eliminated the aperture/slot and the Driver offset, basically turning it into an ML-TL) and then i added that series Helmholtz Lowther hybrid network (fashioned out of duct elements) and i can see how it fully tames the first peak/ripple (3rd harmonic region) pipe resonances that would normally be located around 180hz in this simulation because the mass loading pushes that third harmonic up a little higher than a plain straight pipe or pipe with expansion (This sim was tuned to a 40hz fundamental) ....... The higher harmonics (what i am guessing are the 5th and 7th etc) are still there causing peaks & ripple but the 3rd looks effectively suppressed with a very small hiccup in response around 125hz which must be some resonant effect created by the network ............ This Hybrid Lowther arrangement seems to be as effective as using a 1/3rd driver offset in a TL or QWP.
Interesting alternative
Ahh ok, thank you, so an acoustic band stop filter ..... Very effective devices
and yes, easily created as a Helmholtz trap or 1/4 wave stub/cavity or pipe and according to Akabak sim either style can work well if the tuning is reasonably close to where it is needed (luckily it doesn't seem to require absolute precision, the tuned absorption has some bandwidth to it)..
Last edited:
FREDDI , Kappa12a
Freddi ,
You had asked about modeling a Kappa12a in a small simple Karlflex....
I tried the Kappa in a 60 liter Karlflex (70L-ish with Fredditech cavity) tuned to 50hz (main vent is shortened to 34cm) .... The driver's specs and parameters are not ideal for this cabinet so it won't take much power before running out of xmax , only about 20v, although it is fairly efficient and the cabinet is compact for sure ......
We get about 112db at 50hz and 114db at 60hz , pretty smooth rise up to 400hz at 117db , a peak at 600hz then it wiggles out to almost 1.5k before nosediving ....
Freddi , if you would like the script for this small Kappa12a Karlflex just let me know and i will post it ... .
Ideally , for better performance the driver would have a higher QTS, lower VAS , stronger motor and more XMAX ...
Freddi ,
You had asked about modeling a Kappa12a in a small simple Karlflex....
I tried the Kappa in a 60 liter Karlflex (70L-ish with Fredditech cavity
) tuned to 50hz (main vent is shortened to 34cm) .... The driver's specs and parameters are not ideal for this cabinet so it won't take much power before running out of xmax , only about 20v, although it is fairly efficient and the cabinet is compact for sure ......We get about 112db at 50hz and 114db at 60hz , pretty smooth rise up to 400hz at 117db , a peak at 600hz then it wiggles out to almost 1.5k before nosediving ....
An externally hosted image should be here but it was not working when we last tested it.
Freddi , if you would like the script for this small Kappa12a Karlflex just let me know and i will post it ... .
Ideally , for better performance the driver would have a higher QTS, lower VAS , stronger motor and more XMAX ...
Last edited:
I see that the tests were done at 54.6 volts and 76 volts, "using a dot on the cone and persistence of vision to measure peak to peak excursion". Careful measurements done by Bjørn Kolbrek some time ago showed that the Hornresp predictions of diaphragm displacement are quite accurate, provided that the driver is operated in its linear range and a sinusoidal input signal is used: http://www.diyaudio.com/forums/subwoofers/119854-hornresp-15.html#post1591527
As indicated in the Hornresp Help file, behaviour at all input voltage levels is assumed to be linear, and no allowance is made for low frequency high power amplitude compression.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.