Horn non-linearity frequency dependent

Got speaker theory questions? This is where you'll find the answers.
Post Reply
Message
Author
MissileCrisis
Posts: 1128
Joined: Sat Dec 15, 2012 1:30 pm
Location: Walnut Creek, California

Horn non-linearity frequency dependent

#1 Post by MissileCrisis »

Hello, so I have a horn subwoofer question that has recently come up due to some measurements I've made with a horn loaded subwoofer I designed. Basically, beyond a certain input threshold the response from 40-50 hz sucks out (compresses) almost completely, as in more input = same output. The enclosure corner frequency (length of horn) is approximately 38 and the pressure peak is centered at 40 Hz (which correlates to the maximum mouth velocity). At the level that it is compressing the pressure is relatively low (simulation says 4.6 kPa throat side, 2.5 kPa differential). Throat velocity is less than 7 m/s and exit around 4 m/s at this point. I'm at a loss as to what is causing this acoustic phenomena and was wondering if anyone else had encountered this. For the record the rest of the passband more or less doesn't compress all the way to full power, including the upper air velocity peak at 85 Hz.

Is there standing wave behavior in play and if so is there something I can do in the design to mitigate or avoid this problem? Also, any help into how I can simulate this phenomena would be appreciated. I'm considering learning how to input the design into ABEC to try and understand it better but that will take some time.

Edit: Note this cabinet has a compression ratio of 2.4~ and uses a 15 inch transducer in a front loaded horn configuration
Built
2 x TLAH
1 Titan 39 (lab 12, 15 inch)
1 Tuba 60 (lab 12, 20 inch)
SLA Center
2 x Mini MLTP subwoofers
1 x 212 TH (custom)
2 x Mini-Karlsonators
Owned, but not built
2 x Jack 12 (2512, melded)

sine143
Posts: 3066
Joined: Sat Feb 06, 2010 3:27 pm
Location: Raleigh NC

Re: Horn non-linearity frequency dependent

#2 Post by sine143 »

Driver, horn length (or hornresp inputs) would be helpful. Materials of construction and bracing amount would also be helpful. Voltage compression occurs would help.
Built:
2x Tuba 30s delta12lf loaded (gone)
4x Otop12 d2512 loaded
8x t48s (18, 18, 24, 24, 30, 30) 3015lf loaded
2x AT (1 mcm, 1 gto 804)
2x SLA Pro (dayton pa6, 6 goldwood piezo loaded)
1x bastard XF208

2x OT212 (delta pro 450a loaded, eminence psd)

User avatar
Bill Fitzmaurice
Site Admin
Posts: 28619
Joined: Tue May 02, 2006 5:59 pm

Re: Horn non-linearity frequency dependent

#3 Post by Bill Fitzmaurice »

MissileCrisis wrote:beyond a certain input threshold the response from 40-50 hz sucks out (compresses) almost completely, as in more input = same output.
The driver has reached xmax.

MissileCrisis
Posts: 1128
Joined: Sat Dec 15, 2012 1:30 pm
Location: Walnut Creek, California

Re: Horn non-linearity frequency dependent

#4 Post by MissileCrisis »

Driver, horn length (or hornresp inputs) would be helpful. Materials of construction and bracing amount would also be helpful. Voltage compression occurs would help.[/quote]
Baltic birch with as much if not more bracing as Bill recommends for his subwoofers (bracing spaced every 6 inches throughout the cabinet with crossbracing in mouth). The thing is vibrating, I guarantee that.

Driver excursion at the frequencies i'm talking about is close to a minima, and well within the linear region of the driver (15mm 10% BL distortion). Onset of compression happens around 50V, measurements were taken with highpass set as shown in sim. The driver can handle 110V thermal for 2 hr in free air, I'd suspect it will do 100V sustained in cabinet, excursion with highpass it should be fine to 120V.
Excursion 50V.PNG
[quote="sine143"]
Horn has two drivers, mirrored down the middle of the cabinet with seperate horn paths with both combining for the last flare (roughly last 3 feet).
Horn length 226.1cm or 89in or 7.42 feet.
Built
2 x TLAH
1 Titan 39 (lab 12, 15 inch)
1 Tuba 60 (lab 12, 20 inch)
SLA Center
2 x Mini MLTP subwoofers
1 x 212 TH (custom)
2 x Mini-Karlsonators
Owned, but not built
2 x Jack 12 (2512, melded)

MissileCrisis
Posts: 1128
Joined: Sat Dec 15, 2012 1:30 pm
Location: Walnut Creek, California

Re: Horn non-linearity frequency dependent

#5 Post by MissileCrisis »

Bill Fitzmaurice wrote:
MissileCrisis wrote:beyond a certain input threshold the response from 40-50 hz sucks out (compresses) almost completely, as in more input = same output.
The driver has reached xmax.
According to the sim that shouldn't be the case, not even close, excursion should be within klippel XMAX to full output.
Built
2 x TLAH
1 Titan 39 (lab 12, 15 inch)
1 Tuba 60 (lab 12, 20 inch)
SLA Center
2 x Mini MLTP subwoofers
1 x 212 TH (custom)
2 x Mini-Karlsonators
Owned, but not built
2 x Jack 12 (2512, melded)

User avatar
Bill Fitzmaurice
Site Admin
Posts: 28619
Joined: Tue May 02, 2006 5:59 pm

Re: Horn non-linearity frequency dependent

#6 Post by Bill Fitzmaurice »

You can believe the sim or you can believe the actual results. Not that the sim is necessary wrong, but if the data entry isn't right, or the translation from simulation to plywood isn't right, then the sim prediction will be off.

MissileCrisis
Posts: 1128
Joined: Sat Dec 15, 2012 1:30 pm
Location: Walnut Creek, California

Re: Horn non-linearity frequency dependent

#7 Post by MissileCrisis »

Bill Fitzmaurice wrote:You can believe the sim or you can believe the actual results. Not that the sim is necessary wrong, but if the data entry isn't right, or the translation from simulation to plywood isn't right, then the sim prediction will be off.
The measured impedance curve and frequency response (at 2.83V) match the simulation to within 1 dB and 1-2 Hz (each peak). I was quite deliberate in matching the enclosure design to the sim. Measured distortion seems to reflect excursion curve in that peak 3rd HD is at 33 Hz and mimimum 2nd HD is at 43.
Built
2 x TLAH
1 Titan 39 (lab 12, 15 inch)
1 Tuba 60 (lab 12, 20 inch)
SLA Center
2 x Mini MLTP subwoofers
1 x 212 TH (custom)
2 x Mini-Karlsonators
Owned, but not built
2 x Jack 12 (2512, melded)

User avatar
Radian
Posts: 2028
Joined: Tue Jul 29, 2008 12:56 pm
Location: Tucson, AZ

Re: Horn non-linearity frequency dependent

#8 Post by Radian »

MissileCrisis wrote: According to the sim that shouldn't be the case, not even close, excursion should be within klippel XMAX to full output.
Then what's the problem? If the sim has all the answers, what does the real world matter?
Work the problem. Keep investigating. Don't debate the data.
Good food, good people, good times.

4 - AT
1 - TT
1 - THT Slim
2 - SLA Pro 4x6 Alphalite

sine143
Posts: 3066
Joined: Sat Feb 06, 2010 3:27 pm
Location: Raleigh NC

Re: Horn non-linearity frequency dependent

#9 Post by sine143 »

post the impedance chart you measured.
Built:
2x Tuba 30s delta12lf loaded (gone)
4x Otop12 d2512 loaded
8x t48s (18, 18, 24, 24, 30, 30) 3015lf loaded
2x AT (1 mcm, 1 gto 804)
2x SLA Pro (dayton pa6, 6 goldwood piezo loaded)
1x bastard XF208

2x OT212 (delta pro 450a loaded, eminence psd)

MissileCrisis
Posts: 1128
Joined: Sat Dec 15, 2012 1:30 pm
Location: Walnut Creek, California

Re: Horn non-linearity frequency dependent

#10 Post by MissileCrisis »

Radian wrote:
MissileCrisis wrote: According to the sim that shouldn't be the case, not even close, excursion should be within klippel XMAX to full output.
Then what's the problem? If the sim has all the answers, what does the real world matter?
Work the problem. Keep investigating. Don't debate the data.
I know the excursion isn't the problem because it is reflected in the distortion (mimimum distortion point near minimum excursion point at ~43 hz). What I can't simulate is the non-linearity of the air and the increased mechanical load caused by the fold of the cabinet. I believe it to be a fluid dynamics problem. I did step up impedance measurements today and the notable effect of increasing input voltage is the lower resonance peak (35 Hz) decreases in magnitude progressively. the upper peaks remain basically the same and the impedance minimum rises and falls based on level (heating of the voicecoil). I'll grab the data tomorrow.

What I ultimately am trying to figure out is if there is any predictable amount of improvement I can achieve with increased volume (maintaining expansion rate)
Built
2 x TLAH
1 Titan 39 (lab 12, 15 inch)
1 Tuba 60 (lab 12, 20 inch)
SLA Center
2 x Mini MLTP subwoofers
1 x 212 TH (custom)
2 x Mini-Karlsonators
Owned, but not built
2 x Jack 12 (2512, melded)

sine143
Posts: 3066
Joined: Sat Feb 06, 2010 3:27 pm
Location: Raleigh NC

Re: Horn non-linearity frequency dependent

#11 Post by sine143 »

MissileCrisis wrote:
Radian wrote:
MissileCrisis wrote: According to the sim that shouldn't be the case, not even close, excursion should be within klippel XMAX to full output.
Then what's the problem? If the sim has all the answers, what does the real world matter?
Work the problem. Keep investigating. Don't debate the data.
I know the excursion isn't the problem because it is reflected in the distortion (mimimum distortion point near minimum excursion point at ~43 hz). What I can't simulate is the non-linearity of the air and the increased mechanical load caused by the fold of the cabinet. I believe it to be a fluid dynamics problem. I did step up impedance measurements today and the notable effect of increasing input voltage is the lower resonance peak (35 Hz) decreases in magnitude progressively. the upper peaks remain basically the same and the impedance minimum rises and falls based on level (heating of the voicecoil). I'll grab the data tomorrow.

What I ultimately am trying to figure out is if there is any predictable amount of improvement I can achieve with increased volume (maintaining expansion rate)
your previous graph displays excursion minima at 49hz. what is sd and mms of driver. full t/s parameters would possibly be helpful. driver may be unsuitable for horn loading. higher voltage distortion measurements may reveal the answer, but I would measure a 2.83v response curve, and another response curve at the highest voltage you are comfortable measuring at (reposition microphone at a distance that will maintain the same magnitude measured). graph the difference between these 2 measurements for your compression graph.
Built:
2x Tuba 30s delta12lf loaded (gone)
4x Otop12 d2512 loaded
8x t48s (18, 18, 24, 24, 30, 30) 3015lf loaded
2x AT (1 mcm, 1 gto 804)
2x SLA Pro (dayton pa6, 6 goldwood piezo loaded)
1x bastard XF208

2x OT212 (delta pro 450a loaded, eminence psd)

MissileCrisis
Posts: 1128
Joined: Sat Dec 15, 2012 1:30 pm
Location: Walnut Creek, California

Re: Horn non-linearity frequency dependent

#12 Post by MissileCrisis »

Edit: I've found a plan of attack for a redesign. I'm removing this data as a result. We can close this thread.
Built
2 x TLAH
1 Titan 39 (lab 12, 15 inch)
1 Tuba 60 (lab 12, 20 inch)
SLA Center
2 x Mini MLTP subwoofers
1 x 212 TH (custom)
2 x Mini-Karlsonators
Owned, but not built
2 x Jack 12 (2512, melded)

Post Reply