AMT sub - using dynamic woofer

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
As some already may know, I love presentation of upper frequency department by AMT in dipole configuration (in particular attached to some diffraction alignment device :) ).

Looking for a solution to do a sub in dipole configuration, it seemed logically to give AMT principle a try here too.
The main disadvantage of dipole operation at the bottom 2-3 octaves is its very low SPL output – which is even a bigger problem if we seek after small footprint too.
One way to counter „natural“ SPL loss towards lower frequencies is to increase power invested and also to increase on moving diaphragm area.

Now – „AMT sub“ isn't any new – though probably not widely recognized as such.
Basically any Ripol, N-sub or W-sub utilizes AMT principle.

Despite being a total no-no from a lot of technical considerations as well as questionable in-room usability below modal region, those dipole sub's have gained some interest among hard core dipole lovers :D due to their mere sonic performance.

Michael
 
Standard Ripol / N-dipol / BMC usually looks like this:


Ripol.png

Ripol_back.png




Now an easy way to incerase Sd to push max-SPL is to simply stack them:


Ripol_side-stack.png


Above should have roughly +6dB of max-SPL compared to a single chassis Ripol and it already looks a little bit closer to AMT pleating



An alternative way to stack two woofers is to put them one behind the other :


Ripol_depth-stack.png



Above should also increase max-SPL by +6dB IMO but also shift low frequency corner of max-SPL down – which in fact would be very appreciated
Top frequency peaking would shift down too - giving lower usable bandwidth unfortunately.


My last config : “X-AMT” - to give it a spacy name :D


Ripol_stack-stack.png

Ripol_stack-stack_transp.png



IMO this kind of “X-AMT” configuration would again increase max SPL and also preserve down shift of max-SPL low frequency corner.
The middle two "pleats" are loaded by three times Sd this way.

:)

So, what do you think, elsewhere seen or experimented with ?

Michael
 
Last edited:
Analysing and measuring AMT subwoofers turns out to be a tricky game.



1.at low frequencies a room behaves “different”, as room modes rather dominate what we can get. On the other hand, a free field FR isn't all that meaningful either – not even to speak from measurements demands to get it.
2.with AMT subwoofer – as is with any framed or unframed open baffle construction – the FR depends on the proximity to the source. Very much like the proximity effect of directional microphones
3.in general *all* speaker constructions not acting as a single point source will show more or less pronounced CMP behaviour. AMT subwoofer – like any framed or unframed OB – being rather a two point source, clearly falls into this category, as do Horns, TL, TH...


Actually, even points 1.) and 2.) are a consequence of CMP behaviour in one way or another:

http://www.kinotechnik.edis.at/page...ine/CMP_Behaviour–Bridging_Points_of_View.pdf

So its not exactly tricky because AMT-subwoofer principle is any complicated to understand but rather because with constructions incorporating delay, the kinda universal tool of „frequency response concept“ gets flawed.

In short: our believe that “flat is flat” once we managed to tune a speaker to flat FR gets absolutely meaningless once CMP enters the picture:

Spectral distribution becomes a way more complex function of the *place* we stick the mic out and also being a function of the *time* we look at - than is properly reflected by simple “FR concept”.


All that said, we nevertheless can draw some *limited* conclusions looking at FR plots of H-baffles and AMT-subwoofer arrangements below.

I'll show some simulations plus “close to mouth” measurements.
Also shown are measurements that are taken close to mouth but are “dipole inverse-EQed” - meaning – the speakers are fed by a signal that accounts for the dipole-typical constructive and destructive comb filter FR.
This allows to take in-door measurements that reflect free space / far-field / non-proximity frequency response of dipoles (a very handy trick at those freezing temperatures out there right now :) ).
Dual-sub's are wired in parallel.


After this rather long wind up – lets jump right into a bunch of „really“ pretty pictures = a lot of heavily flawed FR plots.

:)
Michael
 
Last edited:
Below for reference a simple H-baffle subwoofer

Chassis is a 15” with Fs = 40Hz and Qts = 0.4
(Selenium 15PW5)


The H-frame subwoofer is roughly 40x40x80cm (WxHxD)


This is how it roughly looks like (top opened):


H-frame.png





This is how it measures:
RED Trace: close mouth
BLUE Trace: close mouth measurement “dipole inverse-EQed”
GEEN Trace: dipole comb filter

15PW5_H-frame_FR.png





and this is how AJ-Horn and Hornresp simulate:


H-baffle_Hornresp.png


H-frame_AJ-Horn.png




Michael
 
Next is a Dual-Ripol subwoofer

Chassis is 2x15” with Fs = 40Hz and Qts = 0.4
(Selenium 15PW5)


The Ripole-subwoofer is roughly 40x40x40cm (WxHxD) which is half the volume of the H-frame subwooofer housing two chassis.


This is how it roughly looks like (top opened):


Ripole_dual.png





This is how it measures:
RED Trace: close mouth
BLUE Trace: close mouth measurement “dipole inverse-EQed”
GEEN Trace: dipole comb filter

15PW5_Ripol_dual_FR.png



Michael
 
Next is a Dual-Ripol subwoofer with the chassis arranged one behind the other

Chassis is 2x15” with Fs = 40Hz and Qts = 0.4
(Selenium 15PW5)


This Ripole-subwoofer is roughly 30x40x80cm (WxHxD) which is about the volume of the H-frame subwooofer but housing two chassis instead.


This is how it roughly looks like (top opened):


Ripole_dual_extend.png





This is how it measures:
RED Trace: close mouth
BLUE Trace: close mouth measurement “dipole inverse-EQed”
GEEN Trace: dipole comb filter

15PW5_Ripol_dual_externd_FR.png



Michael
 
Last edited:
Last is a N-Frame subwoofer with the chassis arranged one behind the other

Chassis is 2x15” with Fs = 40Hz and Qts = 0.4
(Selenium 15PW5)


This N-frame subwoofer is roughly 40x40x80cm (WxHxD) which is about the volume of the H-frame subwooofer but housing two chassis instead.


This is how it roughly looks like (top opened):


N-frame_dual_extend.png





This is how it measures:
RED Trace: close mouth
BLUE Trace: close mouth measurement “dipole inverse-EQed”
GEEN Trace: dipole comb filter

15PW5_N-frame_dual_externd_FR.png



Michael
 
Blue curve?

Excellent work! One pretty picture equals a thousand words including the sketchup shots. They make the layout instantly understandable. Which of the graphs, red or blue, will be more representative of the anechoic far field response? How much does the dual ripol shake? Is the dipole distance "d" the same for a three port, vibration canceling dual ripol of the same footprint as the four port dual ripol?
 
Below for reference a simple H-baffle subwoofer

Chassis is a 15” with Fs = 40Hz and Qts = 0.4
(Selenium 15PW5)


The H-frame subwoofer is roughly 40x40x80cm (WxHxD)


This is how it roughly looks like (top opened):


H-frame.png





This is how it measures:
RED Trace: close mouth
BLUE Trace: close mouth measurement “dipole inverse-EQed”
GEEN Trace: dipole comb filter

15PW5_H-frame_FR.png





and this is how AJ-Horn and Hornresp simulate:


H-baffle_Hornresp.png


H-frame_AJ-Horn.png




Michael


I'm amazed your Dipole Peak is so so high when I calculate the peak for a 80cm H frame it is 160Hz yours appears to be over 200hz?

David
 
DFD / Deep Frame Dipole sub-woofer

DFD / Deep Frame Dipole sub-woofer



Summing on the effects of DFD / „Deep Frame Dipole“ sub-woofers there are obviously several mechanisms to observe.

Being a dipole loudspeaker, first to look at naturally is the typical dipole FR which basically is related to the separation distance.

In the context of aiming after high SPL output, the separation distance – in our case the depth of the „Deep Frame Dipole“ woofer - can and should be optimized with respect to the bandwidth of desire.
Meaning – for a given XO - we should put the first dipole peak slightly (½ to 1/1 octave possibly) above XO frequency, as this directly affects the attenuation at the lower end of the bandwidth related to the well known comb filter effect seen in the green traces.
Comparing the Ripole speakers – stacking side by side versus stacking one behind the other – shows most clearly what to gain.

Another effect entering the picture with DFD woofers is that we basically create sort of pipe / transmission line / quarter wave speaker as well.
Here a cute little trick comes in handy to upshift pipe-peaking.
Comparing the N-frame and the Ripol (depth-stacking) shows most clearly that with conical pipe we get the nasty peak out of bandwidth.
This is a nice benefit in balancing bandwidth restrictions due to dipole behaviour and due to pipe behaviour.

The last effect that is easy to observe is kind of suck out + some loading occurring with the Ripols.
Mainly the sensitivity at middle frequencies is lowered, giving a more flattish FR in this area, and also the low and lowest department is pushed a few dB.
Sadly this comes at the expense of heavy peaking. Unfortunately, my first guess that distributed arrangement of speakers in the pipe will prevent the pipe from peaking, did not hold


All in all - with respect to SPL maximization / volume minimization - we possibly can say that its best to make “Deep Frame Dipole” sub-woofer as deep as possible, keep the pipe in conical shape and stuff it with speakers.

:)

Michael
 
Last edited:
DFD / Deep Frame Dipole sub-woofer

Excellent work! One pretty picture equals a thousand words including the sketchup shots. They make the layout instantly understandable. Which of the graphs, red or blue, will be more representative of the anechoic far field response?

Thanx.
The blue trace is what you are looking for *if* you are interested in anechoic far field response – but – as said in the intro posting – both FR are kind of academic, considering a typical in room listening scenario.


How much does the dual ripol shake? Is the dipole distance "d" the same for a three port, vibration canceling dual ripol of the same footprint as the four port dual ripol?


These speakers were built first hand to get an impression of the effects that can be optimized.
As for shaking - you know I'm a fan of swinging speakers – I don't think its becoming a problem.


Is the dipole distance "d" the same for a three port, vibration canceling dual ripol of the same footprint as the four port dual ripol?

?? there is no three port speaker ??
The dipole distance is seen in the green traces and is also equivalent to the depth of the speakers.

All speakers have roughly 80cm separation, with the only exception of the ripole stacked side by side which is half of that: 40cm.


Michael
 
Last edited:
DFD / Deep Frame Dipole sub-woofer

I'm amazed your Dipole Peak is so so high when I calculate the peak for a 80cm H frame it is 160Hz yours appears to be over 200hz?

David


Well possibly I'm wrong - I calculated with 340m/s which gives 2.35ms time of flight for 0.8m which I dialed in and which equals roughly 210Hz ?

Anyway, the good thing on doing "comparative research" is that some 10-20% off do not affect validity of trends and conclusions to draw from.
:)

Michael
 
Last edited:
I calculated with 340m/s which gives 2.35ms time of flight for 0.8m which I dialed in and which equals roughly 210Hz ?

Michael,
as with BR pipes, the geometric length of a H or U frame needs to be corrected by an end factor resulting in the effective length. MJK assumes the effective length to be = geometric length + 0.6 x reffective. reffective would be the radius of a circle with the same opening area as the H or U frame.
U and H Frames page 3

Rudolf
 
Michael,
as with BR pipes, the geometric length of a H or U frame needs to be corrected by an end factor resulting in the effective length. MJK assumes the effective length to be = geometric length + 0.6 x reffective. reffective would be the radius of a circle with the same opening area as the H or U frame.
U and H Frames page 3

Rudolf

Thanx - and what would you calculate from that ?

Got confused for a moment ...

The green trace is the overlay of "inverse EQing" - this is not affected by any "opening factors" IMO (except I may add half of the mouth "diameter" to the path length probably).

Might be, far field is not "exactly" what the blue traces tell (its a work around after all) - but on the other hand - how *exact* do we need it finally in circumstances we never listen to ?

Might be - when temperatures are more friendly, I'll try to do a outdoor check, just to get a feeling ...


Michael
 
Conical

All in all - with respect to SPL maximization / volume minimization - we possibly can say that its best to make “Deep Frame Dipole” sub-woofer as deep as possible, keep the pipe in conical shape and stuff it with speakers.
Any ideas on the cause of the dip/peak at 160Hz of the N frame? Aside from the bump at 90Hz, the deep RiPol looks better up to 150Hz as the big peak at 280 could be easily notched with digital EQ. Would you expect increasing the dimensions slightly to allow a golden ratio of 80*50*31 for each side of the deep ripol to smooth the wiggles 85-210Hz? And maybe a slight taper, somewhere in between straight and full conical which in this N frame is currently about 1:2. Taper at 1:5? Magnets at the throat or the mouth? As you can see I am very interested in a RiPol layout for two AE IB15s per side to be used under a swinging Eminence Kappalite 3012 mid with DCX and/ or lin phase pc cross.
 
Lower

Michael,
as with BR pipes, the geometric length of a H or U frame needs to be corrected by an end factor resulting in the effective length. MJK assumes the effective length to be = geometric length + 0.6 x reffective. reffective would be the radius of a circle with the same opening area as the H or U frame.
U and H Frames page 3

Rudolf
So this is good news. The in room bass extension may be even lower than the current blue traces show.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.