A Bookshelf Multi-Way Point-Source Horn

[xrk971]

Weird when seen in light of for SB12NRX25-8 our sims were in agreement, then i only see reason one of us have wrong T/S parameters set for 6FE100 or accidently choose closely named 6FE200, but let it be SB23NRXS45-8 will do the job nicely and when lowered in system spl their 60 watt thermal power specs is not pressed.

Let me run it through my stand alone bandpass script to see if I have made a mistake somewhere. You are right, our sims should sort of agree and thanks for double checking. I am not error free!

 

[xrk971]

Just throwing an idea/question in here.

Is it possible to convert the back chamber into a folded TL and then gaining more low-end? And/Or then use 3"/4" full range drivers instead of a woofer. With this you could probably shift te ports closer to the cd mouth.

It's a valid and a good idea - but would add bulk volume to the box. Since Bushemeister here wants a "bookshelf", albeit he must have some very large books he reads I am trying to keep things as simple and small as possible.

For my smaller uTrynergy that uses four 5.25in woofers, they currently are in a semi dipole open back enclosure. The bass actually reaches an amazingly low level of mid 30Hz. It is typical dry OB bass that doesn't pressurize the room. I was thinking of adding on a TL tail to the box - a stuffed and tapered one and then snake the output back up to the periphery of the horn mouth. It would make the box larger and the new TL would basically be able to turn it from a stand mount to a free standing tower speaker.

The location of the bandpass injection ports do not need to be so deep because this is a FAST concept with a low XO frequency. If the mid to tweet XO frequency was in the 1.2kHz range as is usual with most synergies, then yes indeed locating the injection ports as close to the throat as possible moves the upper frequency cancellation notch higher above the XO frequency. For a 500Hz XO frequency the current setup is predicted to have the first notch well above around 1kHz.

 

[xrk971]

I love this guys!

I go to bed, wake up and you have done amazing work!

So, conclusions:
1. keep SBA 8 inchers.
2. Ports closer to front much better from diffraction viewpoint and minimal effect on tuning.
3. Port size a real compromise between SPL, mechanical port noise, and diffraction effects- likely the Goldilocks's port to be chosen- not too big, not too small!

Ok, after work I will start cutting some white paper port outlines and take some photos on the horn so we can see what looks best.

Thanks again.

You are welcome! Byrtt has been very helpful and keeps me on my toes. Let me double check the Faital Pro 6FE100 again. But you can't go wrong with the SB23's.

 

[bushmeister]

Good - They are on their way!

Also apparently the SB65s have landed in the UK - so hopefully we will get some hard measurements soon.

Regarding box alignments - it looks like for a normal ported enclosure these woofers would be best in about 90L each volume (for an F3 of around 30 hz) so that would be a massive box for the two of them.

For sealed enclosure - 45L per woofer would give a 0.7 system Q with F3 of about 50Hz. So again - a large enclosure.

I guess I could try the woofers dipole - Much like Xrk has done and see - I will be taking many measurements as we go along anyway, so plans may change as we get more hard data!

 

[BYRTT]

.....I will be taking many measurements as we go along anyway, so plans may change as we get more hard data!

Good plan for control look forward first run of data SB65WBAC25-4 coupled XT1464.

Picture 1 is xrk971's sketch. For fun edited that sketch to band pass forward fire ring port for woofer, detail is something we talked about some time ago. Band pass port is not there to replace electric XO but to bring down ctc distance to less than real center of woofer and improve XO area vertical power response.

 

[xrk971]

Good - They are on their way!

Also apparently the SB65s have landed in the UK - so hopefully we will get some hard measurements soon.

Regarding box alignments - it looks like for a normal ported enclosure these woofers would be best in about 90L each volume (for an F3 of around 30 hz) so that would be a massive box for the two of them.

For sealed enclosure - 45L per woofer would give a 0.7 system Q with F3 of about 50Hz. So again - a large enclosure.

I guess I could try the woofers dipole - Much like Xrk has done and see - I will be taking many measurements as we go along anyway, so plans may change as we get more hard data!

Yes, those woofers have large box alignments to get them to their true potential. However, we are using them to "squeeze" bass out of a couple of little holes and all that direct radiator stuff for traditional box volumes goes out the window. The Akabak model handles all that and says it will make 82Hz f3 with 16 liters for both. That is still within realm of a "bookshelf". You might try and open back box like I did and the bass will probably reach an impressive 35Hz I expect. But there is no cone control and Max SPL will be limited. It may be fine for 90dB SPL though as I found out and can be quite pleasant.

 

[xrk971]

Good plan for control look forward first run of data SB65WBAC25-4 coupled XT1464.

Picture 1 is xrk971's sketch. For fun edited that sketch to band pass forward fire ring port for woofer, detail is something we talked about some time ago. Band pass port is not there to replace electric XO but to bring down ctc distance to less than real center of woofer and improve XO area vertical power response.

Your artistic skills are pretty good Byrtt. You should make more drawings like this

 

[bushmeister]

I thought the 'normal' box calculations were probably pointless!

I suspect we will end up with more than 16 litres in the 'box' unless I slope the walls back to make it wedge shaped like the horn. How sensitive are the alignments to the sealed back volumes?

Would it matter if hugely if they were in 10 litres per woofer or 15 litres?

Byrtt - I agree with Xrk - great drawing skills! I am a little intimidated by the beautiful curved forms you guys are outlining for the enclosure though - I am gonna have to get handy with fiberglass and epoxy by the looks of it!!

I suspect my efforts will be a little more drab....

 

[USRFobiwan]

" am a little intimidated by the beautiful curved forms"

why not K.I.S. first and look at shapes later.

 

[xrk971]

I thought the 'normal' box calculations were probably pointless!

I suspect we will end up with more than 16 litres in the 'box' unless I slope the walls back to make it wedge shaped like the horn. How sensitive are the alignments to the sealed back volumes?

Would it matter if hugely if they were in 10 litres per woofer or 15 litres?

Byrtt - I agree with Xrk - great drawing skills! I am a little intimidated by the beautiful curved forms you guys are outlining for the enclosure though - I am gonna have to get handy with fiberglass and epoxy by the looks of it!!

I suspect my efforts will be a little more drab....

Prismatic boxes are fine - use a few angled walls to reduce reflections of possible but egg crate foam and fiberglass work wonders on straight walls.

We have to go crazy and dream of curved surfaces. Foam core can do that curved stuff just fine. Look at bendy plywood or bendy MDF.

Larger volume is Ok - it will reduce Max SPL. So if you are aiming at hitting 113.6dB Max prediction - then add volume filler blocks (bricks, wood blocks, rocks, etc).

 

[xrk971]

" am a little intimidated by the beautiful curved forms"

why not K.I.S. first and look at shapes later.

Mainly because it looks cool but it may be needed to keep volume down if you want to get close to 16 liter target volume. A rectangular prism is full of volume.

 

[BYRTT]

Your artistic skills are pretty good Byrtt. You should make more drawings like this

Have good teacher to learn from : )

 

[xrk971]

Weird when seen in light of for SB12NRX25-8 our sims were in agreement, then i only see reason one of us have wrong T/S parameters set for 6FE100 or accidently choose closely named 6FE200, but let it be SB23NRXS45-8 will do the job nicely and when lowered in system spl their 60 watt thermal power specs is not pressed.

TS parameters used were correct:

Def_Driver '6FE100-8' | 6in woofer 5.25mm xmax, 100w
fs=61Hz
Re=5.4ohm
Le=0.5mH
Qms=6.0
Qes=0.60
Vas=14.1L
Bl=6.8Tm
Mms=14g
Sd=143cm2

So I went and sim'd a plain bandpass box with 8 liter rear chamber and 165cc front chamber and 65mm dia x 30mm long duct and get a pretty flat response. Although the f3 is now 91Hz. This is with 65Hz BW2 high pass filter so that should not be affecting low end extension.

So I will have to see why it sags when put in the xBush horn.

 

[BYRTT]

Thanks nice piece of work configuration and results there looks right.

.....This is with 65Hz BW2 high pass filter so that should not be affecting low end extension.....

Think all EQ and filters sometimes affect more than we like : (

Below example grey trace is BW2 HP 80Hz and same as acoustic output if a 6FE100 live in a Q 0,7 sealed box.
Orange trace is BW2 HP 65Hz filter.
Black trace is sum when they cascaded and looks close to what shown in Akabak.
Think for bushmeister's build target is LR4 HP 80Hz to integrate with sub so show that LR4 HP filter in red trace, cursor at 80Hz -6dB.

Tried with more precision because when 6FE100 live in 8 liter sealed volume its actual Q0,9/101Hz as in grey trace.
Orange trace is BW2 HP 65Hz filter.
Black trace is sum when they cascaded and looks as Akabak sim.
Black trace is close to red trace target LR4 HP 80Hz to integrate with sub, looks as one strategic right set PEQ would correct black trace to same as red.

.....So I will have to see why it sags when put in the xBush horn.

One comes to think about interference pattern because woofers radiate some against each other verse if side by side on a plane baffle radiate in same direction.


Picture 3 is Winspeakerz 64 watt into 6FE100, one plot is sealed 8 liter the other is symmetric band pass 8 liter back box 0,165 liter front box. Goes lower than Akabak because lack of BW2 65Hz HP filter.

 

[xrk971]

I did take another look in Akabak and it still slopes down on left like previous sim. I think what is happening is that the WG is providing a mild gain to mids above 200Hz to the right and this makes it look non-flat as if it is the woofer's fault. It's just that the bass over shoot on the Faital is less than the SB23 that it looks that way. I could not find any typo in the script. So probably as you say, a little PEQ can help and probanly either driver can work with the Faital more suited for a flatter Q. The SB23 just actually has more self induced bass bump which in the end - works out flatter acoustically.

So for Bushmeister, if you have the SB23's on the way, keep them and remember they are thermal limited to 65watts x 2 = 170 watts. They should work fine.

If you wanted to switch to 6FE100's that can work too, and that handles more power at 200watts.

However if you don't play loud continuously like a pro sound speaker, the thermal limit is not an issue.

6FE100's also cost less and are smaller so probably easier to integrate.

 

[bushmeister]

SB23's should arrive in 2/7.
No sign of SB65's yet.....I think they may have got held up in customs...

 

[xrk971]

Bushmeister - man, stuff takes forever to get to Middle Earth Hobbit land! I would go nuts waiting that long. How long does it take for Faital Pro's to come?

You need to check on customs with that driver. They are tiny packages - should fly right through one would think.

 

[xrk971]

xBUSH v1.1 with 6FE100-8

I wanted to look at all the other aspects of using the 6FE100's to see where there are going to be differences. The biggest thing with the 6FE100's is that they are smaller, so the displacement for given stroke is less and this means that they will be *displacement xmax* limited at max SPL vs the SB23's which will be thermally limited before hitting their xmax. The 6FE100's will be more sensitive though. The f3 of the 6FE100's in this system will be in the 90Hz range rather than 80Hz range for the SB23's.

Here are the results for the predictions with a qnty 2 per driver x 50mm dia x 12m long injection port (making it 12mm vs 1mm gives a little more bass extension). 65Hz BW2 HPF to protect over excursion and cross to sub.

Max SPL at 5.25mm xmax:

Cone displacement at xmax and 18.5vL

Velocity through qnty 4 x 50mm injection ports:

Electrical power for 6FE100's and for SB65WB:

Electrical Impedance for both drivers:

System Sensitivity at 2.83v:

Here is predicted Impulse Response - looks pretty clean ( assuming the SB65WB is well behaved in WG):

Here is predicted Group Delay without any EQ which should flatten it more, and once the sub is integrated the GD rise in the bass should be reduced. It shows that the speaker has a native time delay of about 2ms (also shows up in location of peak for IR), but in general, the GD is flat and will provide nice tracking of frequencies over a broad range with less time smearing.

 

[xrk971]

Same sim with SB23's

Just for comparison with the above sims for 6FE100's. Here is same sim but for the SB23's in the same setup with 50mm dia x 12mm long ports, 65Hz HPF, and now thermally limited to 170 watts or 22.8v.

Cone displacement at thermal max:

Velocity through 50mm ports at thermal max is high and will chuff. Probably need to make bigger:

Electrical Power Input (170 watts max):

Predicted electrical impedence:

Sensitivity at 2.83v:

 

[xrk971]

So why go through all this trouble to make point source horn besides it has better directivity control and seamless multi-way integration?

Look at the cone displacement at a typical listening level of say 94dB (2.83v) - quite loud actually. The horn loading does a great job of reducing excursion and cone moves 100 microns, while woofers (although not horn loaded) are also only doing 0.75mm of movement. These are all at a small fraction of their linear xmax cone travel so the distortion that normally comes from cone motion is greatly reduced. It should sould super clean at 94dB, and in fact will sound so clean you think it is not as loud and will keep turning it up.