"Filler" driver ala B&O

[xrk971]

Maybe soemone would like to have a Bandpass which is 4th order, i.e. having 2nd order slopes. This could be achieved using a 4th order polynomal. This should be feasible in the analog domain and in the digital domain as well. The high and lowpass could even be pulled further apart if necessary.

I anyone is interested I will derive it.

Regards

Charles

Can it still sum up to be transient perfect with 4th order filters?

 

[phase_accurate]

Yes it will, but I am currently without simulator so I can't say what the cost would be.
You must know that it is possible to do transient-perfect crossovers with ANY desired order but they come at the cost of increased overlap and lobing. Your MTM driver arrangement will help a bit regarding the latter.

With the fourth order polynomal the lowpass and the highpass will each be 3rd order with a hump. And a passive version would be difficult.

Regards

Charles

 

[BYRTT]

xrk971 is it right if i sense there to be very much room ambience into the sound clips and in that track for me is unknown have no reference point to judge it.

You pointed to Econowave route, can I point to below thinking : )
I very much like the performance you often have shown with 10F/8424 so here another suggested scenario for this XO and the Eminence woofers running OB. Put 10F in center, above and below the 5" Faital Pro in below picture, and lastly place the two Eminence woofers symmetrical above and below MF/HF section. OB for woofers and your good Dagger or Nautaloss box at behind of baffle for MF and HF drivers. Going down to 500Hz (400-600) with XO point will azure CTC spacing is below 1/4 wave length as Barleywater often suggest is good. Think 10F and woofers will perform pretty well when cut LR2 at 500Hz, question is if suggested Faital can do it. It will not be a 20-20kHz but nice 10 octave 32-16kHz system.

 

[wesayso]

Can't we build that out with a pair of woofers below the eminence and a tweeter up top and use a pair of 10F's, in other words use 2 of these kind of filler drivers, the 10F as the upper pair and Eminence as the lower pair with some LF woofers below them.
WWMMTMMWW like . It needs more channels on the miniDSP though

Throw the tweeter in a waveguide, and some processing on top and you'll have a very nice expanding array with transient perfect response.

Not the first to think about this: http://www.htguide.com/forum/printthread.php?t=34839

 

[xrk971]

xrk971 is it right if i sense there to be very much room ambience into the sound clips and in that track for me is unknown have no reference point to judge it.

You pointed to Econowave route, can I point to below thinking : )
I very much like the performance you often have shown with 10F/8424 so here another suggested scenario for this XO and the Eminence woofers running OB. Put 10F in center, above and below the 5" Faital Pro in below picture, and lastly place the two Eminence woofers symmetrical above and below MF/HF section. OB for woofers and your good Dagger or Nautaloss box at behind of baffle for MF and HF drivers. Going down to 500Hz (400-600) with XO point will azure CTC spacing is below 1/4 wave length as Barleywater often suggest is good. Think 10F and woofers will perform pretty well when cut LR2 at 500Hz, question is if suggested Faital can do it. It will not be a 20-20kHz but nice 10 octave 32-16kHz system.

Yes, you are right about the ambiance, tons of it due to OB - stuff bouncing off back walls. But then it is like a band in your room - sound comes from 360 deg around them.

Are you suggesting a WMTMW with the 10F as the hole filler mid and the Faital as the tweeter? Is this to get lower extension on HF? I could use my 5MR450NDY as the tweeter as I have that already and it goes to 15kHz also.


I have one more piece of identical plywood. It would be easiest at this point to swap out the TG9's for the 10F's. I was thinking about getting the XT25 tweeter, but find I am not lacking highs with the Dayton soft dome - I actually very much like its sound. No ringing at all with all the goopy stuff they coat the silk dome with (it appears to be a viscous non Newtonian fluid applied that is permanently tacky (wet). Bad for lint and sawdust as they stick like crazy to the dome. I think for $20, you would be hard pressed to find a smoother sounding tweeter.

I could get an $11 waveguide for the PRV D280Ti CD and use that as tweeter. Dual 10F's. Now it is not such a cheap speaker anymore once the 10F's are thrown in.

 

[freddi]

which $11 waveguide? - I've played with the MCM version of this one a long time ago Denovo Audio DW-62S round waveguide DIY Sound Group

 

[wesayso]

10F as the tweeter is what I'm reading there, Faital Pro as the mid filler.

 

[Zero D]

@ xrk971

Yes i designed the BW1 active crossover with op amps.

The key thing to get right are the overlap factors and central frequency.

I spent some tweaking the values of components, until i acheived the flat response & zero phase shift etc.

Individual gains can easily be added.

Yes, U1 = W U2 = BP U3 = T

Yes it would be able to be simmed with TINA (spice) I've had FilterPro for some years, but since i got Multisim i havn't used it. Might dig it out & see.

However, if you want to build one for me and I will hook it up to my system, I am all for that

Ok, if you pay my return air fare

I designed with all BW1 to see if it could be done, when a lot of the talk was for BW2-BW1-BW2 filters etc. Also to try & acheive All in phase filters & drivers, transient-perfect, & linear this & that etc. In short the best of all worlds.

As i mentioned, you could change the f3's, for eg, W f3 = 1kHz, BP f3 = 2kHz, T f3 = 4kHz. This might suit the appropriate drivers better.

@ phase_accurate

Your 4th order idea might be interesting to compare, to see how linear etc etc it might be ?

 

[xrk971]

which $11 waveguide? - I've played with the MCM version of this one a long time ago Denovo Audio DW-62S round waveguide DIY Sound Group

Well, $11.99 to be exact...

PRV Audio WG16-25-B 1" 45 x 45 ABS Elliptical Tractrix Waveguide 2/3-Bolt

 

[diypass]

should be instresting posting links to pdf....web...and other sources to know better this type of solution ?
where could i find theory about it ?

 

[xrk971]

10F as the tweeter is what I'm reading there, Faital Pro as the mid filler.

Ok, so the 10F is a tweeter with super wide bandwidth, and the 5FE120 is a mid with super wide bandwidth and reaching into the bass, thus easier to get textbook slopes on all drivers.

Coincidentally, it turns out that the woofer from a coax is in general, designed to have pretty wide bandwidth to start with in order to xo with a CD around 2.5kHz. So I lucked out in that regard by using it an octave below its usual XO point.

That 5FE120 looks pretty nice. Hmm... I like the XT25 too - both about the same price.

 

[BYRTT]

Yes, you are right about the ambiance, tons of it due to OB - stuff bouncing off back walls. But then it is like a band in your room - sound comes from 360 deg around them......

Wov sounds nice, the closest i come to listen OB is playing thru Guitar tube amp and it has far from a full range audio bandwidth.

.....Are you suggesting a WMTMW with the 10F as the hole filler mid and the Faital as the tweeter? Is this to get lower extension on HF? I could use my 5MR450NDY as the tweeter as I have that already and it goes to 15kHz also......

As wesayso say i meant 10F as tweeter 2 x 5FE120 as mid fillers, thought was not to get lower extension on HF that just get a compromise to pay of this config but to open a wider window for vertical lobe by going lover XO point and 10F as tweeter can go down there and you shown plenty times it sound very good, also could TG9/TC9 be used as tweeter but 10F seems the cream of those three. Hope in the below pictures the way i use free XDir from tolvan data to sim vertical lobe at XO point is right. The first picture should be approximate your present setup and reveal small listeners vertical window to receive the nice SR from this XO type, in picture two a wider window is shown and XO point is reason.

Regarding 5MR450NDY and 15kHz instead of 10F then do you have 2 x PA130-8 laying on the shelf and can it go low enough, it will for sure play loud : )

 

[xrk971]

Byrtt,
You are a genius!

Regarding 5MR450NDY and 15kHz instead of 10F then do you have 2 x PA130-8 laying on the shelf and can it go low enough, it will for sure play loud : )

I have all of these - Good idea! Two PA130-8's in parallel will get me 94dB, a good match with 5MR450NDY at 95dB. Two independent Beta 12cx will get me 97dB + 3dB or 100dB and -5dB for losses due to baffle gets back to 95dB. I may just buy another pair of PA130-8's since I like them in the 0.53x Karlsonator's so much don't want to mess with them.

The system overall sensitivity is -1.5dB below sensitivity of the mids which if are 88dB+6dB=94dB-1.5dB=92.5dB
It will be a 92.5dB system! That will have some great dynamics. Probably run XO center freq at 500Hz with Gamma=2? So Beta 12cx LR2 LP is at 250Hz (np problem), 5MR450NDY LR2 HPF is at 1kHZ, and PA130-8 with 250Hz-1kHz BW1 bandpass needs to go from 125Hz to 2kHz smoothly - which it does.

Only problem is beaming with a 5in tweeter.

Or I can try two PA130-8's as the mids and keep the tweeter as is - seems to work well at 1.2kHz. Thus PA130-8's will give me the little extra mid range boost that I need.

What xo is used for above calc?

 

[BYRTT]

.....

What xo is used for above calc?

Thought was for that 10F / 5FE120 / Beta12cx combination Fx 500Hz with gamma 1,0 in WMTMW config.

Then it gets longhaired because the left one says 250Hz which is the real acoustic XO point between the LF and MF and right one says 1000Hz which is the real acoustic XO point between the MF and HF where center for filler is 500Hz.

 

[xrk971]

Thought was for that 10F / 5FE120 / Beta12cx combination Fx 500Hz with gamma 1,0 in WMTMW config.

Then it gets longhaired because the left one says 250Hz which is the real acoustic XO point between the LF and MF and right one says 1000Hz which is the real acoustic XO point between the MF and HF where center for filler is 500Hz.

Thanks.

How ugly is it with my current MTMWW setup as described on post #1 with Fx of 1200Hz?

 

[BYRTT]

Thanks.

How ugly is it with my current MTMWW setup as described on post #1 with Fx of 1200Hz?

Its actual the first picture in post 112 set at 600/2400Hz acoustic, for the CTC distances took a guess and must be close enough. Think the LSR305 monitor you tested some time ago would have vertical lobe mid between the two shown at post 112. Having transient perfect acoustic lobe think will pay back when wide as possible versus horizont dispersion to battle it not just happens at one point in space.

 

[xrk971]

should be instresting posting links to pdf....web...and other sources to know better this type of solution ?
where could i find theory about it ?

The theory is in the first few pages
http://www.southampton.ac.uk/~apm3/diyaudio/Filler_drivers.html

John K's website with spreadsheet:
http://www.reocities.com/kreskovs/GenFiller.html

The B&O patent:
http://www.google.com/patents/US4031321

and the later links in this post
http://www.diyaudio.com/forums/multi-way/88135-filler-driver-ala-b-o-7.html#post4381538

 

[xrk971]

Sometimes Less is More

Well, I tried to figure out how to improve the step response. I have a sharp peak followed by a dip then the ramp. Changing the time delays did not help, and by looking at the SR from just the mid and tweeter, I could see the problem was that there was excessive ringing in the HF that was causing the dip. Removing the two PEQ's in the tweeter made the SR much better. The Frequency response of the tweeter is not as smooth now and has a downward tilt. But I could not hear the difference really in the lack of highs. The subjective sound overall though was improved as it sounded even more cohesive now. The guitar on SRV's Tin Pan Alley sounds especially good.

Here is the HF PEQ diagram - just a mild high shelf is left:

Here are the measured acoustic XO and response curves:

Here is the measured impulse response, note the reduction in ringing:

Here is the measured step response:

However, I tried the square wave at 1200Hz and it is not working out. Looks more like a triangle wave. Is this indicative of bad time alignment still? What am I doing wrong that it is not even close?

 

[BYRTT]

.....
However, I tried the square wave at 1200Hz and it is not working out. Looks more like a triangle wave. Is this indicative of bad time alignment still? What am I doing wrong that it is not even close?

Well think you on the road below is the old B&O review linked to at post 70.
Your 1200Hz is much looking like the 100Hz there, isn't this figure indicative we have more HP filter function than we liked to : )

 

[phase_accurate]

Thats quite strange because the step response indicates that it should be at least as good as the B&O example.

Can you do the same measurement outside ? As BYRTT indicates it looks as if we have a highpass function somewhere. But maybe it is something else like a superposition with delayed sound that is causing it.

Regards

Charles