[matevana]

As promised, here's an update to post #19 with final crossover values. It was necessary to wait almost 3 months for the drivers to settle-in.   

The value changes are slight and the variable potentiometer has been updated with a fixed resister value for best fidelity.

Low/Mid:
Dayton 10" Pro Sound Driver 8 ohms
L1 = 1.80 mh
C1 = 33 uf
R1 = 7 ohms

Mid Coupler:
Celestion 5" Neo 8 ohms
L2 = .22 mh (18 ga)
C2 = 50 uf
P1 = 8 ohms (or variable 8 ohm potentiometer)

High Frequency:
Vifa XT19TD00-04 4 ohms
C3 = 6.8 uf
R2 = 1.50 ohms

[sjhomey]

Your Shopping Cart
     
   Dayton Audio DMPC-6.8 6.8uF 250V Polypropylene Capacitor
Part Number: 027-424 In Stock$2.93    $5.86    
   L-Pad 15W Mono 1" Shaft 8 Ohm
Part Number: 260-250 In Stock

      $6.75    $13.50    
   Dayton Audio DNR-1.5 1.5 Ohm 10W Precision Audio Grade Resis
Part Number: 004-1.5 In Stock

      $1.25    $2.50    
   Dayton Audio DNR-7.0 7 Ohm 10W Precision Audio Grade Resisto
Part Number: 004-7 In Stock

      $1.25    $2.50    
   Jantzen 0.33mH 18 AWG Air Core Inductor
Part Number: 255-220 In Stock

      $5.26    $10.52    
   Jantzen 1.8mH 18 AWG Air Core Inductor
Part Number: 255-264 In Stock

      $11.79    $23.58    
   Dayton Audio DMPC-50 50uF 250V Polypropylene Capacitor
Part Number: 027-443 In Stock

      $13.80    $27.60    
   Dayton Audio DMPC-33 33uF 250V Polypropylene Capacitor
Part Number: 027-441 In Stock

      $10.14    $20.28    
45 Day No Hassle Return Policy Order Subtotal:    $106.34
   
   

[sjhomey]

Ed
Somehow I managed to post the above list before I was able to edit it. It is a list of what I think I need in crossover parts. Hopefully it is legible. If not I can repost. Am I close?
Thanks

Peter

[matevana]

Ed
Somehow I managed to post the above list before I was able to edit it. It is a list of what I think I need in crossover parts. Hopefully it is legible. If not I can repost. Am I close?
Thanks

Peter

Exciting!  Someone's got an SL in their future  

I think I can save you a few bucks. I ended up using the Dayton 100V Non-polarized caps for the 33uf and 50uf applications. In this particular config I could not hear a difference.

027-350  $.98 cents each (33 uf)
027-354  $1.11 each  (50 uf)

Also, I used the 8 ohm pot on the Celestion Neo temporally until the speaker was broken-in and then settled on a fixed 8 ohm resistor as the best overall permanent value. You can take advantage of this knowledge and go straight to the 8 ohm resistor in lieu of the potentiometer, while saving yourself a few more dollars and stepping up the SQ slightly. The pot allowed me to dial-in the amount of attenuation while I was tweaking the x/o.

Hope that helps!

Ed

[sjhomey]

Hi Ed

Thanks for the heads up. I just placed an order with the electrolytics. I read a lot of conflicting opinions regarding the them. What sticks in my mind most is that the tolerances can be way off. Do you have any qualms about using them?

Also, is the layout using the 8 ohm resistor (R3?) (+ _ L2 _ C2 _ R3 +)? Are you siliconeing your parts to the board or just using wire ties? I suppose the resistors should stand proud of the board slightly for cooling. I see that some like to eliminate any connection that is not a point to point solder connection. I'm thinking about using some barrier terminals at the inputs and then outputs to the speakers.

I still have quite a way to go, but I thought I would make up the crossover layout before I start gluing the bottom structure together. Hopefully that will lessen the chance of any unwanted surprises.

Thanks,
Peter

[matevana]

Hey Peter,

Please see my opinions below, after your questions:

Hi Ed

Thanks for the heads up.

(Q) I just placed an order with the electrolytics. I read a lot of conflicting opinions regarding the them. What sticks in my mind most is that the tolerances can be way off. Do you have any qualms about using them?

(A) Not in this application. I actually bought both and really didn't hear any discernible differences with this crossover, though it can be quite clear in other apps. Regarding their tolerance, perhaps I got lucky in the pairs I was sent, but here too I didn't detect any variances from side to side.  Another slight advantage with the electrolytics is their size; they will fit neatly on the crossover board and allow the overall assembly to be smaller. You can always start with the electrolytics and upgrade if you feel improvements can be realized, but I really didn't hear much difference.

(Q) Also, is the layout using the 8 ohm resistor (R3?) (+ _ L2 _ C2 _ R3 +)? Are you siliconeing your parts to the board or just using wire ties? I suppose the resistors should stand proud of the board slightly for cooling. I see that some like to eliminate any connection that is not a point to point solder connection. I'm thinking about using some barrier terminals at the inputs and then outputs to the speakers.

(A) I'm not quite sure what you are asking regarding the layout of the 8 ohm resistor? I used hot glue to secure the components and prevent them from vibrating. Less of a PITA than silicone if you want to reconfigure later. Heat should not be much of a factor, as all three top drivers are very efficient and basically comprise the midrange on up in the SL. They will likely see minimal wattage. You will find that the only driver that gets a bit of a workout will be the LF (Usher). If I touch any of the resistors after hours of play time, they are no warmer than any other component on the board. I used terminals as a matter of convenience but agree that anything you can do to minimize the signal path is a good thing. You may even want to consider three separate boards, which could then be placed in proximity to their drivers, or arranged strategically on the inside rear wings (where a single large board might not fit).   

(Q) I still have quite a way to go, but I thought I would make up the crossover layout before I start gluing the bottom structure together. Hopefully that will lessen the chance of any unwanted surprises.

(A) Always a good idea. When you actually wire them up, be sure to listen to each driver individually post crossover, and also be sure to check the phase relationship between drivers. I've made this mistake and find that a test cd, with signals recorded both in and out of phase can be very helpful here. This is especially true in OB, since the rear wave summed with the front wave can sometimes alter the perceived phase relationship at the listening position, even when the drivers are physically wired in phase. If you experience this, sometimes even minute adjustments in distance to any reflective surfaces can correct this. 

Thanks,
Peter

[sjhomey]

Thanks Ed

Is this the layout for the midrange crossover? I erased the P1.

[matevana]

Thanks Ed

Is this the layout for the midrange crossover? I erased the P1.

Exactly. The 8 ohm resistor is the last component in series prior to the Celestion mid coupling driver. It attenuates that driver a few dB's and adds some additional resistance to the cap and coil circuit.

[sjhomey]

Oops. I bought the .33mh inductor instead of the .22mh specified in the revised components list. Is there much difference in performance? They are inexpensive, just a little more time.

[matevana]

Haha, I have to laugh. I spent more time going back and forth between the .22 and .33 coils than any other component in the x/o design.  There is a difference but it may come down to your preference. They actually measure fairly close regarding F/R, but because of the range they affect it seemed to translate into how forward sounding instruments appear within the sound stage; almost like a presence control on a guitar amplifier. I would suggest building it with the ones you have. Since they are fairly cheap just order the others the next time you go to place a parts order. I would love to know your opinion of the two since I agonized over this decision. If you do decide to swap them out later it's a pretty painless procedure. 

If it helps in the decision process, the .33's sounded a little less bright in the upper mids by comparison; the center of the sound stage will move back a few feet from the listening position, which was about 8 feet in my case.

[sjhomey]

Ed
I'll take your advice and start with the .33 coils. I am planning to go point to point with all the wiring with no terminals so undoing the soldered bundle of positive leads at the input to replace the .33 with a .22 in the future might be cumbersome but not undo able. I am in no hurry so I could order the .22s and insert them to begin with if you think I'll end up there anyway.

On another note I'd like to pick your brain on stereo imaging. I heard a very expensive system last week and the placement of instruments was very pronounced, almost too much. Solo horns seemed to come from either one side of the stage or the other with no one standing in the middle. On my system with the Hestias and even with previous speakers I don't get much in the way of imaging. The Hestias give a deep soundstage but not much differentiation from side to side. I know everything starts with good source material but where else in the chain can good imaging be lost? I am using a Logitech Squeezebox Touch to bring in internet radio or my music stored on my computer and playing through a DIY Nelson Pass F5 amp. Any thoughts?

Peter 

[matevana]

Hey Peter,

For starters one of the original Hestia's design goals of low cost dictated a single sub amp which is also crossed fairly high . This design compromise will smear the stereo image a bit, since the left and right channels are summed in frequencies that are still fairly directional. After living with the Hestia's for a bit, I decided to build the Hestia-SL's to improve upon a few things; namely imaging and off axis frequency response. I do like the idea of point to point wiring, and would suggest the .22's if you think substituting them later will be a pain.

Your question about the audio chain is a good one. I have been doing a lot of tinkering these days with audio components and have come to the conclusion that they all play a role in imaging, however the digital to analog conversion stage can be most significant.  Whether you use an outboard DAC or rely on the transport's methods of conversion, it is the sound stage qualities that often differentiate between lesser and greater converters. Have you every considered taking the digital feed from your Logitech and passing it through an external DAC?  Outboard DAC's run the gamut in price and features, but its very possible to get a decent DAC these days for under $500, and quite a bit lower on the used market. The new crop of asynchronous DAC's such as the Schiit Bifrost, Peachtree Dac-it or Micromega MyDac are all decent. If you have an opportunity to try them out, you may find the contribution they make toward imaging is quite significant. They also handle streaming from a PC or Mac and will usually improve most S/PDIF devices such as CDP's and other players (DVD, BluRay, etc). 

I have become enamored with the $99 Pure i-20 digital dock. I use it with my iPhone and iPod. It can extract a pure digital signal from these devices (if you have your own outboard DAC) or it's internal 24bit 192khz DAC is actually quite good as a stand-alone, though you are largely limited to redbook CD quality (16 bit, 44.1) in native mode. I am impressed with its imaging capabilities, given what it costs. While my Hestia SL's are placed about 10 feet apart, I can easily identify instruments that appear dead center, when they are recorded that way. I can also identify a deeper sound stage that I could not so readily do with the original Hestias.   

[matevana]

Over the past 6 months I've been consulting for a sonar company that makes acoustic transducers. Having access to large scale measurement systems and anechoic chambers, I've been able to apply some newly acquired skills to crossover design.

In doing so I decided to revisit the Hestia-SL crossover. I was fairly confident that the original driver selection would still lend itself well to 1st order electrical (2nd order acoustical) slopes. It's amazing what state-of-the-art measurements can reveal and how much untapped potential there was. Another revelation over the past half year has been that not all passive components are created equal. Far from it in fact! I have really learned to appreciate how components can provide a significant edge in systems that are capable of revealing fine detail. I'm not necessarily referring to boutique priced components; just great sounding caps and coils that wont break the bank. 

The specific component combinations listed below provide significant improvement over the initial design.

Low/Mid:
Dayton 10" Pro Sound Driver 8 ohms

L1 = 4.00 mh, Solen 14 AWG Perfect Lay air core inductor 
C1 = 31 uf, Bennic non polarized electrolytic, 100V
R1 = 6.8 ohm, Mundorf MOX (metal oxide resistor)

Mid Coupler:
Celestion 5" Neo 8 ohms

L2 = .22 mh, Solen Perfect Lay 14 AWG
C2 = 5.6 uf, Clarity Cap ESA, 630V
R2 = 5.6 ohm resistor, Mundorf MOX

High Frequency:
Vifa XT19TD00-04 4 ohms

C3 = 5.6 uf Mundorf EVO aluminum oil capacitor, 450V 
(No further attenuation required in new crossover design)
 

[Juhazi]

Hi Matevana,
a multiway speaker's xo slopes /nad thus off-axis performance) will get a major bnefit from Duelund approach. It is kind of twin xo to make  acoustic responses to maintain coherent phase shift even past xo point. Then of-axis measurements were greatly improved, with added clarity when listening to music. A dipole's low end naturally makes that but low-pass needs more shaping.
http://www.musicanddesign.com/Duelund_and_Beyond.html

I found moving from LR2 to LR4 acoustic (with duelund approach) very beneficial with my AINOgradient project. I use minidsp for xo and eq - much easier than passive, but I need 8 amplifier channels to drive them!  http://www.diyaudio.com/forums/multi-way/231353-aino-gradient-collaborative-speaker-project-37.html#post3695824

[Juhazi]

..and just one more thing!
        After going  ot "Duelundish" xo my dipoles stopped sounding "OB" Kind of sad

[matevana]

Yes, I read about the concept when John K put it on his site. Interesting stuff and nice set-up by the way! My experience with MiniDSP wasn't all that positive, but it was a first generation model. Since that time a lot of folks have done some impressive things with bi-quad programming.

Can you elaborate about your speakers sounding less like OB with the different slope? 

[Juhazi]

OB-sound is kind of joke... I don't have any other dipoles at home but I have listened to some others.

When I started my project I thought that acoustic LR2 would be best. Then I went outdoors and made measurements with higher level (distortion check) and off-axis up to 180¤ Distortion results  made me to try LR4 and response shaping beyond xo. To my surprise off-axis response was much more even (better phase match) and also distortion was lower (easier job in low end) This with same on-axis response!.

When I carried the speakers back in, they sounded really different. Airyness was nearly gone but sound was clarified and bass-mid sounds were much better controlled. Kind of like less reverbs in the room. They sounded almost like my 2-way monitors, still better though!

[mcgsxr]

OB-sound is kind of joke

You lost me, can you explain?

I ran OB mains for over 7 years, and while no speaker I have owned was perfect, boy were they entrancing with certain music.

For the record I am currently using 2 way monitors and sub, my circumstances no longer allow for 36 inch baffles and the nest of wiring my speakers used to be!

That said, I am not done with this approach, once the kids leave and the basement is mine again (finished it over the winter, and suddenly more users are appearing, and sharing their opinion!) I will be back.

[Juhazi]

This how my "Duelundish" slopes look like. There is also "inverted" curve to show summing of each drivers and non-inverted overlap of W to HM and LM to T. Step response too.

"OB-sound" was a joke in my message! The sound itself is not a joke. At least I like them.
See my previous post for explanation

[matevana]

I disagree with the premise that the low pass filter always benefits by having more shaping. One of the goals of the Hestia project was the use of low slope first order electrical crossovers. Granted these can't often be used. But with the use of highly compatible drivers and proper crossover points they can sound better than anything else. Linkwitz was criticized for using first order electrical between the two mid bass drivers, but for those who have listened to the LX-521, they report that the driver coupling is flawless. A drivers dynamics are often times less at risk of compromise with a single component crossover then with a more sophisticated approach.