Go Back   Home > Forums > Loudspeakers > Multi-Way

Multi-Way Conventional loudspeakers with crossovers

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 30th May 2007, 02:33 AM   #911
diyAudio Member
 
Lynn Olson's Avatar
 
Join Date: Jul 2004
Location: Northern Colorado
Quote:
Originally posted by Variac

-This probably is an all-out, proof of concept design, but think of how manufacturers start out with the all-out model, then there are "trickle down" benefits to less ambitious speakers. There are lots of things that he can determine that will help in future designs- possibly by him in the future:
You've read the tea leaves correctly. As mentioned earlier, this system is about halfway between the Ariel and the Karna in terms of how "out there" it is - the expense, complexity, and risk of complete or partial failure (in terms of reaching design goals).

I've screwed up before; the concept for the Aurora amplifier seemed good in principle, but turned out to be very tempermental about the choice of input tube, a sensitivity that pretty much undermined the whole project. The Karna amplifier uncovered a sensitivity to the exact choice of filament voltage, with a 2% shift being quite audible. The sheer transparency of the design - which probably has some of the lowest open-loop distortion of any amplifier - uncovered problems that were unsuspected with more conventional topologies. When you crawl out on a limb, sometimes it falls off, taking you with it.

My first speaker, the Audionics TLM-200, was in retrospect far too complicated, and the choice of drivers I was forced to work with (by management) were not the best choice, even for 1975.

Rest assured I'm building this thing. I've outgrown the Ariels, thanks to the Colorado extreme-audio crowd and the high-dynamic sound they like. They listen to custom-made SET amplifers with unobtanium tubes like prewar Telefunken DHT drivers and Eimac 75TL transmitter triodes, built with HP laboratory power supplies and custom-tuned Dave Slagle interstage and output transformers. The speakers are out there, too: Lowther/AERs with BIG Oris/Azura Horns and Klipschorn bass.

I really love the effortless Big Horn dynamics and the vivid tonal colors, but the close-in spatial presentation (and residual mid colorations) aren't something I can live with. That's where dipoles come in. They just sound right.

Combining Klipschorn dynamics with dipole sound may sound like an impossible goal, but why not try? This takes me far afield from a Linkwitz Orion, Visaton NoBox, or other popular dipoles, but that's OK. I've been kicking around the idea of combining prosound drivers with dipoles for a year and half now, and it's time to move forward and make it happen.

The huge contributions from all of the diyAudio crowd have pointed me in directions I wouldn't have thought of on my own - for which I am very grateful. I also can't emphasize enough that underlying concepts in this thread can be applied to any type of dipole you want, from big to small, from audiophile-inefficient to dance-club ultra-dynamics.

Quote:
Originally posted by rythmikaudio


Nice thread. The directservo configuration uses a combination of sensing feedback and current feedback. The purpose of the current feedback is to prevent local high Q peaks that can lead to instability at extreme operating conditions. It also provides a means for graceful degradation (the accelerometer based design does not have this, which is covered by our patents). In the vented box, it turns out the same concept can be used to flatten the output around the tuning frequency.

For simulation, one can take a regular driver, keep all other parameters the same except with the Re value reduced to 1/3 to 1/4 of its original value, update the Qes and Qts values, and plot the FR. You can see it produces a basically flat frequnecy response with some shelf-up characteristic at mid bass. Our servo network can flat that out too. If the result still have slight peak at the tuning frequency, we uses a 2nd order HP to make it a flat 6th order FR.

Most people have a very narrow interpretation of the servo. In my view, servo is based on the principle of remote sensing. As long as one can insert a 10ohms resistor into the speaker wire and still have the same FR (max output will change though), that is servo. It does not need to have sensing feedback from all radiators. The advantage of this remote sensing configuration is that it increases the cone movement control, together with all other benefits I put on the web site. Reducing the Re to its 1/3 value is like using a magnet 3 times as big.

I once filed a patent application for a servo based on PR with sensing feedback from both active and passive radiators. It is mathematically sound and perfect. But I cannot get it to sound right, which to me is far more important. I think I figured out why it will never sound right. Later in the process the patent examiner incorrectly cited a Japanese patent, I didn't bother to pursue further.

Personally I like "sealed" servo over "vented" servo. I can voice the latter with a couple of different components to sound close to sealed. I am constantly puzzled by this, BTW.

Brian

Rythmik Audio
A big thanks for chiming in, Brian. I was attempting to read between the lines on your website, trying to figure how "Direct Servo" really worked. With no accelerometer or microphone, a sense coil appeared to be all that was left, just as shown in the animation. Combining VC current sensing with smart limiting looks like an elegant way around some really nasty stability issues that have long troubled active-feedback woofer systems. If the servo has the ability to synthesize an Re 1/3 that of the real value, I'm guessing we're talking about 10 dB of direct servo feedback.

If I read your letter right, the VC current sensing is used to evaluate the proportion of woofer vs vent radiation by analyzing the skirts and depth of the impedance minima at the box frequency, and uses that additional information to compensate for the big drop in voltage from the independent velocity sensor.

I get the impression an additional circuit within the feedback board is acting as a multiplier or some kind of dynamic level-sensor, and applying this shaped, signal-processed voltage to the summing node of the feedback network. All this is pure speculation on my part, but something like this would have the "adaptive" properties mentioned on the website. It would be intelligent feedback, able to sense and avoid the disaster of oscillation, and also smart enough avoid damaging the driver from excess heat or excursion.

If this extra circuitry is there, I imagine you've had a lot of fun looking at the dynamic error terms on a scope while playing music. If my guess is right and you've been looking at the dynamic error terms, the difference between sealed and vented systems was probably rather enlightening, and well outside the limited information in the published literature.

P.S. My instinct, having thought about it a bit more - is to go with a pair of 12-inch sealed subs, one per channel. Small cabinets, less annoying box modes. Keep it simple.
  Reply With Quote
Old 30th May 2007, 02:38 AM   #912
diyAudio Member
 
Join Date: Feb 2004
Location: Texas
Quote:
Originally posted by AJinFLA


Hope things are going well for you in TX.

cheers,

AJ
Glad to hear from you too. I am still working on my own version of OB subwoofer though. Hope they will be available soon.

Brian

Rythmik Audio
  Reply With Quote
Old 30th May 2007, 02:57 AM   #913
diyAudio Member
 
Join Date: Feb 2004
Location: Texas
Quote:
Originally posted by Lynn Olson

If I read your letter right, the VC current sensing is used to evaluate the proportion of woofer vs vent radiation by analyzing the skirts and depth of the impedance minima at the box frequency, and uses that additional information to compensate for the big drop in voltage from the independent velocity sensor.
It is actually simpler than that. I am not a big fan of complex audio systems, which is very different from any other industry.
Quote:

I get the impression an additional circuit within the feedback board is acting as a multiplier or some kind of dynamic level-sensor, and applying this shaped, signal-processed voltage to the summing node of the feedback network. All this is pure speculation on my part, but something like this would have the "adaptive" properties mentioned on the website. It would be intelligent feedback, able to sense and avoid the disaster of oscillation, and also smart enough avoid damaging the driver from excess heat or excursion.
I hope I didn't mislead you. The term adaptive is merely to describe the difference between the EQ effect from servo vs the EQ effect from LT. There is no complex system in there. I normally advice against complex systems because they are harder to analyze.


Brian

Rythmik Audio
  Reply With Quote
Old 30th May 2007, 03:13 AM   #914
diyAudio Member
 
Lynn Olson's Avatar
 
Join Date: Jul 2004
Location: Northern Colorado
Quote:
Originally posted by rythmikaudio

I hope I didn't mislead you. The term adaptive is merely to describe the difference between the EQ effect from servo vs the EQ effect from LT. There is no complex system in there. I normally advice against complex systems because they are harder to analyze.

Brian

Rythmik Audio
That's reassuring, and refreshing, too! I'm not going to poke any further with groundless speculation about DirectServo and what's inside. What matters is that it works, and works on the problems that have troubled woofers for many decades.

I really like the appeal of servo feedback versus Linkwitz Transform. The feedback is just saying, "do this", and the desired FR and distortion improvement happen automatically.

With a Linkwitz Transform, you get the desired FR, but only at small signals. Large signals skew the Theile/Small parameters, perturbing the filter alignment. Distortion is also magnified due to substantially increased excursion, and you flirt with destruction of the driver if you get greedy and ask for too much.

The alignments that really bother me are the mathematical stunts like 6th-order active-equalized vented systems, or even worse, bandpass systems. All of these rely on linear small signal models - to implement rather tricky high-order filters. All it takes are suspension and magnetic nonlinearities to creep in and the model falls apart.
  Reply With Quote
Old 30th May 2007, 12:22 PM   #915
Pano is offline Pano  United States
diyAudio Moderator
 
Pano's Avatar
 
Join Date: Oct 2004
Location: Milliways
Blog Entries: 4
The servo thing is cool.

Years ago I used to tour with a rig that had processors that plugged in between the amps and the speakers. So it was sensing the current on the speaker lines, then looping some info back to the amps.

Worked pretty well, IIRC. Mostly it would just compress if the system was run too hard. Kept me from ever blowing it up, though!
__________________
Take the Speaker Voltage Test!
  Reply With Quote
Old 30th May 2007, 05:51 PM   #916
Banned
 
Join Date: Feb 2004
Location: Portal 2012
Lynn.

In my experiance with multiamped BIG open baffles using pro drivers, Fully horn loaded systems and large and small format compression drivers I recommend doing one of two things -

One, if you want dipole radiation use multiple 10" mid QTS pro drivers (like the Madison Warriors- (they kick butt with no eq) for bass (six of them, stacked three high pairs) 40-50 cycles up to a coax (like the Radian 8-10") and cross between 150 and 250 cycles. I have found multiple smaller drivers closely spaced or in a line sounds better than 15's in an open baffle.

I have the 8" radians here and am tempted to build this- problem is I'm just finishing up on a big hog horn system and it sounds sooooo good I think I'll just keep it for the summer.. The little Radians show great promise but aren't really as sensitive as I like-

Two, use a large format compression driver with at least a 3" dome in a suitable 300 Hz flare round horn crossed at 500-600 cycles to a 12" or 15" front loaded straight hypex horn with a 40-50 Hz flare. A super tweeter may be needed.

Either system will require a subwoofer system. The horn system will be more demanding of the sub - and probably will require more subs or a lower compression sub system to 'keep up' and integrate with the bass. The horn system will also be easier to drive, be less complex and be more dynamic, giving you more of the 'ease' you mentioned in an earlier post.

AFA the midrange colorations well I know they can be tamed to meaningless amounts with the right horns - what you will loose with the horns compared to the dipole is 'open' - They will be more directional in the midrange (smaller wide range listening area) but that can be a blessing.

You just won't get the dynamics with out the horns - I have been there a few times. Surely what you propose is superior to most commercial systems in low compression and low coloration but it won't have the ease of the horns..
  Reply With Quote
Old 30th May 2007, 06:39 PM   #917
ScottG is offline ScottG  United States
diyAudio Member
 
Join Date: Feb 2003
Location: US
Quote:
Originally posted by rythmikaudio

In the vented box, it turns out the same concept can be used to flatten the output around the tuning frequency.

If the result still have slight peak at the tuning frequency, we uses a 2nd order HP to make it a flat 6th order FR.

Personally I like "sealed" servo over "vented" servo. I can voice the latter with a couple of different components to sound close to sealed. I am constantly puzzled by this, BTW.

Brian

Rythmik Audio
The question I have here is:

Precisely why do you like the sealed over the vented servo? (..for a given freq.)

Also, whats the overall design of each?

My overriding concern with the design is quality in the bass range, not sub harmonic. At the same time though, I recognize that the lowest freq. band does effect the rest of the spectrum, so here you really have to know what the design is doing.

In particular I would NOT want a flat summed response from the port tunning freq.. In fact I wouldn't even want such a design to be flat in-room at the very low tunning freq. (..rather I would want it several db's lower than the average in-room).

Again, the design I suggested (though inclusive of achieving bass extension), was primarily concerned with increasing the mechanical dampening of the driver to extend a more mass controlled character to it throughout the fundamental *music* range (i.e the bass range - excepting organ music, low bass techno music, and the like).

The problem here, at least as I see/hear it, is keeping up with the subjective clarity of the dipole system in the midbass while adding a degree of tactile sensation ("punch" and "slam") that the dipole is lacking - and capable of doing all this at a higher output level than any reasonable sized dipole can when considering size and distortion. (..also providing the ability to eq. the system in this problematic/modal freq. region without affecting the dominant tonal character of loudspeaker & amp pairing.)

A true sub then could be added in addition to such a design (..and here my preference would be for an IB room corner loaded).
__________________
perspective is everything
  Reply With Quote
Old 30th May 2007, 08:24 PM   #918
AJinFLA is offline AJinFLA  United States
Banned
 
Join Date: Feb 2005
Location: Tampa
Quote:
Originally posted by rythmikaudio


Glad to hear from you too. I am still working on my own version of OB subwoofer though. Hope they will be available soon.

Brian

Rythmik Audio
Hey, I had to wait a couple years, but I bet it will be worth it. Figured you'd come around sooner or later .
Open baffles, waveguides, servo subs... interesting thread here eh?

cheers,

AJ
  Reply With Quote
Old 30th May 2007, 08:27 PM   #919
mige0 is offline mige0  Austria
diyAudio Member
 
mige0's Avatar
 
Join Date: Jan 2005
Location: Austria, at a beautiful place right in the heart of the Alps.
Hi

Quote:
I have found multiple smaller drivers closely spaced or in a line sounds better than 15's in an open baffle.
Magnetar, it seems you like it VERY loud sometimes ?


http://www.diyaudio.com/forums/showt...289#post963289

Do you also have a nice picture of such a 10" OB line array like mentioned above?
You would not recommend a deep sub ( 20-50Hz ) in any kind of dipole configuration not even with double 21", or did I misinterpret your numbers ("Dual 21" Woofers 45-330 cycles 99 db sensitive") ??

Measurements that were published by a German DIY magazine several times now show that the lowest subwoofer octave can be preserved by aligning the axis of the dipole in PARALLEL to the front wall. They place a dipole W-style sub close to the wall and claim a more or less linear FR down to 20-25 Hz .

Do you have any experience with something alike?

Greetings
Michael
  Reply With Quote
Old 30th May 2007, 09:01 PM   #920
diyAudio Member
 
Lynn Olson's Avatar
 
Join Date: Jul 2004
Location: Northern Colorado
Default On The Other Side of the Rainbow

Here's the HF crossover for the compression driver, using a minor variation of the Klipsch and ALK crossovers.

The transformer (or autoformer) has some interesting advantages compared to an L-Pad:

1) The load as seen from the HP filter falls in very narrow range between 8 and 10 ohms, no matter what the compression driver is doing (the transformer impedance ratio is the square of the turns ratio). This essentially gets rid of the interaction between the triple impedance peaks (induced by the horn cutoff) and the crossover. The crossover is so insensitive to the compression driver it makes very little difference if the driver is 8 or 16 ohms; both look pretty much the same on the other side of the damping resistor.

2) The source impedance as seen from the compression driver is equally favorable. Instead of seeing the series impedance of the crossover and the optional equalizer, it sees the 10 ohm damping resistor divided by 16, which equals 0.625 ohms. This virtual source is in parallel with the amplifier source impedance, so the source impedance is actually well below 0.625 ohms in the passband of the crossover. The driver behaves very close to a multi-amped system in terms of source impedance (damping factor).

3) The transformer is not going to be particularly large, since there is no requirement to handle low frequencies, nor does it need to handle any more than 50 watts (under the most extreme conditions). The Klipsch autoformers are tiny, less than a cubic inch. I'll be asking Bud Purvine and Dave Slagle to make this transformer for the project - it isn't that different than prosound 70-volt distribution transformers, just a higher-quality application.

4) The inductance compensation (Zobel) is there to be nice to the transformer, giving it a resistive load beyond 100 kHz. This simplifies the transformer design, and makes it easier to get good square-wave performance. The Zobel also simplifies the crossover design, as well as making a gentler ultrasonic load for transistor amps with marginal feedback stability.
Attached Images
File Type: jpg dipole_hp_filter.jpg (31.0 KB, 894 views)
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



New To Site? Need Help?

All times are GMT. The time now is 08:45 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2