In the works...

5-7-2007:  Finally, the new house is in the works!  This place will include a 19' x 27' dedicated home theatre which off course will double as a test bed for listening evaluation of new speaker projects.  

Recently I visited with Siegfried and listened to the new Orion ++ with the rear firing tweeter.  I had been wanting to do this for some time because I could not get a good feel for the effect on my system at home due to space limitations.  In my current room, the speakers are only about two feet from the wall with the Sony 60" TV sitting between them.  I had been skeptical for some time and in fact had been considering the possibility that it was a mistake to use dipole higher than a few hundred Hz.  Some argue that dipole or omni above a few hundred Hz produce an artificial depth in the sound stage.  Leaning toward believing this I had considered the possibility of making the new system monopole starting at about 400 Hz.  After some experimenting with this however I concluded that it's simply not possible to control the off axis radiation with a monopole even above several KHz.  Consequently I have concluded that Siegfried is correct, the entire bandwidth must be dipole!  However, our reasons are very different.  I still believe that in fact, the net result must be that the speaker acts more like a Cardioid and that both side and rear reflections are a bad thing.

Listening to the Orion++ and the Pluto, I found similarities between the two speakers as Siegfried suggest but I also found differences.  To my ear, the imaging of the two speakers were polar opposites.  The Pluto defined positions of instruments and vocals with razor sharp clarity not only in position but in size.  The Orion ++ on the other hand broadened the size in all three axis and blurred the edges relative to the razor sharp edges of the Pluto.  Comparing the two side by side actually made both effects somewhat unpleasant.  On the Pluto, some instruments and vocals were so sharply defined in space it was though there was no body, just a floating mouth.  On the Orion ++ it seemed difficult to define position or size.  The one thing that stood out was the perception of depth.  The problem for me was that it did not seem that instruments and vocals were spaced in a larger depth but rather each one occupied more depth.  They were elongated!  So, much to Siegfried's surprise, my listening to the two speakers seemed to confirm my earlier assumption that both side and rear reflections are bad.  Unfortunately, time constraints limited my listening experience and Siegfried believes that listening to other material in longer sessions might change my mind but I don't think so.  At this point I still believe the ultimate goal would be to obtain more of a cardioid dispersion pattern that does not vary with frequency.  The only way I can think this might be approximated would be to make the entire speaker dipole and then try to get rid of the rear wave.  Finding a way to eliminate the rear wave uniformly across frequency will be the challenge.  Absorption material tends to be a low pass filter so this will need to be overcome.

I'm continuing to work on the new 6-way system that is planned for the theatre in the new house.  The board is coming along but it will still be some time before I'm ready to have the boards made.  The design of the speaker itself is also coming along but I'm going to need to get out of Asia and get back home so I can experiment with absorption materials.  I did listen to a set of dipoles while in CA where absorption material was being used to absorb the rear wave and the experience seems to confirm my assumption. 

Update 2-27-2007

The improvement in detail and clarity found in the 4-way Seas dipole has lead me to conclude that dividing the bandwidth up more is a good thing.  Proponents of single cone point source systems would probably disagree but I definitely have a different opinion as to the values in the balances of tradeoffs.  In fact, I believe the advantages of point source and other arguments for single full bandwidth drivers are questionable at best to start with.  The disadvantages of multi-way high order active crossovers are certainly minimal relative to the overwhelming disadvantages of any full bandwidth single driver approach.  So, a new 6-way crossover PCB is in the works...

One of the first big changes in the new crossover board will be the addition of a protection circuit.  Some time ago one of the Seas tweeters in the 4-way system was destroyed when a bad connection at the back of the crossover chassis caused the sudden loss of the -12 volt supply leaving the + 12 volt supply on.  The resulting sudden 50+ volt DC applied to the tweeter made very little noise and the tweeter never tweeted again.  So, on the new board, if either supply rail drops below about 10.8 volts, the entire power supply to the crossover will shut off.

The circuit to do this is very simple.  Q1 combined with R400 and D2 simply forms a selectable voltage regulator to match the relay coil voltage.  Q2 and D1 form a cutoff switch.  D1 is selected for the desired threshold.  When the voltage across D1 is less than the value of the zener, current stops flowing and Q2 turns off the relay shutting down power to both rails.

click to enlarge

The new crossover itself also has some differences.  Previously, each driver section was parallel and each driver saw only it's own high pass sections.  In the new crossover, it will be jumper selectable to pass each driver through all the lower high pass filters.  For example, the tweeter might see it's own high pass filter at 2 KHz.  It will then see the low-tweeter high pass section at perhaps 900 Hz, then the upper midrange high pass at 300 Hz and so on all the way down to the rumble filter on the subwoofer.

Another change will be to no longer use simple 4th order LR high pass filters.  Each high pass section will consist of a bi-quad (Linkwitz Transform) and a 2nd order filter.  So in the case of a tweeter for example, the natural 580 Hz roll off of the tweeter would be pole shifted to the 2 KHz crossover frequency using the bi-quad.  This will create an acoustic second order response that extends below the original 580 Hz resonance.  Then, the 2nd order filter will be applied at the crossover frequency.  The net result is a true 4th order acoustic roll off all the way down.  If a simple 4th order LR filter had been used at 2 KHz then the roll off would have been 4th order to the 580 Hz natural resonance and would then become 6th order below this.

The jumper selectable sequencing of high pass filters does not include the bi-quad as this section is considered a driver correction and not part of the effectively 2nd order electrical filter.  Each driver sees only it's own bi-quad, it's own 2nd order high pass filter and all the lower 2nd order high pass filters.  By doing this, each driver above tracks the driver roll off below.   Once the midrange for example starts rolling off at it's crossover frequency, the residual from the tweeter will attenuate at the same rate.  That is the 4th order acoustic roll off of the tweeter will be rolled off electrically at the same rate as the midrange electrical roll off.  The phase shift of the two drivers track!

I've been working on the board layout for some time but it's going to take a lot of time to complete it.  It's more art than science!  Finding a way to put 500 parts on a board so all the traces don't criss-cross excessively isn't an easy task.  I will eventually get it done, but it's probably going to be a few months down the road from this 2-27-07 update.  My new house with the dedicated Home Theatre is due for completion around October this year so it needs to be done before then.

Canceling woofer frame energy...

1.   One example of opposing woofer mounting for vibration cancellation..

.