[Aether Audio]

Hey Guys,

Just a little update:  For those of you that heard our stuff last year, get ready for a bit of a shock.  For those that haven’t, get ready for a much bigger one.  Our products are now on a whole other level since this most recent tweeter upgrade.  Not only that, but I’ll confess a little discovery I made.  I know most of you understand the importance of loudspeaker impedance – especially at low frequencies – and it’s effects on amplifiers.  Even most designers don't consider that lower impedances at higher frequencies  to be too much of a concern,  After all, there’s not much power actually delivered to the speaker at the higher end of the spectrum – right?

Don’t kid yourself... like I did.  In my efforts to develop a working crossover for this new tweeter, I didn’t concern myself too much with optimizing the high-pass (tweeter’s circuit) section for optimal matching with the amplifier.  Most modern amplifiers from reputable companies can deal with such issues without much trouble.  During development we mainly worked to get our traditional flat response while seeing to it nothing “bad” was done in the process.  I suspected that it might be possible to optimize the impedance further, but after initial listening sessions, there seemed to be no point.  The system sounded absolutely wonderful so "what the heck, if it ain’t broke – don’t fix it.”

Well, before too long my neurotic obsession with optimizing performance got the better of me… and back to the lab I went.  After just a few minor modifications I had a circuit that significantly raised the impedance seen by the amplifier when driving the tweeter.  I thought, “This should help sweeten things up just a bit more,” but I was in for a shock.

The first implementation of this new circuit was in an upgrade for a pair of Continuum A.D.’s belonging to Joe Jurzec of JAM’n Audio.  Joe was here with me that day and he heard the first results the same time I did.  From the first “wack” on the ol’ Chinese Drums track – I jumped - it actually startled me.  I could hear the difference in that very first sound emanating from the system.  Never in my wildest dreams could I have imagined how much difference that little “mod” could make.  Let’s put it this way – To me the improvement was almost as significant as the new tweeter itself.

But don’t take my word for it.  Just ask Joe, he’ll tell you.  We were both blown away.  His “little” Continuums actually blew away a pair of un-modified Revelations we had just been listening to - and not only in the highs either.  His Continuums sounded better than my reference Revelations in EVERY way... even in the bass.  

It all makes sense now though,  the higher impedance means the amp doesn’t have to work as hard.  The error amp stage isn’t so busy trying to correct for back emf generated by the more highly reactive circuits in the old design.  But that’s amplifier design jargon.  

Let’s put it this way:  Pay close attention to the impedance curve over the ENTIRE frequency range of any speaker you may be considering…yes, even at high frequencies.  If it exhibits somewhat excessive electrical phase and/or low impedances – watch out!  That’s were paying attention to those little plots generated by John Atkinson at Stereophile can really give you some insight.  

You know what’s funny?  Many audiophiles almost dismiss the importance of measurements and such.  I even fell into that trap to a certain degree - at least in this case.  Well, here’s just another case of when the end result is a by-product of the underlying physics and is both clearly audible and easily measurable.  Go figure.  Just thought you might be interested.

Take care,8)
-Bob

[ctviggen]

Interesting. I've always heard that tweeters need more power than bass drivers, although most people think it's the other way around.  Now, whether that's because of the impedance curve or not, I'm not sure.

[Karsten]

Interesting. I've always heard that tweeters need more power than bass drivers, although most people think it's the other way around.  Now, whether that's because of the impedance curve or not, I'm not sure.

The tweeter does not need large amounts of power, but it will need some voltage to get the good dynamics.

Brg,
Karsten

[JeffB]

I am kind of curious what sort of design your test amplifier has.  I have heard that class D amps are very sensitive to impedance issues.  I also recently read an article about negative feedback and how inaccurrate the feedback can be.  I suspect a more even impedance might help with the accuracy of the negative feedback.  I guess I wonder if the fix will show improvement across all amplifiers.

[Christof]

I could be wrong but I think Bob is using one of the Dave Belles' amps in his test set-up?

cg.

[Aether Audio]

Jeff,

Thanks for asking!  This is a very good question because it directly relates to our own experience.

First of all, we listened to the previous version of the crossover on three different SS amps.  One was a high power, later model Crown Comtech amp that uses Bi-Polar outputs and relatively large amounts of negative feedback.  We used it mainly just to test for dynamic range issues.  With 800 W/ch. you can really push the limits.  It souded fine, well...as fine as an amp made for soud reinforcement can sound.  The high frequency performance of that amp was never what you could call sweet or audiophile quality.  Even on other designs that didn't use this crossover, it never sounded all that great.  Still, the Revelations with the original X-O sounded pretty darn good on it.

Well, then we listened on a lower power (130 W/ch.) Belles amp.  It has MOSFET outputs and somewhat (I think)lower feedback levels.  It sounded much sweeter and detailed - so much so that I figured I was done with the design.

Then we listened on an old Crown DC300.  It sounded somewhere between the two amps above.  Of course, it has Bi-Polar outputs and high feeback, but it still sounded better than the Comtech.

Then Joe came over for his upgrade.  As a little treat he brought a Ref 9 Nuforce amp and a Modwright preamp.  Two notes:  First, I would have never guessed the Modwright had tubes - it souded as clean, quiet and accurate as any SS I've ever heard - totally transparent.  Second, I LOVE THAT NUFORCE AMP!!!  That amp was spectacular.

Well, we listened to his gear on my reference Revelations before we got his upgrade installed.  He couldn't get over how good everything sounded and the increadible dynamics and zero distortion at high levels.  My listening area is in our production facility and the only thing we have to divide that space from the rest is some acoustic panels set up around the perimeter.  We're talking 11,300 sq. ft. of space in this building and 15 foot high ceilings made of fiberglass insulation laying across truss beams.  His electronics and the Revelations were so loud we had to yell at each other from 50 feet away where I was working.  All this was going through the un-modified crossover.  So...I'd say that little Nuforce amp held it's own just fine, even driving the more difficult load of that X-O.  I never had a clue that there was anything wrong with the sound - and Joe, well, he was so impressed he could have never guessed either.

After all that, we swapped out the Rev's for his newly updated Continuums with the modofied crossover and...the rest is history.

So, there you go.  Four amps - 2 Bi-Polar, 1 MOSFET and 1 digital, and they all seemed to be fine.  Of course, we didn't ever try a tube amp on the old X-O, but that's about the only thing lacking in this little "accidental" test.  The upshot was that the impedance issue never stood out as a problem on any of the amps to our hearing, but then everything changed with the higher impedance mod.  Maybe a zero-feedback amp would have shown a bigger difference.  With the typically low damping factor those designs exhibit, it would not have been able to correct for the back emf generated by the X-O and the sound would have been worse, right from the start.  I don't know.

All I know is that raising the circuit impedance transformed our already exceptional performance into one I would have never dreamed of - and one I have never heard anywhere before in my life.  I'm not trying to "blow my own horn" here about our products.  It's just a fact.  I've listened to a lot of different speakers over the years and I've NEVER heard anything like this before.  OK, I'm gushing now. I'll shut up.

The main point is that the speaker impedance at all frequencies appears to be a very important issue - far more than I would have ever guessed.  I'll bet there are a number of designs out there that could be significantly improved if their designers re-visited the impedance issue.  There are always things you can do to compensate for impedance.  Maybe not so much for it's absolute "real" component, but you can definately nullify excessive reactive components that lead to large amounts of electrical phase lead or lag.  

I can tell you one thing for absolute certain.  I know of one designer that is going to from now on.

-Bob

[Danny Richie]

Welcome to the club Bob.  

[Aether Audio]

Danny,

You just had to get your "digs" in  - din't chya?  Hey, it's not that I didn't hear what you were saying, it's just that I worked my tail off to get a design that worked and I really didn't have the time right then to investigate it further.  After all, it already sounded fantastic.  How much better could it get?

I guess I have to credit this one to you thiugh.  Our little conversation stuck in my craw and I couldn't get it out of my head.  Yes ol' "Wise One," you were RIGHT!  Thanks!!!

There everybody, Danny told me so.  Maybe not in so many words, but he did try to warn me in general.  

Danny, in my own defense, I still bet even you wouldn't have guessed it would have made this much difference if you would have heard the original - it was already "that good."  So there.

G'night,
-Bob

PS. I'll call ya.

[ekovalsky]

There is some good reading about this topic on Paul Speltz's website

Almost everyone I know who has used his transformers on their low impedance speakers (mostly classic Apogees) has been thrilled with the results.  Anyone with speakers of 4ohm or lower impedance should check into these, particularly if they are using tube amps.

[Danny Richie]

Bob, I can't wait to hear them. I know they'll be great.

[Aether Audio]

Danny,

Thanks!  Can't wait to get together in a week.


ekovalsky,

Thanks for posting that link.  I checked them out a while back and the autoformer makes perfect engineering sense.  

There a couple of cautions I would add though.  Anybody seriously considering them must remember that the device is a transformer.  Any type of transformer that has to deliver power and operates over the audio band almost enevitably has a steel core (typically silicon steel).  Steel, as well as most other "non-air" core material, has an upper limit to the amount of magnetic flux that it can accept.  Above that point it begins to "saturate" (kind of like amplifier clipping) and will then generate harmonic distortion.  The autoformer above looks to have been engineered about as good as reasonably possible.  Still, at least from a purely technical standpoint, it will not be as "pristine" as connecting your speakers directly to the amp output.  Nevertheless, it's benefits may out weigh any drawbacks in this regard.

The bigger issue is one of reduced amplifier headroom.  If you use such a device to raise the "apparent" impedance of the speaker, you are essentialy trading current for voltage.  The amplifier will not have to source as much current, and that's what suposedly produces the improved sonic performance.  Basically, the amp doesn't have to work as hard delivering large amounts of current.

But...In order to achieve a certain volume or SPL, the amp will have to source a higher output voltage to achieve that SPL than it would have driving the speaker direct.  Every amplifier has a maximum output voltage that it can deliver.  Above that point - it clips.  If it has to deliver more voltage to reach a certain volume, then there's less voltage left for short duration transient peaks.  The upshot is that you could end up clipping the amp far more often than before you inserted the autoformer.  Of course, one way around that is to use an amp with a higher 8-ohm rating.  But then, you may not need the autoformer anyway.  It's a toss up.

If a guy is really curious though, it's worth a shot if you have the spare $$$.  As for me, I would just opt for the bigger amp and be done with it.  But then, I may be just a litlle power mad to start with.

If anybody has any experience with using this device, I'm sure we'd all be interested in hearing from you.  I know I would!

-Bob

[JeffB]

I know that giorgino1 has used an autoformer with a Decware SET amp(Zen) and a pair of hornshoppe horns.  The Zen prefers 2 ohm loads and the autoformer was used to reduce the impedance.  I believe he thought it helped the sound in his system.  Decware came out with a modification for their Zen so that it worked better with 8 ohm loads.  giorgino1 had the modification performed and stopped using the autoformer.  If you really want the details, I would send a message to giorgino1.