Getting the full SDA Effect from Polk L800 Speakers

[Matthew J Poes]

Interesting, how is it mathematically and electrically possible? I can see how a differential preamp to output twice the voltage, but we are talking about class AB power amp when you referred to rail voltage right?

If so, can you kindly elaborate on how you can get twice the output (again, I assume you meant the power amplifier output) using half the rail voltage, simply by using differential drive?

The Fully Balanced Power Amplifier - Advantages and Design Challenges - HomeTheaterHifi.com

The Fully Balanced Power Amplifier - Advantages and Design Challenges

hometheaterhifi.com

Also read this, I can't equal this guys article. It's one of the best coverages of this I have ever seen.

Bridged and paralleled amplifiers - Wikipedia

en.wikipedia.org

This also talks about the voltage doubling due to bridging.

 

[PENG]

The Fully Balanced Power Amplifier - Advantages and Design Challenges - HomeTheaterHifi.com

The Fully Balanced Power Amplifier - Advantages and Design Challenges

hometheaterhifi.com

Also read this, I can't equal this guys article. It's one of the best coverages of this I have ever seen.

Bridged and paralleled amplifiers - Wikipedia

en.wikipedia.org

This also talks about the voltage doubling due to bridging.

I read that article a few time, did not see that he make such a claim that you can get twice the output for half the rail voltage for fully balanced amps. I can see the argument that the swing from +30 V to -30 V is of the same magnitude as from +60 V to 0 (ground) but you don't get twice the output power because the total swing is the same in absolute sense.

 

[PENG]

Because the total swing is what matters. So here is how a differential output amplifier using this bridge tied load approach works. Take two identical push pull amps (as an example). Each has 75 volt rails. For the sake of argument, you have 10 amps of current from the power supply. That is 1500 watts based on the swing between +/-75 volts (150 volts), but we need to divide by 1.41 to get to RMS. Class AB would only be 65% efficient. Class D would be 90% efficient. Let's go with Class AB. Now we are down to 975 watts peak and 691.5 watts RMS and that would be pushing things pretty hard. Now we go to a bridge tied load version, the second amp is inverted (phase is reversed. You tied the positive terminal of the second amp to the negative terminal of first amp. That is now referenced to ground. It doesn't touch the speaker. The positive terminal of the first amp goes to the positive terminal of the speaker. The negative terminal of the second amp goes to the negative terminal of the speaker. Again, no ground connection here. The two swing between these amps is now going to be much higher, 300 volts, like having one amp with +/-150 volts. Now we have basically 1300 watts of RMS power output ability. Or...you could go down to 35 volt rails, now the caps are very small and way cheaper/way more reliable.

Here is a schematic of one of the older SDA speakers:
View attachment 46434

Basically it is pulling the sound information from just the positive terminal, which on a single ended output amplifier, is fine, all of the signal is present on the positive terminal of a single ended output amplifier.

View attachment 46435
Here is a PP amplifier. You can see there is a positive output, load, ground. All the signal is on the output here.

But a differential amplifier is different.
View attachment 46436
Half the signal is on the + output, half is on the - output. Go back to the rail voltage comment. To obtain the total swing, the sinewave isn't swinging up and down the full amount on just the positive output. Instead, half the voltage level is on the positive terminal and half is on the negative terminal, with inverted phase. That is what gives the total voltage swing. So for a given wattage of output, the stereo array speakers would only see half the voltage swing that the main drivers see, thus 6dB quieter.

Does it make sense now?

Thank you very much, now I have something to chew on, and see if it makes sense.

 

[Matthew J Poes]

Thank you very much, now I have something to chew on, and see if it makes sense.

"for example the maximum output voltage swing of each amplifier is between a peak of + and – 10 volts, when the output of one amplifier is at + 10 volts the output of the other will be at –10 volts, which means that the load (a loudspeaker) now sees a 20 volt peak difference between the “hot” (normally red ) output terminals." From the wiki article.

 

[PENG]

"for example the maximum output voltage swing of each amplifier is between a peak of + and – 10 volts, when the output of one amplifier is at + 10 volts the output of the other will be at –10 volts, which means that the load (a loudspeaker) now sees a 20 volt peak difference between the “hot” (normally red ) output terminals." From the wiki article.

Yes I said that too using 30 V for example, it is obvious, but you don't get twice the power output. The power fomula P = V^2/R still applies. so a 20 V to 0 V single ended Vs +10, -10 V legs still give you the same max power output, not twice. I guess you must agree, no? May be this time we both need our 2nd, 3rd cup, wish I can buy you one for trying to answer all my questions, but you would have to come to Canada, after Covid obviously.

 

[Matthew J Poes]

Yes I said that too using 30 V for example, it is obvious, but you don't get twice the power output. The power fomula P = V^2/R still applies. so a 20 V to 0 V single ended Vs +10, -10 V legs still give you the same max power output, not twice. I guess you must agree, no? May be this time we both need our 2nd, 3rd cup, wish I can buy you one for trying to answer all my questions, but you would have to come to Canada, after Covid obviously.

You are still missing the point.

The rails of each individual amplifier of half of what the normal amp would be. Even you have to agree with that, right? If each amps voltage swing is referenced to the 0 volt state here, the ground, then an amp that is swinging 30 volts, when combined with another inverted amp in a BTL approach, is now swinging 60 volts.

 

[Matthew J Poes]

Yes I said that too using 30 V for example, it is obvious, but you don't get twice the power output. The power fomula P = V^2/R still applies. so a 20 V to 0 V single ended Vs +10, -10 V legs still give you the same max power output, not twice. I guess you must agree, no? May be this time we both need our 2nd, 3rd cup, wish I can buy you one for trying to answer all my questions, but you would have to come to Canada, after Covid obviously.

I just realized something when you noted that formula. Rails refers to the power supply. Not the voltage swing at the amplifier output. Normally when someone uses the terms rails they means the amplifier power supply rails. Not the amplifier output. So yes, the amplifier output voltage swing is the same in both an SE and differential map. It’s the power supply that has half the voltage.

 

[ski2xblack]

Thank you very much, now I have something to chew on, and see if it makes sense.

The networks are extracting a "L - R" ambient signal (as opposed to normal stereo L + R) passively, a la the old Hafler dynaquad approach. That explains the need for the floating ground on the amp and extra wire connecting the speakers together. The positive amp terminals are connected to the networks/speakers, but the negative speaker terminals (on the drivers that recieve the extacted signal) are connected to each other, not the amp, hence the different power draw from the rails. Apologies for any incorrect technical terms.

Does that help clarify whats going on electrically?

 

[TLS Guy]

The networks are extracting a "L - R" ambient signal (as opposed to normal stereo L + R) passively, a la the old Hafler dynaquad approach. That explains the need for the floating ground on the amp and extra wire connecting the speakers together. The positive amp terminals are connected to the networks/speakers, but the negative speaker terminals (on the drivers that recieve the extacted signal) are connected to each other, not the amp, hence the different power draw from the rails. Apologies for any incorrect technical terms.

Does that help clarify whats going on electrically?

You know I think you are right. I think this is not really sending out of phase output from the opposite channel. That would null the center information.

So in fact it probably is the old Hafler circuit used as "pseudo surround" back in years done by. I experimented with that and abandoned it smartly. It did have somewhat of a vogue. So what they are doing with these speakers is using that crude passive Hafler circuit to derive difference signal between left and right channels, and then feed that out of phase to the outside drivers of the opposite speaker.

Here is a good description of the Hafler circuit. It does explain why an amp without a true ground for at least one speaker terminal would not drive that Polk contraption. In addition the fact that it does not work with balanced output amps, gives the game away as you have correctly surmised. Very well done six2xpack!

Really, Polk (Sound United should just be honest about what they are doing and stop the confusion. However I suspect that if they had been honest about regurgitating this old and largely discredited approach, they would not have much of a market.

 

[Irvrobinson]

The networks are extracting a "L - R" ambient signal (as opposed to normal stereo L + R) passively, a la the old Hafler dynaquad approach. That explains the need for the floating ground on the amp and extra wire connecting the speakers together. The positive amp terminals are connected to the networks/speakers, but the negative speaker terminals (on the drivers that recieve the extacted signal) are connected to each other, not the amp, hence the different power draw from the rails. Apologies for any incorrect technical terms.

Does that help clarify whats going on electrically?

I'm not convinced yet. You might be correct, but the entire design is patented, and I have trouble believing the strategy is just an L - R signal. Here is the patent. This site does not include figures referenced by the patent, and since the patent is verbose it makes it annoying to follow. I'm tied up most of the weekend, but perhaps one of you can make sense out of it.

US Patent Application for Head Related Transfer Function Equalization and Transducer Aiming of Stereo Dimensional Array (SDA) Loudspeakers Patent Application (Application #20180317034 issued November 1, 2018) - Justia Patents Search

An enhanced Stereo Dimensional Array (“SDA”) loudspeaker system 250 preferably including a mirror image pair of loudspeaker enclosures 280L, 280R configurable by a user or installer as a left-channel SDA loudspeaker and a right channel SDA loudspeaker uses a novel driver array aiming...

patents.justia.com

 

[PENG]

The networks are extracting a "L - R" ambient signal (as opposed to normal stereo L + R) passively, a la the old Hafler dynaquad approach. That explains the need for the floating ground on the amp and extra wire connecting the speakers together. The positive amp terminals are connected to the networks/speakers, but the negative speaker terminals (on the drivers that recieve the extacted signal) are connected to each other, not the amp, hence the different power draw from the rails. Apologies for any incorrect technical terms.

Does that help clarify whats going on electrically?

Thank you for the information. It probably would clarify things enough for me but I have to digest all the information provided so far that includes a schematic Matthew posted that he said was for a different (older) model, and the patent info Irv just posted. That's a lot to read and digest, so thank you all for giving me something to read for another boring Covid weekend.

 

[ski2xblack]

I'm not convinced yet.

From your link:
"The outer-baffle set of drivers (aiming on an axis away from the centered listener) play the crosstalk cancellation or SDA dimensional effect signals. Crosstalk cancellation (or SDA dimensional effect) signals are generated by crossover circuits connecting the loudspeakers to the amplifiers such that the left tower gets an “L-R” signal and the right tower gets an “R-L” signal. An electrical crossover network is used to make the crosstalk cancelling signals used to drive the dimensional or SDA effect tweeter/midrange driver array by matching the main tweeter/midrange driver array's signal and compensating for the headshadow. In the prototype a simple R-L shelf circuit was used to achieve this."

 

[ski2xblack]

Wow, that patent is a bizzarre, postmodern masterpiece. It makes quite a few claims and goes into a lot of analysis without actually supporting the claims. TLS basically nailed the fundamental design, but they've added a first order filter to the ambient arrays, resulting in the associated phase shift, and have renamed that as "head shadow compensation" in true postmodern fashion.

While it's beyond ridiculous, I can see how the combo of arrays, driver spacing, and appropriate listener positioning would produce some kooky, unconventional spatial effects. Genuine crosstalk cancellation though?

 

[Irvrobinson]

From your link:
"The outer-baffle set of drivers (aiming on an axis away from the centered listener) play the crosstalk cancellation or SDA dimensional effect signals. Crosstalk cancellation (or SDA dimensional effect) signals are generated by crossover circuits connecting the loudspeakers to the amplifiers such that the left tower gets an “L-R” signal and the right tower gets an “R-L” signal. An electrical crossover network is used to make the crosstalk cancelling signals used to drive the dimensional or SDA effect tweeter/midrange driver array by matching the main tweeter/midrange driver array's signal and compensating for the headshadow. In the prototype a simple R-L shelf circuit was used to achieve this."

I'll be damned. Occam's Razor is correct again. Thank you for persevering through the most poorly-written patent I've ever read.

 

[mns3dhm]

Thank you for reviewing this speaker and explaining the technical design aspects in detail. I'd love to see another Audioholics speaker roundup like the one you did not too long ago on floor standing speakers under $2000. I'd be curious what Audioholics thinks of available floor standing speakers at a bit higher price point, say $3000 - $5000 a pair. Thanks again!

 

[TLS Guy]

I'll be damned. Occam's Razor is correct again. Thank you for persevering through the most poorly-written patent I've ever read.

You have that right. That is basically a massive smoke screen to provide cover for what basically is a very crude piece of frippery. I will tell what that head shadow is for. I was thinking about one problem last night. The outside drivers and inside drivers have to have pretty much opposite phase with respect to the woofers. So that circuit will in fact be to split the difference.

I do have a fundamental issue with this approach. Most speakers are literally awash in time shifts. This is the most neglected problem in most commercial designs. I know that this is an issue in speech intelligibility and producing really convincing sound fields. So now we have a major company producing a design that is literally piles on time shift on time shifts. What they should actually be doing, is devoting their energy and resources to eliminating time shifts. This whole episode is corporate gimmickry, elevated to the nth degree.

 

[Golfx]

You have that right. That is basically a massive smoke screen to provide cover for what basically is a very crude piece of frippery. I will tell what that head shadow is for. I was thinking about one problem last night. The outside drivers and inside drivers have to have pretty much opposite phase with respect to the woofers. So that circuit will in fact be to split the difference.

I do have a fundamental issue with this approach. Most speakers are literally awash in time shifts. This is the most neglected problem in most commercial designs. I know that this is an issue in speech intelligibility and producing really convincing sound fields. So now we have a major company producing a design that is literally piles on time shift on time shifts. What they should actually be doing, is devoting their energy and resources to eliminating time shifts. This whole episode is corporate gimmickry, elevated to the nth degree.

So you gentlemen see only negative? Negative with the review, negative with the premise, negative with the wording of the patent application, negative with Polk for designing it. Geez.

 

[shadyJ]

So you gentlemen see only negative? Negative with the review, negative with the premise, negative with the wording of the patent application, negative with Polk for designing it. Geez.

It's worth noting that only the guys with negative conclusions on the speaker are also the guys who haven't actually heard it. The guys who have, Matthew, Gene, and myself, all think it's a great speaker.

 

[Irvrobinson]

So you gentlemen see only negative? Negative with the review, negative with the premise, negative with the wording of the patent application, negative with Polk for designing it. Geez.

It's worth noting that only the guys with negative conclusions on the speaker are also the guys who haven't actually heard it. The guys who have, Matthew, Gene, and myself, all think it's a great speaker.

Since I think I'm being included in this group of gentlemen, I would like to make sure my views are being represented properly. I am not negative on the speakers themselves, since I've never heard them. Being negative on a speaker without hearing it is silly. In fact, I'd like to hear them, if anything to see what all the hubbub is about.

I did have questions about Matthew's room. He answered them. His room is so different than mine, or anyone else's room I've experienced, I'm honestly not sure what conclusion to draw about how the speaker would sound in a more normal room. On the other hand, the fact that the Polk's are claimed to throw a remarkable soundstage with only 6-8 feet of separation while being close to the front wall is intriguing for L/R use in HT systems. Most speakers don't throw an awesome soundstage positioned like that. On the other hand, the narrow sweet spot could be equally problematic. I'd like to experience them and make this judgment for myself.

A couple of us had several questions about the amplifier terminology used in the review, and I'm still hazy on what to conclude, other than I'm a stickler for precise terminology. That discussion got a few of us curious about how an amplifier's output stage topology could affect the output of the SDA arrays, which led to an investigation of just how the SDA circuitry worked. I think we figured that out, though the precise nature of the reason for why the SDA arrays are affected with certain amplifiers remains a mystery, IMO, and the mystery is muddied by the lack of precise terminology. I also think it could be a significant question to answer for prospective buyers. I think the review could be tightened up in this area, and if that's a problem for anyone, so be it.

As for the patent itself, as someone who had read and often reviewed for filing hundreds of patents in a long career, I can say with confidence that the patent text didn't please me much. But that's neither here nor there. Those were editorial comments about the prose, and the real objective was to figure out how the SDA circuitry worked, and I think that objective was met, thanks to ski2xblack.

In fact, there is only one person here, TLSGuy, who has said anything negative about the speakers themselves, and in my opinion TLSGuy is perfectly capable of defending himself and usually does.

 

[shadyJ]

Since I think I'm being included in this group of gentlemen, I would like to make sure my views are being represented properly. I am not negative on the speakers themselves, since I've never heard them. Being negative on a speaker without hearing it is silly. In fact, I'd like to hear them, if anything to see what all the hubbub is about.

I did have questions about Matthew's room. He answered them. His room is so different than mine, or anyone else's room I've experienced, I'm honestly not sure what conclusion to draw about how the speaker would sound in a more normal room. On the other hand, the fact that the Polk's are claimed to throw a remarkable soundstage with only 6-8 feet of separation while being close to the front is intriguing for L/R use in HT systems. Most speakers don't throw an awesome soundstage positioned like that. On the other hand, the narrow sweet spot could be equally problematic. I'd like to experience them and make this judgment for myself.

A couple of us had several questions about the amplifier terminology used in the review, an I'm still hazy on what to conclude, other than I'm a stickler for precise terminology. That discussion got a few of us curious about how an amplifier's output stage topology could affect the output of the SDA arrays, which led to an investigation of just how the SDA circuitry worked. I think we figured that out, though the precise nature of the reason for why the SDA arrays are affected with certain amplifiers remains a mystery, IMO, and the mystery is muddied by the lack of precise terminology. I also think it could be a significant question to answer for prospective buyers. I think the review could be tightened up in this area, and if that's a problem for anyone, so be it.

As for the patent itself, as someone who had read and often reviewed for filing hundreds of patents in a long career, I can say with confidence that the patent text didn't please me much. But that's neither here nor there. Those were editorial comments about the prose, and the real objective was to figure out how the SDA circuitry worked, and I think that objective was met, thanks to ski2xblack.

In fact, there is only one person here, TLSGuy, who has said anything negative about the speakers themselves, and in my opinion TLSGuy is perfectly capable of defending himself and usually does.

Point taken, Irv, and yes, try to give these speakers a listen if you get the chance.

As for the way the review deals with the SDA array versus certain amplifier topologies, keep in mind that Matthew was the ONLY reviewer who even picked up on this issue. No other reviewer understood the speaker well enough to consider the matter at all. Many other reviewers had some bizarre and contradictory comments on the nature of the speaker in light of our experience. We did some blind testing on the speaker with and without the SDA effect, and that is how Matthew identified the issue. I would say a lot of other reviewers heard the sound of their own expectations. One of the things that helped us to identify the issue was that we approached the L800s a bit skeptical of the SDA technology ourselves. When I first heard the speaker, I didn't think it imaged better than normal speakers, but that was before Matthew realized that the amp he was using wasn't really compatible for the SDA effect.