Really depends, but most Class D amps do pretty well with these. Hypex nCore is especially well suited to low impedance loads.
Can you say RUNNING TOO HOT! It may seem like a good idea but in the end the amp will run too hot and if played this way overtime will burn up. Infinity speakers are known for tearing up amps that are incapable of dealing with the impedance swings. I would try something more like a good mosfet design, like a ADCOM 555, or even better 2 of them bridged!
Mitch, are you using the speakers with their "extended/normal" switch in the "extended" position (which iirc seemed to be the most commonly preferred setting), or in the "normal" position?
In the "extended" mode the Kappa 9 goes down below 0.8 ohms at multiple bass frequencies, where lots of energy is typically required. The "normal" mode is not quite as severe, but is still very challenging.
Also, regarding the mention that was made of operating Adcom 555s in bridged mode, keep in mind that a bridged amplifier will "see" a load impedance equal to the speaker impedance divided by 2. In other words, less than 0.4 ohms at multiple bass frequencies if the speaker is in "extended" mode.
I play them in extended mode. I have a very big room.I would advise caution in extrapolating from the results others report with different speakers!
Even with the Scintilla. Not only does the impedance of your speakers go a bit lower than the impedance of that speaker at some frequencies, but I wouldn’t be surprised if the phase angles of the Scintilla’s impedance may be less demanding.
Anyone have experience running them with Quad electrostatics, which go down to 2 ohms in the high frequencies. Considering for a pair of stacked quads.Most Class D amps behave as voltage sources, which is to say that they can double power as the impedance is cut in half.
This does not work in your favor in the case of ESLs, the reason being that the typical ESL has an impedance curve that varies by about 10:1 from the bass to about 20KHz. In addition, the impedance curve is not based on a driver in a box as you know.
When the impedance curve **is** based on a driver in a box, the impedance curve is also a map of the efficiency of the speaker; so for example the resonant peak of the driver in the box will be a peak in the impedance curve.
ESLs have their impedance curve based on a capacitor so its not a map of the efficiency of the speaker. IOW, the speaker has the same efficiency from the bass all the way to 20KHz or wherever it rolls off.
We all know that solid state amps and also class D amps can often double power as impedance is cut in half. Think about that happening over a 10:1 range- the amp will be making about 8x more power than it should at high frequencies. This is why solid state amps often sound bright on ESLs.
When a good solid state amp with good current ability, drives a wild impedance curve, like many esl’s have, and speakers like Wilsons have, it will stay close to flat across the frequency range, giving least colouration.
Where a many a tube amp and some mosfets will start to behave like a tone control giving a far from flat frequency response, giving more colouration to the sound.
This is why with Martin Logan ESL as I have, which go down to 1ohm in the high frequencies can sound recessed in the highs with tubes, but with good solid state with good current drive the high are better defined not recessed and extended. Same applies to the bass if low in impedance.
This is proved many times by Stereophile’s amplifier tests into simulated speaker load conditions.
When a good solid state amp with good current ability, drives a wild impedance curve, like many esl’s have, and speakers like Wilsons have, it will stay close to flat across the frequency range, giving least colouration.Because of the inclusion of ESLs in this statement, it is false. The coloration of solid state and ESLs is brightness, and often accompanied by a lack of bass energy although the extension is usually there.
Because of this fact, for best performance the match of amp to speaker is something you have to be careful about with ESLs.
This is proved many times by Stereophile’s amplifier tests into simulated speaker load conditions.There is an inference here that a simulated load pertains to an ESL; it does not. It pertains to a box speaker; Stereophile is not claiming that the response on a simulated load translates to anything regarding an ESL. The simulated load does not represent an ESL panel in open air. Its an entirely different technology! As I pointed out earlier, in a **box** speaker the impedance curve is also a map of its efficiency at any given frequency.
This is not true of ESLs which do not have a resonance of a driver in a box.
Now in the case of Martin Logan, they are trying to capture the largest market possible for their speakers. To this end they have set the bass impedance of their speakers at 4 ohms, which means they are in the neighborhood of 0.5 ohms at 20KHz. Naturally no tube amp can play that properly, neither can most solid state amps. At this impedance, the speaker cable becomes a significant portion of the output impedance of the amplifier. Most solid state amps can't continue to double power as impedance is decreased to this low level; the idea is to prevent the constant voltage character of most solid state amps from actually working.
It sort of works IMO, but not very well. There is no point in making any amplifier work hard as distortion is the result, and ESLs are pretty transparent. The type of distortion likely to show up is higher ordered harmonics, which will contribute to brightness and harshness. So while there are some benefits from this approach, it is not without tradeoffs.
Ralph, since since the 70’s I have had nothing but mostly ESL’s, Stax F81, F83, Acoustat 2, 6, 2+2, 1, 1+1, Martin Logan CLS, Monolith, Monolith III.
Tube amps I’ve had Triode, Ultra Linear, Pentode, Tetrode, SE, Paralleled SE, massive Push/Pull, some built by me some not, but heavily modded.
Nothing drives these ESL’s as well as a good bi-polar solid state with good current ability, period sorry.
And please don't start with the Zero stuff on OTL's
Sounds to me like despite all that time, you've still not unlocked all the performance available.
What you're not taking into account is how variable ESLs are! Old Quads are higher impedance as are Sound Lab and certain Acoustats. Other Acoustats are low impedance and Martin Logan a bit lower.
My point is no one single solution is correct and the math bears it out. FWIW, about 80% of our MA-2 production is on Sound Labs and I too have been working with ESLs since the 1970s.
Please don't put it down to one way and one way only. Such generalizations are simply false.
The OP has Infinity Kappa 9’s who’s impedance drops to
.8 ohm at 35hz
1 ohm at 40hz
2 ohm at 55hz
2.8 ohm at 60hz.
Upper mids lower highs
2.5 ohm at 2.5khz
1.2 ohm at 8.5khz.
All this and we haven’t even taken into account any negative phase angles yet, which will make these figures even more severe, pity there’s no graph on this.
This is clearly a speaker that need bags of current for them to work at their best!!
Ralph, thank you for your excellent explanations, as always. My Martin Logans definitely sound more balanced in the top octaves when driven by my McIntosh MC2200 as compared to my other SS amp that doubles down with halving of the load impedance. I'm speculating that the autoformer design in many of the McIntosh amplifiers is why the Martin Logan pairing with McIntosh gear is very popular. I wonder if a pair of MC601s could be a good match for the OP's speakers.