• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Why would anyone use an AB amp for a tweeter?

KEF now has 4 speakers in the active category !?
Nope, more than that! There are at least three variants of KEF LSX extant, and LS60... and at least one other...
 
Nope, more than that! There are at least three variants of KEF LSX extant, and LS60... and at least one other...

Oh I thought that currently they produced LSXii and LSX LT and LS50 wireless and LS60 :) and there are older variants of LSX and LS50 wireless? :) I missed some .
 
Didn't wade through this entire thread but am posting my 2 cents, since this is something that I have considered myself (class-AB on tweeter, class-D for all other bands in a multichannel active speaker).

Cheap-er class-D amps, or maybe I should say those that use older technology, often have frequency response problems above 3kHz due to the interaction of the output filter with the load impedance. It's only relatively recently that better amps from Hypex and Purifi have become load-independent "up there". So you can get away with a less expensive class-D amp using older technology for all bands except the tweeter. Here is where class-AB would work great, since there is far less load interaction even with relatively low cost implementation. And of course the power demand in that band is lower as has been mentioned several times.

Sorry if this point was already brought up, but I am in a rush to do some Xmax related stuff!
 
Didn't wade through this entire thread but am posting my 2 cents, since this is something that I have considered myself (class-AB on tweeter, class-D for all other bands in a multichannel active speaker).

Cheap-er class-D amps, or maybe I should say those that use older technology, often have frequency response problems above 3kHz due to the interaction of the output filter with the load impedance. It's only relatively recently that better amps from Hypex and Purifi have become load-independent "up there". So you can get away with a less expensive class-D amp using older technology for all bands except the tweeter. Here is where class-AB would work great, since there is far less load interaction even with relatively low cost implementation. And of course the power demand in that band is lower as has been mentioned several times.

Sorry if this point was already brought up, but I am in a rush to do some Xmax related stuff!

Yep it’s has been brought up , so nowadays you can get class-D modules that’s relatively load independent not only from hypex and purify .
KEF has compounded the normal issues by designing the LS60 with an upper fr response of 36kHz .

And typically class-D modules are limited to ca 20kHz for the filter and other design considerations ( Bruno P has stated that he rather optimise inside the audible band than make the designs more wideband ).

So I Guess there are limited suppliers that provides a 40kHz class D module , it’s not only the filter but also switching frequency et all involved ? So you then have to do some in-house development or contract that to someone else . I would further guess that developing a traditional small class AB amp is known entity, you can probably find it in a textbook.
 
Simple they used AB class for HF and now they don't in most cases and big suprise the self noise went up.
This is so far from the truth.

1734948787985.png

1734948827983.png

KH 120 A:

1734948884422.png
 
Thank you for your explanation! It starts making sense. I never realised before that at the tweeter frequency class AB amps perform very well compared to the best class D.

It even shows (too some extent) in the Hypex Nilai 500 vs the NAD C316 BEE.
Fascinating. It's like my work. When you get to a very deep understanding of the technology, and somebody asks for the best solution, the answer almost always start with: it depends...

View attachment 415734

View attachment 415735
Interesting examples.

First the Nilai above maxes out worst case at -80 dB while the NAD is around that range entirely. Looking the other way, the Nilai is better in all other frequencies and powers other than 2 highest ones.

Second, as I have written a thousand times plus wrote a FAQ about it, the Nilai (Hypex and Purifi in general) did this on purpose. Better to have the feedback loop strongest where it is audible and let go where it is inaudible (2nd/3rd harmonics of 10 kHz are 20/30 kHz, of 15 kHz are 30/45 kHz, etc). Especially around 10 to a 100W in a tweeter, where musical signal is dozens of dB below full scale.

These examples just show good engineering at its best: solve a problem where it matters. :)
 
Last edited:
Perhaps quite simply, the size, weight and heat release become problems in a space so constrained as a small acoustic enclosure, even if it is a column...

When we see the size of the cooling radiator of a class AB amplifier making 500 watts, when we see the size of the transformer to power it and that of the filter capacitors (which can admittedly be replaced by a switching power supply lighter)... I still wonder if resorting to class D is not a necessity...

Because the two amps producing the most power and the most heat are those dedicated to the bass and the midrange... and this Kef precisely contains a 500 watt amp for the bass and one of 100 watts for the midrange in class D ...

And when you produce on a large scale, the weight of each packaged unit is not such a negligible data...

So the reason would be economic-technical-logistical

And the 100 watt amplifier for the tweeter is not expected to deliver more than 2 or 3 at medium speed... so it does not need to be cooled as if it were amplifying the complete signal
 
Last edited:
I have tried A & AB with 4- AMT ESS H.F. drivers and subjectively the SET amps sound much better with AMT's.
 
I would say whatever KEF’s done with the LS60 I would not say that self nois is a problem IME , have had other active speakers hiss more like Adam or Fostex .
 
All 4 amps are in the Primary speaker in the LSX II LT. The secondary speaker doesn't have a power cord connection the only connection is the usb c tether cable carrying the audio signal from the primary. They're probably the same amplifiers only difference is the LSX II has them in each speaker so they can be used wireless without the tether connection cable.
Having caused the initial confusion, I've also been reading. The Australian website actually clearly states differently: for the LSX II LTthe secondary speaker is active, powered from the primary speaker. The LSX II is active with each speaker powered. It couldn't be any clearer:




1734957895485.png
 
Just out of technical fetish (far,FAR from audibility) it's reasonable for one to want an AB class amp up high,specially a low power one as the SINAD list top champions which against class D they maintain a straight line at the abysses of the chart up to ultrasonics.

However that's not the case with Kef amps I suspect as this straight line can probably be a lot higher.
That's how the myths sustain though.
Until we see a Purifi or a (as they promised) icepower Conductor doing the same up there the myth will go on.

And of course all that are technical,again,far from audibility.
 
And the LSX II LT manual doesn't say anything about where the amplifiers are, at least not directly.
What it does say is that the signal sent to the secondary speaker is upsampled to 96/24, indicating a digital signal rather than analogue signal to the drivers.
 
And the LSX II LT manual doesn't say anything about where the amplifiers are, at least not directly.
What it does say is that the signal sent to the secondary speaker is upsampled to 96/24, indicating a digital signal rather than analogue signal to the drivers.
True, I made an assumption and it's probably wrong, not the first or last time, I see the interspeaker USB c cable for the LT also provides power via PD3 as well as the digital signal.

A1.10. Can I use my own USB-C cable for inter-speaker connection?​

We do not recommend using USB-C cables from other manufacturers for LSX II LT interspeaker connection. The supplied interspeaker cable has proprietary configurations to meet the performance that we aim to achieve.

As there are numerous commercially available USB cables with different configurations, it would be hard to determine if certain cable models work with LSX II LT, ensuring no safety concerns.

If you really want to use your own cable, make sure that the cable supports USB 3.0 or above, and PD 3.0 or above

Note that KEF will not be liable for malfunctions or property damage caused by third party cables.
 
Last edited:
Perhaps quite simply, the size, weight and heat release become problems in a space so constrained as a small acoustic enclosure, even if it is a column...

When we see the size of the cooling radiator of a class AB amplifier making 500 watts, when we see the size of the transformer to power it and that of the filter capacitors (which can admittedly be replaced by a switching power supply lighter)... I still wonder if resorting to class D is not a necessity...

Because the two amps producing the most power and the most heat are those dedicated to the bass and the midrange... and this Kef precisely contains a 500 watt amp for the bass and one of 100 watts for the midrange in class D ...

And when you produce on a large scale, the weight of each packaged unit is not such a negligible data...

So the reason would be economic-technical-logistical

And the 100 watt amplifier for the tweeter is not expected to deliver more than 2 or 3 at medium speed... so it does not need to be cooled as if it were amplifying the complete signal
This is bridged 3.4 KW, tho stereo is far less impressive 950 Wx2.
It's G class PA amp not aiming at star's but just good enough, SoM's can't handle even close. It represents design example on scaling basic AB with many transistor's (take a look at lower end from same series) perfectly suitable for s full range and with built in DSP (PEQ). Even prices are not that bad given how primary transformer's are toroidal. Design point whose long run harsh environment and arguably even serviceable. So it's not topology at first place but design goals.
 
Back
Top Bottom