Wow. I go away for a few days and lookit what happens.

Digital amplifiers operate by definition by switching fully on or off quickly enough that the average output over some time constant is roughly equal to the desired output.

The switching 'on duration' and/or frequency can be modulated to accomplish this. The practical effect is a variable duty cycle.

An unavoidable side effect of this switching is switching noise. This will be manifested as a tone or range of tones above the signal frequency. If the load is known and/or guaranteed to be inductive the amplifier can be configured to safely drive the load directly, but if the load has or may have significant resistive or capacitive components, especially near the range of 'carrier' frequency(s) an inductor must be placed in series with the load to filter the switching current.

If the carrier frequency(s) is sufficiently higher than the signal frequency(s) the filter may be designed such that it is well out of the way of the signal, but if they are anywhere near each other this may not be possible. In this case the filter must be carefully matched to the load such that it is effective without distorting the signal. If this is still not possible it may be possible to alter the input prior to amplification in a manner complementary to the filter effect.

If this carefully matched system is exposed to unexpected loads the upper frequency response will tend to droop as load decreases. This effect can be observed in this ICEpower document:

http://www.icepower.bang-olufsen.com/sw1166.asp

The frequency resonse is down 1-2 db for 2.7 ohms. Some loudspeakers have made it to market that dip lower than 1 ohm.

Now it is true that this would be exceptional frequency response performance for a loudspeaker or maybe a few other classes of components, and the linear distortion is minimal when taken in a holistic 'impact on system' view, but as consumers we are accustomed to better from our amps.

Of course the amplifier may drive such a load without obvious distress from a power distribution sort of view, but this isn't power distribution.

The folks at Acoustic Reality and the reviewers at High Fidelity both assert the eAR has this fixed, but haven't backed it up with any published tests or IP that are germane to the issue. The base ICEpower module is better than most, but still exhibits the issue. So maybe it's fixed, but I'm a 'show me' kinda guy.


SLL:



Not needed. You are also entitled to free expression. Nice of you to say though.

I suspect AzRyan is probably a very nice and personable fellow in real life. Real life interpersonal skills are often not easily mapped to a forum like this and it should be expected. Let it roll off your back, and try to concentrate on any content that may exist. I've learned a few things from him and I'm thankful for it. I'd be thankful for more relaxed presentation, but I'll take what I can get. If we were all the same the world would be a boring place.

EDIT:

The massive improvements in this area as represented by the ICEpower module and others make me think the future is bright for class 'D' audio, but for me it's got to get a bit more mature before I'd jump, that's all.

[This message has been edited by charlie (edited December 21, 2002).]

[This message has been edited by charlie (edited December 21, 2002).]