Well Carl, you will love the CIA room at RMAF! They just announce their new D-500 amp will debut there. $4500(?) pair, 500watts into 4 ohms, can play into 1ohm all day. Two UcD400s running together. It is not a UcD700 module. Nobody has heard it yet, so should be fun. He demos with Von Schweikert. I've never heard either brand!
I think the high frequency issues are more apparent when listening through very revealing speakers. Being that the price is more accessible on some of these switchers than equivalent spec linear amps, they will tend to paired with more flattering speakers that might not tell all. And without the audible screech, they will sound pretty good and continue to grow in popularity. High enders' intrigue will continue and when someone makes a great one, look out!
So T amps are switchers - A HA! Now I know. Tripath, ICE-Power, and Hypex are the three major topologies, with a couple others built up discreet. These three brands are switchers, all make modules for manufacturers to incorporate into their own designs. A lot of amp makers have dipped in their toe, some got their toes bitten off, like Cary, ARC.
Chip amps I think are all-in-one linear solid state amplifiers on a chip. You provide signal, power, and cooling and they do the rest. Some were developed for high end car stereo use and sound damn fine! Read about Charles Altmann's BYOB amp for some gushing customer reviews. They have brutal overload characteristics but can be bridged, etc. They can handle a lot of current.
As I understand it, a linear amp (transistor, tube, chip) draws the output waveform by varying voltage directly and proportionally to the input signal, by direct amplification. It is a linear gain. A digital amp, or "analog switching" amp as their marketing departments prefer to be called is not linear. The output device (transistor) draws the output waveform in quantum steps of equal +/- voltage gain magnitude which looks like sawtooth. The switching frequency is fast enough to allow the comparator controlling the output devices to draw frequencies beyond audible 20kHz. A filter after the output device washes off the sawtooth frequency leaving a very accurate final output signal.
Hypex in particular is, on paper, and on test bench, nothing short of awesome. For a true audionervosal type, this brings out the perfectionist like no other amp. If someone can get it to sound good, we will all be using them. But even Hypex acknowledges that the original modules could be improved and they have done so with their new versions which use audiophile grade parts based on DIYaudio gang's free R&D service. CIAudio also specifies stripped modules which he finishes with his own parts. Being a new technology, we can envision trying out new versions of output devices, power supplies, comparators, input stage chips, signal caps, everything has to be worked out from scratch from the audiophile's perspective. They are gradually taking over the low end of $300 home receivers, and will eventually move up the ladder as always happens.
My personal opinion is that the switching frequency has to go up up up. Transistors can handle high freq, but maybe not at high power? Hypex seems to think their 350kHz is fast enough, and maybe it is. Seems like the faster the switching frequency, the lower amplitude the switching "fuzz" will be, to the point that
a small choke or some audio transparent filter could be used, or not even necessary. Of course, same goes for digitally stored audio data. With PCs running at 3GHz in every home, audio sampled at 44.1kHz seems archaic!
Rich
