Class-D amps - a different re view

Maybe I shoud stay out of this one, especially as I do not have time to read all of the posts, but.......

The critcism that not all of the vendors disclosed all of their specs is no reason to find fault in them. Most here have been around high-end long enough to know that specs, i.e. measurements, are just design tools for the guy who design the stuff. (We have been making and selling amps, conventional and switching, for over 20 years. Without once "disclosing" all of the specs. Is there any reason to?)

Yes, a lot of guys like to look at the numbers, but since when could anyone say how something will sound based solely on numbers? If we could, tubes wouldn't be as popular as they are.

Several have already said so, but let me reiterate as a manufacturer of switching amps:

Synergy.

Not every amp will sound good in every system. If they did, we would all own Carver amps driving Bose speakers. But obviously, that is not the case.

Yes, these amps are not for everybody. I do not know of anyone making them that claims such.

Likewise for fine companies like AtmaSphere. I don't see Ralph telling anyone that tube amps are the only way to go. (I only point out Ralph as I have friends that own his products, and they sound good in the right system.)

Martin Collums and his friends may not like them, and I can accept that. To each his own. To dismiss them all, based on his perceptions would not be an enlightened positition. Listen first, then make up your own mind.

To be perfectly honest, I have heard listeners say that our (switching) amps sound "very tube-like".

In the right system.

I have personally heard them sound like crap.

In the wrong system.

Synergy. Simple as that. Need I say more?

Kijanki:

Your post mentioning jitter and PLL seems to be off-base, and probably ill-suited to this thread. Perhaps you would like to clarify it, but may I suggest in a different thread.

RF gumby:

Yes, I think I am in the right place. I am pretty sure I did not hit the bookmark for the Insane Asylum.

Or it might have something to do with my being an RF engineer, partially out-of-place in the role of front man for a minor high-end company. (It fell upon me by default. Long story.)

The "talking points" that were mentioned as proof that Class D amps are by design inferior seem to be a mix of technical points and sophistry. I could make a similar list for any number of other technologies. Equally convincing, equally misleading. If I didn't know better, I would come to the conclusion that someone had an agenda.

However...........some points do have a bit truth to them. Some Class D amp modules have an output frequency that does vary with speaker load. The maker(s) of these modules clearly show this in their data sheets. I would think that anyone who truly has a background in electronics design could understand how changing the impedance a second-order LPF operates into could have an effect on its damping ratio.

I believe the voltage on the output that he refers to is the Vcc/2 in the ICEpower-based units. They operate with only one polarity of power supply, and being a full bridge design, will have 1/2 of the rail voltage on BOTH output terminals.

Translation:

Don't grab your speaker cable in one hand, and ground in the other! Not a good idea. (Maybe that is why so many have shook-proof terminals.)

Also, there will be a certain amount of the carrier frequency on the output. Yes, it may be a few volts. Somehow, I don't know of many tweeters that will generate tons of IMD with that sort of voltage present.

As I said, some technical facts, seemingly massaged to possibly scare off the uninformed.

Kijanki:

If you have a 1000ASP, then it will have much more than 22 V on the output terminals, so I would not advise touching them. (Rowlan uses that module in some of their amps.)

As for the output network, I was talking about the damping factor, or Q, of the filter. Not the damping factor of the amp.

If Colloms believes that they have sampling stages inside, then he is ignorant as to their operation, or else one can conclude that he has an agenda. Modern Class D amps of the self-oscillating style (like ICEpower and Hypex/UcD) are basically an amplifier that is unstable, and use that oscillation to generate a carrier wave. (Unlike clocked designs, which no one seemed to like.) They use some simple circuits, in some rather clever manners, to modulate that carrier with the audio signal. Over-simplified story on how they work.

The answer is "yes", but.........boy, am I going to step into it here..........the module that one would expect to be best for that particular speaker, based on its low impedance and power requirements, really doesn't work that well in that appication. (There is a reason that I know that answer.) Long story, and perhaps best not dealt with in detail in a non-technical forum.

Anyone desiring a more detailed answer should contact me privately.

I have not been following this rambling thread very closely, but perhaps I should clear up some common misconceptions about making Class D amps:

1.) They are cheap to build, and
2.) They are easy to build, because all you have to do is put it in an expensive box.

Well, yes, that can be the case. Indeed, I am sure that there are outfits doing just that. But to say that everyone does that is a gross oversimplification.

Let me give you an example of something similar. I'll let the readers draw their own analogy and conclusion.

About 15 years ago, National Semiconductor came out with a line of parts that has come to be known as "chip amps". We looked into them, to see if they could lead to building a very affordable amp. One that would come close to high-end expectations, without the high-end price.

Well, they sounded good enough to merit consideration, but..............

How much could we knock the price of an amp down by using them?

Depends. Depends on how you want to go about it.

If it has to look nice, you need a decent enclosure. Not cheap. You would need a large power transformer, and decent filter caps. Not cheap, either. Still needs a heat sink, cut down from an extrusion, milled, and anodised. Ok, we can cut down on the size of the PCB somewhat, won't have to buy a boatload of discrete parts and output devices. But how much will that be less than the cost of buying the chip and making a PCB for it?

In reality, not much. You still need the same size power supply. Same amount of heat sink radiating area. After all, it is basically just like any other amp, only with most (if not all) the electronic guts inside one chip.

So, when you get down to it, it is not really all that much cheaper.

In fact, it can cost more.

"Huh? Did he say 'more'? That can't be."

Sure it can. Think about this:

Let's say through "research" that you find that it doesn't work as well as expected with only one chip. There is an awful lot heat to get rid of, and from a small source. Heat sink or not, it is probably going to get hotter than several discrete output devices. And it is all in one chip.

Well, heat is the enemy of ICs. So, you futz around, and through your "research" you find out that for several reasons that it sounds better if you run 4.......or 6........or 8.......in parallel for each channel. Wait......we are back to where we started from: same enclosure, same power transformer and filter caps, same heat sink, and now the same size PCB with the same number of parts. Only 8 ICs cost more than 2 outputs and a handful of discrete components.

Sure, you could cut every corner in the book, and make something inexpensive. Sure, there would be a market for it. Especially if you made a gazillion of them to get the price under $1K.

Or, you could end up with a novelty that cost as much as a traditional amp. Perception being reality for most, how many folks would buy a $3000 amp with a bunch of ICs in it, when you can buy the same thing that actually looks like a real amp for the same price from the same company?

Guess what: we never made any.

OK, fast forward to 2000 and something.........

You can buy a Class D amp module..........cheaply if you buy enough...........and start making amps with them.

But wait! You still need that fancy box. And probably same ol' power supply. Unless you want to go for "novel", and put a switcher in it. You need to market to a different crowd if you go that route.

So, what have you saved? Obviously, no heat sink. And probably not much of a PCB.

You could take those savings and use them for nice touches, like say..........EMI control..........or well, there are lots of tricks that an experienced designer can do at this point. (Sorry, no inside info gets divulged at this point, gang.)

Bottom line is this:

Class D modules are a tool that a designer can use. How they choose to use them is the difference between inexpensive and over-priced. Good sounding, or just good enough. Or something in between.

Yes, you can make an amp without a lot of bother and expense by using them. The opposite is also true.