Feel the power! 1.5 KW amp

[argofanatic]

I just came accross this and figured some of you might enjoy this.
ESP 1.5 KW Amplifier.

[rcavictim]

Careful. That 1500 watts quoted is actual toaster watts. In more customary consumer watts (watt is claimed by manufacturers of overhyped consumer audio products in big box stores that magically output more power than they consume) that would be 15,000 watts!

Actually I am in the process of gathering bits on my lab bench to make a 1500 watt amp that will do this into 8 ohms with just two output devices, but they run at about 3400 VDC, and being tubes, may not warm the hearts of folks in this forum who prefer their amps to radiate solely in the long wavelength, completely invisible infra red.

[tiltedhalo]

Not big enough, got to add another 6 pairs of outputs then bridge the dam thing over 3KW all you need is a decent sub station

[Eva]

There are much clever ways to do it, namely class G and class H:

The Class - H Amplifier

Starting a High Power Design. . .

There is also class D, but I'll leave it apart because it's seen as pure vodoo by conventional "linear" minds (who think amplifiers only in terms of big toroids and bulky storage capacitors).

[Nigel Goodwin]

Originally posted by Eva

There is also class D, but I'll leave it apart because it's seen as vodoo by conventional "linear" minds (who think in terms of big toroids and bulky storage capacitors).

You still have "big toroids and bulky storage capacitors" in class D as well, although not quite as big as class AB or B amplifiers. It's converting the PSU to switch-mode that gets rid of the "big toroids and bulky storage capacitors", and you can do that on a linear amp if you like?.

[Eva]

I mean that class D power consumption is ridiculous. For example, some days ago I measured a bare 5W to 15W of input power being used while putting +/-28V peak of music (loud disco/house) into a 8 ohm two-way speaker with slight clipping, and the gate drive circuit was already using 3W of that.

Class G and class H are also remarkable improvements, as power consumption and dissipation with music can be easily halved in comparison with plain class B.

It's a bit pointless to give more or less merits to amplifiers depending on how much iron, copper and electrolytics they have inside.

[richie00boy]

I usually like your posts a lot Eva, but how can 100W into a load draw only 15W from the power supply? Have you solved the world energy problem?

[Nigel Goodwin]

Originally posted by Eva
I mean that class D power consumption is ridiculous. For example, some days ago I measured a bare 5W to 15W of input power being used while putting +/-28V peak of music (loud disco/house) into a 8 ohm two-way speaker with slight clipping, and the gate drive circuit was already using 3W of that.

Then you must have been measuring wrong! - class D is more efficient than AB, but NOT more than 100% efficient, so it has to take more power than it outputs (as I'm sure you know!).

[Eva]

Remember that music signals don't ask the amplifier to output maximum voltage amplitude to the load all the time. Average power content of music signals is usually between 1/5 and 1/20 of what it would be with a pure sine wave of similar amplitude (we would become deaf otherwise!). Note also that class B circuits start to become a bit efficient (p_out > p_loss) when output voltage is above 50% of rail voltage, but with music signals that happens during a small portion of playback time. In practice, this means that power efficiency of plain class B amplifiers fed with music is around 15% (I wrote a program to analyse wav files in order to obtain that figure, and it didn't simulate a reactive load, just a plain resistor).

Also, remember that loudspeaker loads are reactive. This means that a fraction of the energy delivered by the amplifier to the speaker (sometimes over 25%) is returned back to the amplifier (nothing strange there, just plain capacitance and inductance). That energy is completely dissipated in the output devices for plain class B systems, but class G and class H systems manage to dissipate only a fraction of it, and class D systems return all that energy back to the power supply (indeed that concept may be a bit shocking for people used to amplifiers that can only draw current from the PSU).

That's why I say that class D stuff is seen as vodoo by most people. You can't understand it well before previously understanding a lot of general amplifier concepts (that most people seems to ignore, since they are too busy dreaming of big toroids and supply capacitors).

[real]

Remember that music is not a sine wave - it may be clipping only on very high peaks, but the average power is much lower.

Still, the figures quoted seem to be too low (especially 5W).