how does this amp do it?

[tedmich]

Powersoft K20: 1U amp that weight 28lbs and produces 18kW into 4ohm load, 120VAC supply = 150A?

K20 | k-series | products

$10k! Magic?

[g1]

Some kind of burst/peak spec.? It lists 15.9A@115V power consumption at 1/8th power into 4 ohms. There are no amps that can deliver more power (aside from peaks) than they use. Otherwise, we'd be using them to sell power back to the utility companies.

[cbarrow7625]

It is "burst" power. 8 cycles at 1kHz (i.e., 8mS). Class D amplifiers are typically specified using this criteria. It is a specified EIAJ spec.

I found some test reports on the internet published by this manufacturer comparing their amps to other competitive Class D amps. The upshot? They all use "burst" power. None of them can sustain any appreciable power into 2 ohms (or 4 Ohms bridged, which is where the 18k @ 4ohms comes from). This manufacturer performs better in a number of areas than its competitors.

These amps are well liked, well respected and are reliable. It is not by any means a "fake" specification but if you think you can run 18kW of subwoofers using it, you will be sorely disappointed. The 10,400W @ 8 Ohms specification is more realistic if you want to draw power for any length of time, and the long term power delivery is about 1.5-2dB lower than that (around 7k @ 8Ohms).

Still mighty impressive for the size and weight but nowhere near the "top number" spec.

Powersoft does everything by the book. There are no lies or cheats in the spec; you just have to know how the specs are created in order to know how to interpret them. THese amps are awesome top of the line products that lots and lots of pros would swear by. Their premium price is worth every penny, 18kW or not.

[Enzo]

Let's say your toe is stuck under a 900 pound piano. You give the piano a mighty yank and it lifts enough to free your toe. But can you sit there and lift 900 pounds? Probably not.

HOW does it do it? Switching power supply removes the need for huge power transformers. After that, it is just a matter of enough output transistor capability to provide the current to the load.

[cbarrow7625]

And high enough DC rails to hit the peak voltages.

[Steve Conner]

I see PMPO is not dead! :-(

On reading the user manual, it has a 45A mains connector. Also there is a "mains current limit" option in the menu system, and the maximum setting is 32A. That means you get no more than about 6kW total sustained power when running it off 230V, or 3kW off 115V.

By the magic of Class-D you probably get this same power into 8, 4 or 2 ohms.

[cbarrow7625]

Everyone seems to get really aggravated by Class D being specified differently than other "traditional" technologies. It is certainly can be confusing to users and a boon to the marketing departments of Class D amplifier companies. When it comes down to it though, the way Class D amplifiers get specified is actually in very good alignment with the way musical signals (and music power) are transferred to a load.

It all comes down to crest factor (the difference between the average & peak signal of a waveform - or music). Almost all signal have a crest factor (except square waves - any other types Electrical Engineering folks?). Sine waves have a 3dB crest factor, music typically has at least 6dB but usually it is higher like 10dB or more. The "crest" portions of the signal only have to be reproduced for a very short duration. Why should the amplifier need to be designed to deliver long term power for events that are short term by definition? The accepted standard for specifying things like long term power draw by music electronic equipment is to use 1/8 power. Only 1/8! That is a pretty good representation of the actual average to peak content of "most" music material. If you can make a 1000W (long term) amplifier that can also reproduce 8000W peaks, then it can be smaller, weigh less, cost less and still correctly reproduce "most" music accurately.

As I said earlier, you have to be really careful and know exactly what you are getting into if if you want to reproduce long synthesizer tones to heavy droning subwoofer signals. The crest factor of those types of signals can be closer to 1/3 music power. You are not going to be able to draw 2600 rms watts (1/3 of 8000W) out of my 1000W/8000W peak amp described above no matter how hard you try.

In the case of the Powersoft amp in question, the difference between the power (via current limiting) allowed by the mains (6000W / 3000W) and the "burst" power rating (18kW @ 4 Ohms Bridged) is just 4.8dB / 7.8 dB respectively for 240V/120V operation. As I mentioned previously, most folks know better than to try to run amplifiers at a 2 Ohm (or bridged 4 Ohm) rating. If that is taken into account, the real difference is only 1.8dB / 4.8dB respectively between the mains & burst audio ratings. Even the 7.8dB difference is within the "real world" crest factor range for most music. The fact that real world application brings both differences down to below 5dB makes it absolutely reasonable that this amp will reproduce most music correctly even at its spec sheet rating.

Most of the time you are going to be fine using Class D amps as specified without giving any thought to the time constant issues that are really involved behind the scenes. Unfortunately, it is not in the best interest of Class D amplifier manufacturer to point out to customers that they may not get the "spec" power under demanding circumstances. There is very little incentive for amplifier manufacturers to make anything but Class D amplifiers these days (for pro audio & other "mass" music applications) so it will soon become just about the only choice. When that happens, the distinction between the "old" way to spec amps and the "new" way will mostly evaporate. We will all get used to thinking about the parameters in a new way and will have to recognize the benefits / limitations. When it comes down to it, some amps will become known for working well with subwoofers and other amps will become known as better sounding for general applications. As long as they sound good and we, the customers are not disappointed with our purchase decisions (i.e., the amps don't appear to under-perform regardless what our engineering minds and calculations tell us they "should" do), then we will get more of these amps. It all comes down to if they perform the way we want them to - and that they just sound good. Say what you want, Powersoft has done a good job on all fronts (FYI, I have no affiliation with Powersoft - I am just very familiar with theirs and other amplifiers).

[Steve Conner]

If "most folks know better" than to run a PA amp at its 2 ohm rating, why does the manufacturer even give it a 2 ohm rating?

The American Federal Trade Commission tried to crack down on inflated "music power" claims for home stereo equipment. They specified a "RMS" (actually sustained average) power test, but with a 1 hour preconditioning period where the amp would drive one-third of its rated power into 8 ohms. Amps designed with massive "music power" in mind would burn up or shut down during the preconditioning, and wouldn't even make it to the power test.

Amp makers didn't like this, so they ignored it and used one-eighth power for the preconditioning instead, in order to keep on selling junk. (Or amps with power ratings that "better reflect real-world use", according to your view.)

Class-D amps tear up the rule book because they are theoretically perfect, non-dissipative power converters. They don't have the dissipation maximum at one-third power. So there was an opportunity to come up with a new way of specifying them, and it's sad, though understandable, to see that it turned out much like PMPO. It will just fuel the myth that the old Peavey amps have "bigger watts", or whatever.

Don't get me wrong, I'm sure this is a serious heavy-duty amp that could destroy any conceivable speaker system. But it's not an 18,000 watt one. The answer to "how does it do it" is, as usual where marketing is involved, "It doesn't"

[cbarrow7625]

Amp manufacturers give an amp a 2 ohm rating because they can. I think I have been pretty consistent in fully recognizing the part that marketing departments play in misleading the consumer. Unfortunately it is not the job of the marketing department to fully explain the engineering limitations of the product, quite the opposite. The products that are typically most respected by knowledgeable users come from companies that have engineering departments that work to explain the products in detail, limitations and all, through White Papers or other technical presentations.

Music power in home audio equipment is a little bit different because home & car audio (as well as chip amp manufacturers) typically specify the power rating at 10% THD. How they think that is reasonable, I have no idea. I don't know enough about the FTC crackdown to intelligently comment on the particulars but I would assume that the crackdown came more for that reason than because of claims of "burst" power. Those manufacturers also typically specify amplifiers based on theoretical power dissipation conditions (i.e., attached to a perfect, enormous heat sink with no thermal resistance). Most chip amp spec sheets will explain somewhere within that the actual usable real world power is less than the stated max power. The home stereo & car guys just publish the theoretical power @ 10% THD as the spec for their piece of equipment. Yep, that would equal blow up city if you try to run it with heavy loads. In those cases, I mostly blame the marketing departments for the chip manufacturers. The misleading specs just get propagated by the "manufacturers" (re-packagers) of the amplifiers. I would love to hear more about the details of the crackdown, especially if any of my assumptions there are incorrect.

For amplifiers that are used for laboratory testing, or in the medical community, or as drivers for industrial equipment (DC to 100kHz type amps) they absolutely need to be specified in an extremely simple way based solely on long term sine wave testing. Audio amps just don't run with those duty cycles.

Comparing what is going on in home audio with pro audio is mixing apples and oranges. In this day and age in the pro market you cannot get away for very long with publishing false power numbers - people get dissapointed and stop buying your stuff. Check out the July-September issues of Live Sound International for a series of articles based on testing a variety of class D amplifiers done by a VERY independent 3rd party, Pat Brown of Synergetic Audio Concepts (SynAudCon). Pro class D amplifiers have their limitations, but unlike home audio amplifiers, the performance specifications are published based on actually testing the amplifiers, not on figment theoretical power and extremely high distortion. They actually hit the numbers they say they will hit, yes some of those numbers can only be hit for short durations, but that is they are all playing by the same EIAJ rules.

Music is not 100% duty cycle, so amplifiers for reproducing music don't necessarily need to be designed for 100% duty cycle. Don't assume that all specs come from lying marketing info. The amplifiers themselves are just more complicated. That is a result of customers wanting amplifiers like this. Lighter, smaller, cheaper, but still get the job done just as well or better than the behemoth amps of old. They are built correctly for the application.

[Steve Conner]

Well, on the one hand, you say pro audio manufacturers can't get away with publishing false power numbers. On the other hand, you say that nobody believes the 2 ohm rating. And we have a 18 kilowatt amp that can never draw more than 6kW from the mains. Surely you agree that these statements can't all be true.

When I'm bench testing an amp, be it hi-fi or PA, I expect it to make its 8 ohm rating on a sine wave test indefinitely. I don't test the 4 ohm rating for more than a minute or so, because it's more than likely "music duty". And I wouldn't test at 2 ohms at all in case it blew up.

I'd like to know how the EIAJ test for class-D amps is representative of real-world signals. What frequency is it 8 cycles of? The main use of Class-D amps is for bass bins, so it should be something like 80Hz, not 1kHz. I'd also appreciate if Powersoft had published one of those white papers with a graph of power output vs. tone burst length. They suggest that it is limited by the current-time curve for a Type C breaker programmed into the DSP, the idea being to avoid blowing breakers, but I'd like to see it set out explicitly.

There ought also to be some way of relating the published output power to a rating of speaker system that would be a good match for it. Speakers are rated by a similar process, a 500 watt woofer can not really absorb 500 watts of actual continuous power, but it's supposed to handle a "500 watt" amp driven to the edge of clipping with program material. Does that still apply to Class-D amps rated under the EIAJ scheme? Given that the typical 18" driver is rated around 500W, are Powersoft intending you to drive 36 woofers with their amp?

[tedmich]

I've always thought this mackie ad was pretty funny

[Steve Conner]

Yeek, a "1 nanosecond brown noise burst".

I'm not trying to imply that modern Class-D amps from reputable brands are somehow bad. (If I wanted to prove that, I'd investigate the THD figures.) It's more of a philosophical question, as to what the power rating actually means. Imagine a guy designing a big PA system for a concert, should he consider one of these new 18kW things as equal to 18kW worth of the old Peavey CS series amps he was used to?

[R.G.]

Here's another way to look at the power available from any device.

You can't get out more than you put in, minus losses. Losses can never be zero, let alone less than zero.

But you can store it up and use a big pulse of stored energy for a short time.

The AC power line can only supply 15A for most USA power lines. You can get 20A if you have specially installed branch power circuits and 20A rated breakers and outlets. If this thing is rated for 15.9A at 115V (hmmm... wonder why it's that number? ) then the average power out cannot exceed 15.9A*115V = 1828W over any period of time bigger than the internal energy storage.

But mankind has a long history of using energy storage devices to deliver short pulses of peak power. The first one was probably a rock or a stick. Accelerate one of these over a long(ish) period of time, and when it hits something it delivers its kinetic energy and momentum at the rate which the rock/stick and target crush each other, which may be very short. The peak power in the delivery pulse is much higher than the average power which accelerated it.

We got a more controllable storage/release mechanism when some bright guy invented the bow and arrow. Conceptually, even fire is the use of stored solar energy in a sudden (compared to the storage time) release.

A modern day analog uses electricity stored in capacitors to fire 100MJ railgun projectiles and laser bursts. It is extremely easy to make a nitrogen laser which has a peak burst power of megawatts. However, the nitrogen laser inherently quits lasing after about 10nS IIRC, which runs the average power down a bit.

But none of these can exceed the average power input. Neither can this amp. It cannot exceed 1.8kW per second. Come to think of it, the Mackie "specification" of 100kW for 1nS into 0.0001 ohm comes remarkably close to a railgun.

[J M Fahey]

In a PA system the Bass bins will be either overdriven a great part of the time (say 20% of the time) or , more probably, heavily limited/compressed/processed/whatever (hey, we want punch, donīt we?) so the REALISTIC (real PRACTICAL use) signal they are called to reproduce is a big , continuous, almost-full-power-clipping 40 to 120Hz range signal.
Skillfully processed (and I count on that), the continuous power will hover 1 or 2 dB below clipping.
And quite important clipping *is* allowed (and even found desirable by PA operators) because
a) it adds to the punch
b) you can indulge in it anyway, with no penalty, because most *loud* subwoofers are low pass or narrow tuned boxes which acoustically filter clipping harmonics ... not to mention that any which reach listenerīs ears will be heavily masked or just swamped by "everything else" in the music program
So, *for me* , RMS rating *still* is the valid spec.
Specs that expect you to play your amp at no more than 10% its rated power, "to keep dynamic range", are good, excellent ... to listen to a Telarc or Deutsche Grammophon direct-to-disc Symphonic recording (or String Quartets, Jazz Trios and such) but are *not* realistic for anything played at a Stadium.

My headbanging clients customarily play in up to 45000 people Soccer Stadiums (River Plate, Boca Juniors, Velez Sarsfield) in Buenos Aires, or in large open air venues.
I compulsively walk *everywhere* listening, including sticking my head (for a few seconds) in the bass bins.
(Yea, I know, itīs not healthy) , and clearly recognize clipped signals.
Yet 15 meters away (and further) everything melts into a smooth, clean, punchy sound.
Note: most PA hiring Companies in Argentina use these Powersoft amps.
They are highly rated, very trusty and well made ... but nobody thinks 18KW Class D (as rated per the spec posted above) is the same as, say, 10 x 1.8 Kw "old type" amps. No way.
jm2c , as usual.

PS: *any* rating that calls a realistic spec measuring 1 millisecond bursts, with, much worse: a less than 1% duty cycle (to allow for PSU capacitors charging) is , in my book, just plain, unadultered, b*llsh*t.
Thatīs how "Music Power" , step by step, turned into PMPO.
Sorry.

[jmaf]

I'm idealizing my own PMPO marketing, what do think of "JMAF amplifiers are built with 50uF ultralow ESR capacitors which can deliver a megawatt of power during a yoctosecond every night, 365 nights a year."

Does this sound OK with the engineers here?