how does this amp do it?

[Steve Conner]

Yeah!!! I'll order 10.

JM Fahey's comments about compressed bass in PA systems are why I was concerned about the EIAJ rating scheme for Class-D amps. Driving bass bins, they will be expected to make a continuous heavy drone/throb at close to their rated power, not little 8 cycle bursts. Somebody earlier mentioned this as an exception, but I think it would be a common use for these kinds of amps.

JM, if you think the 18kW rating doesn't compare to 18kW RMS of "old style" amps, how many kW would you say? Does the 6kW we calculated from the mains current seem more realistic?

If so, it is still a big improvement, because a 6kW rack of old-style amps would draw quite a bit more than 32A from the mains. They're no better than 70% efficient, and don't have power factor correction.

[J M Fahey]

Really canīt answer with precision, but I *think* the equivalence will be somewhere between those 6KW you found possible and the almost 2 KW implicitly confessed by themselves when they rate their amps as "15.9 A @ 115V".
Take your pick, at the end of the day the winner will get a nice 2000 Lb (PMPO) roasted Turkey. (about 20Lb at the shop counter).
Although I really donīt know why they do that; *nobody* will use one of these at home , and Big Name shows do NOT , repeat NOT rely on commercial power; they have for ages used hired big power generators, as large as a Shipping Container (in fact often housed in one), not because of Sound needs, but because of Lighting.
So they have *huge* Industrial type outlets available around stage.

This is where the Big Guys plug their PA amps.
No puny 15/20A home outlets, plus we use 220V (and since itīs their own generated power, what the "official" line voltage is does not really matter, they can easily have 220V or even 380V triphasic (best for power amps) in *any* country in the World).

This is part of the Amp Rack (this picture is old, still conventional AB-Systems and Crown amps)

This is part of the flown mid/high array (the subwoofers were floor level at the stage sides)

This is my very scared daughter Victoria, when she saw from the stage the "sea" of people, slowly filling the 45000 seat River Plate Soccer Stadium. (really Football all over the World, but being in an American based site, for courtesy letīs use their convention )


EDIT: those "analog" power amps shown are *more* than 70% efficient, because at those power levels *all* are Class F , H or G with quad power rails and different switching systems between them (acording to class).
So thereīs not *that* much difference in power consumption Vs. delivered audio power.
And average power delivered can never be higher than average power taken from the power line.
Just thinking with an open mind, I may accept that the amps *might* be absorbing , say, 2KW continuously from the power line, no matter what, and delivering it in, say, 50% duty cycle floor drum kicks (after all, itīs a *rythm* instrument, not an organ Bass Pedal).
In that very specific (but more related to real world sound) scenario, a 2KW continuous rated "trick" Class D amp may behave exactly the same as a 4KW RMS conventional one, both pulling 2KW average from the outlet.
The conventional one pulling them "on demand" only and the "trick" one, filling a big reservoir, with input signal or not, worth for, say, a couple seconds audio.
That would be a *big* capacitor bank, I leave to anybody interested the task of calculating some Farad and Volt values.
I doubt it would fit in a 1U rack
And if you do not have the raw capacitance, the power can only come from the outlet, on demand, with the severe restrictions mentioned by even themselves.

Every time I think about it, I find it closer and closer to the PMPO mentality.

On December 3, many of my clients and friends will play at "Buenos Aires Rock 4 All" Heavy Metal Festival.
Iīll take some fresh pictures if possible.

[jmaf]

I was a roadie for a friend's band in 1999. There was a riot on one of their shows(small gigs, nothing compared to J M Fahey's photos) because the power company did not arrive to connect the power lines to the stage. Generators: good idea.

[J M Fahey]

Iīve been doing some research, but nowhere to be found the "official" EIAJ spec, although being something invented in Japan to rate consumer (as opposite to Pro) Audio stuff, it stinks of Marketing, not Engineering department stuff.
Close to the PMPO spirit.
Found quite a few "big/thin/light" power amps on line, *all* spec EIAJ, *none* RMS, which stinks for me.
They should, at least as comparison.
I am also *very* pissed off that *nobody*. repeat, nobody publishes important "other" specs(such as THD, power consumption, temperature) at full power, always at "1/8 omax. output power" , at best "1/4 max rated power" which for me is an indirect way of saying that "1/4 max rated power is in fact real power" if you catch my drift.
Finally I found the datasheet for a Philips Car Audio chip.
Philips, being a "serious" company but also trying not to shoot its own foot with respect to possible sales to massive Asian manufacturers, states both in the same page (God Bless Anton Philipsī spirit).

Click on it to make it readable.
And what do we find?
The chip is a quad bridge Car amp, run straight from Car battery, no up converters, so we already know they are good for 16 clean watts, maybe 20W with visible (10%) clipping ... and that at 14.4V, car voltage with the motor (and alternator) running; engine stopped means 12.6V.
The chip is sold as "4 x 40 EIAJ watts", the datasheet 4x16W RMS guaranteed, so the conversion factor is 40/16= 0.4 (40%)
By the way, EIAJ means input overdriven 6dB by a square wave (just to make *sure*the output is a perfect square wave too), into minimum impedance, and at 1KHz only, so *any* piece of cr*p can comply.
So, re-reading the Powersoft specs, 18KW EIAJ mean a more reasonable 7200W RMS.
To be more precise, since they imply that although they provide a 2 ohm rating, you *should* load it with 4 ohms for long life, the real rating is: 2x5200Wx0.4=4160W (total) or 2080W per channel.
Not bad for a 1U amp, but, guys !!, spare the cr*p, will you?

PS: as a side note, their using a square wave, "maximum distortion" rating means that my personal observation of subwoofers heavily compressed and clipping most of the time, was uncomfortably close to reality.
Oh well.

EDIT: so "how does this amp do it?" should be answered as:
"simple, it doesnīt"
PS: 9000 W per channel compared to reasonable 2000W per channel is a 9:2 Marketing/Engineering ratio. We are really in PMPO World here.

[cbarrow7625]

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.

Here you go. I found this document out on the internet when doing some research on Class D amplifiers and how they react under certain signal conditions. I don't know who the author is, although it seems pretty clear that it was written by someone at Powersoft. It is a full performance comparison of some of the most popular Class D amps on the market. There are burst measurements, 2 Ohm performance, THD, and a bunch of other stuff. This paints a pretty good picture of how each of these amps performs overall.

I think this amp manufacturer's site has a pretty fair description of the design approach to class D amplifiers: Sales & supports | D-amp digital amplification . I found it useful.


The discussion about how much "power" an amp can deliver or a speaker can handle is really a red herring. Since, as we all know, the impedance of a speaker is rarely one resistive value the "power" delivered is extremely difficult to determine. What really matters is how much voltage a speaker can handle and how much voltage and current an amp can deliver. I wish we would start specifying amplifiers and speakers based on voltage as the primary value, not power

These amplifiers have high enough DC voltage rails to allow them to generate very high AC audio signals, higher than in a lot of Class AB (or variant) amps. While you can make a good argument that these amplifiers cannot put out more than about 5000W constantly, they also can put out short duration peaks 2-3 times higher than that; something a typical class AB 5000W amplifier can't do. You could make a Class AB amp that does that, Class G & H variants would be the most obvious way to do it. But manufacturers don't build these amps to handle peaks. Why not? My guess is because the amps would just get larger and even less efficient.

Should manufacturers only be allowed to list long term power handling power? I think it is obvious that they should be providing these numbers. But how do they then differentiate their products that can deliver higher voltage for shorter periods of time? Would I rather buy a 5000W amplifier with 3dB of crest factor headroom or a 5000W amplifier with 10dB of Crest factor headroom? I'll take the 10dB of crest factor just about every time. I want the dynamics to come through.

Is it strictly just PMPO? Read the attached document and decide for yourself.

[tedmich]

quite an interesting study! Whats up with the odd "fold back" traces on some of the Lab Gruppen tests?

[jmaf]

quite an interesting study! Whats up with the odd "fold back" traces on some of the Lab Gruppen tests?

There's two distortion figures: one for pre-limiter operation and another with the power limiter turned on. When the limiter kicks in, distortion remains high, so you have dual traces on the same power region, one when the power is without limiter, the other when the power is being limited(the higher line).

The continous line implies a time relationship, the above line happens after the lower one, after something happens at about 2.2kWatts - that "something" is the power limiter.

[Jazz P Bass]

The chart reflects the "foldback limiter" in the amp.
"During long-duration use, the
limitations of the power supply in this amplifier will be
evident. Their fold-back limiter will turn on and drastically
reduce the signal."

[tedmich]

I would have used dual traces; you shouldn't have 2 y-values for a single x-value on a single trace, IMHO

[jmaf]

I would have used dual traces; you shouldn't have 2 y-values for a single x-value on a single trace, IMHO

I agree. IMO they should have been two charts. A separate one with the final trace, where the limiter steps in.

[tedmich]

in my mind the "massaging" of amp power numbers is similar to the computer industry treatment of some data. The HD industry keeps coming out with faster and faster bus systems, ATA133, SATA, SATA 3, Thunderclap etc. and then glossing over the fact that no HD read or write mechanism can come close to filling their data pipes. They have however included larger and larger (16mb->32mb) memory buffers, essentially SS disks, that can provide big buffer transfer numbers (burst speed!), which then have to be filled up with... the same slow magneto data transfer... As long as the numbers keep going UP people stay on the upgrade train and although energy savings do help somewhat I venture the net energy used with new products is almost always positive barring a tech revolution; very rare.

[J M Fahey]

As a real world sample of what type of signals PA systems must handle, here is a clip from one of my clients, Argentine Power Trio "Almafuerte".
The singer/bass player uses 3 of my SS "FAHEY B300" bass heads, driving his old Ampeg 8x10" cabinets (he sold his SVT Classic Tube heads to buy mine).
As you can see, although not a "bass solo" by any means, bass sound is steadily there, never either overwhelming nor dissapearing, thanks to the heavy compressing and waveform rounding built-into my amps.
The bass signal is much closer to a "steady drone" as was described above, than to "1 KHz 8 msec bursts", thatīs for sure.
*Maybe* a Fusion guy, or a Funky player, with lots of snapping, popping and tapping (I have those clients too) would have a low duty cycle, "bursty" signal, but for a Heavy Bass player, using his old trusty Precision Bass and trying to avoid "holes" (silences) because itīs a Trio, no way.
The Fahey B300 are the ones to the right of his head, the ones with white front panels and bright red Leds, and are clearly seen (and heard).

[jmaf]

The Adobe Flash compression on youtube videos doesn't help. My $ 10 speakers also anihilate any chance I had to comment on the bass on that track

[tedmich]

JM that was great! Bass is always there and has to fill a lot of sonic space, like Lemmy when MH was a trio. Love the hands and Hulk guy at the end. Why can't they play stuff like this on "Tengo Mucho Talento"? just ranchero, always ranchero...

[J M Fahey]

Maybe because Mexico shares a border with you, while we are 8000 miles away
Oh well.
And yes, we *do* have a very strong "Argentine Rock" movement, 40 years old and still alive and kicking.
I was very lucky for having started at about the same time and developed in parallel along the way.
It has sure been a powerful driving force trying to meet my customerīs needs.
As an example, one of them bringing to my shop the first Alembic bass in South America (way back then), a couple Stanley Clarke records, and stating: "this is the Bass ... this is the sound .... I need the Amplifier (and speakers)"
I had *no* access to the EVM15L and even more, the VMR 6" mid/highs driver, essential for that bright slappy punchy sound.
In due time, I had to pester my voice coil maker until I got him to make me a few edgewound aluminum ribbon voice coils.
Or the time when another customer brought his Stick, with its own requirements.
Or when a very gifted but poor young Bass player tore the frets away from his Jazz Bass copy, filled the voids with Bondo, and visited me with some Jaco Pastorius records.
One thing is for sure, lifeīs never been dull.