Efficiency, where's the data?

[Gary Cook]

A post in another thread prompted me to think about about claims by amplifier manufacturers that their amplifier design is more "efficient", meaning more of the power that goes in comes out as sound. For example, Class H power supplies (SMPS) as found in Emotiva XPA Gen 3's are claimed to be more efficient than their previous linear (toroidal based) power supplies. But how much more? Is it worth the substantial price difference? For power limited situations (110 volt 15 amp circuit) is it worthwhile swapping for a Gen 1 or Gen 2 to a Gen 3 and getting more watts out? Class D amplifier manufacturers claim up to 90% efficiency, compared to Class AB amplifiers that struggle to reach 80%. Is that 10% points extra efficiency claim proven in the real world? If I input, say, 110 watts (110 volts at 10 amps) into a Class AB amplifier and get, say, 90 watts out does that mean I will get, say 105 watts out of a Class D amplifier with the same 110 watts input? Yes, I know that watts in isn't the same as watts out, but in a side by side comparison that's not relevant. Similarly I know that it's not linear, twice the power in does not = twice the power out, but again in a side by side comparison it's irrelevant.

It would be nice, to support their claims of higher efficiency, if the amplifier manufacturers (not just Emotiva of course) quoted watts in versus watts out that way we could tell if it's a worthwhile efficiency gain. After all it's not that hard to measure on the bench while performance testing THD, output watts etc.

Cheers
Gary

[AudioHTIT]

I am interested in this as well, and agree that some standard measurements would help in evaluating PS efficiency. What I would like to see (and which would be VERY SIMPLE for all vendors to do now) is for each product of any type to list it’s idling wattage. Much of the time our equipment is turned on it’s doing nothing, it’s this idling draw that ends up being the primary contributor our systems make to our electric bill; and whether you’re interested or concerned about how much power you use (or you feel it’s insignificant compared to your HVAC system), this is useful information that is easy to supply. We shouldn’t have to buy the gear and put a kill-a-watt meter on it to find out this information.

I’ve seen some vendors give this information in the past and some probably still do, but it would be very simple to at least supply:
* Idle Power (power on, no signal) xxx Watts
* Standby Power (power off, remote / triggers ‘listening’) xxx Watts
(I used Watts to account for different voltages, but if there’s an argument for current go ahead and make it)

[audiosyndrome]

Idle and standby power should be negligible unless you're heavily biased into Class A. Most A/B amps idle at just a few watts, as little as one or two. HiFi News magazine does list those measurements plus max power.

Russ

[AudioHTIT]

Oct 1, 2017 14:14:27 GMT -5 audiosyndrome said:

Idle and standby power should be negligible unless you're heavily biased into Class A. Most A/B amps idle at just a few watts, as little as one or two. HiFi News magazine does list those measurements plus max power.

Russ

I would disagree that 50 to 100 Watts idle for most amplifiers is negligible (much higher for Class A). We will each value the information differently, but since it requires very little effort to supply it, why not? It’s great that some reviewers provide this info, but I should be able to find it along with the other specs for each piece of gear.

[Jim]

50-100W describes the idle consumption of most Emotiva amps I’ve measured. All with big iron transformers...non SMPS.

[leonski]

Red Herring issue, IMO. Going from Linear to Switcher MAY net you a few points, but I'd be surprised if you saved A-Buck-A-Year on your electric bill.

Class 'A' amps essentially draw Full Power @idle.

Efficiency @idle? ZERO. Or in a 'no signal' state.

sound.whsites.net/efficiency.htm

As I see it, you have 3 inter-related issues. 1. Output stage efficiency. 2. PS efficiency 3. Total Package efficiency

[garbulky]

My XPA-1 draws something close to 400 watts at idle in class A mode. Aww yes!
My UPA-2 drew much much less.

But these new amps are of course not class A amps.
Come on guys, you are going to spend more than a grand on an amplifier. Who gives a hoot about 10% more efficient?
I sure don't. If you can afford that price for one copmonent, you can afford the $5-10 increase per year or whatever it impacts you.
Being more efficient is a buzz word that doesn't really impact reality.

[pknaz]

Other than curiosity (And I'm all for curiosity) what real world application does this analysis provide?
These are taken directly from the PDF manuals of the amps. They assume 120VAC.
Example 1
XPA-2 Gen 2
300 watts per channel; both channels driven; into 8 Ohms
500 watts per channel; both channels driven; into 4 Ohms

XPA-2 Gen 3
300 watts RMS per channel; THD < 0.1%; into 8 Ohms
490 watts RMS per channel; THD < 0.1%; into 4 Ohms

Example 2
XPA-7 Gen 2
Power output (all 7 channels driven):
200 watts RMS per channel @ 8 ohms (0.02% THD)
315 watts RMS per channel @ 4 ohms (0.2% THD)

XPA-7 Gen 3
Power Output All Channels Driven; THD <0.1% ; into 8 ohms
200 watts RMS per channel


Doesn't seem like their change in power supply has yielded any increase in output power, at least not that their advertising

[Cogito]

Oct 1, 2017 15:28:28 GMT -5 pknaz said:

Other than curiosity (And I'm all for curiosity) what real world application does this analysis provide?
These are taken directly from the PDF manuals of the amps. They assume 120VAC.
Example 1
XPA-2 Gen 2
300 watts per channel; both channels driven; into 8 Ohms
500 watts per channel; both channels driven; into 4 Ohms

XPA-2 Gen 3
300 watts RMS per channel; THD < 0.1%; into 8 Ohms
490 watts RMS per channel; THD < 0.1%; into 4 Ohms

Example 2
XPA-7 Gen 2
Power output (all 7 channels driven):
200 watts RMS per channel @ 8 ohms (0.02% THD)
315 watts RMS per channel @ 4 ohms (0.2% THD)

XPA-7 Gen 3
Power Output All Channels Driven; THD <0.1% ; into 8 ohms
200 watts RMS per channel


Doesn't seem like their change in power supply has yielded any increase in output power, at least not that their advertising

It resulted in MUCH cheaper shipping costs for Emotiva, period.

[leonski]

Oct 1, 2017 15:56:52 GMT -5 Cogito said:

Oct 1, 2017 15:28:28 GMT -5 pknaz said:

Other than curiosity (And I'm all for curiosity) what real world application does this analysis provide?
These are taken directly from the PDF manuals of the amps. They assume 120VAC.
Example 1
XPA-2 Gen 2
300 watts per channel; both channels driven; into 8 Ohms
500 watts per channel; both channels driven; into 4 Ohms

XPA-2 Gen 3
300 watts RMS per channel; THD < 0.1%; into 8 Ohms
490 watts RMS per channel; THD < 0.1%; into 4 Ohms

Example 2
XPA-7 Gen 2
Power output (all 7 channels driven):
200 watts RMS per channel @ 8 ohms (0.02% THD)
315 watts RMS per channel @ 4 ohms (0.2% THD)

XPA-7 Gen 3
Power Output All Channels Driven; THD <0.1% ; into 8 ohms
200 watts RMS per channel


Doesn't seem like their change in power supply has yielded any increase in output power, at least not that their advertising

It resulted in MUCH cheaper shipping costs for Emotiva, period.

You BET!     Delete 20lb or more transformer and replace the entire PS with a switcher.     New PS runs at a very high frequency compared to a linear so needs a fraction of the huge capacitors otherwise needed.  Yet another cost and weight saving.  

Based on numbers, and it could be pretty high, you start saving on shipping, too, and returns, and damage.   Balanced against higher cost of switcher BUT you gain back in economies of scale, having essentially replaced 3 or 4 or even 5 different amps with ONE module and ONE power supply and ONE case.      Once you have the amp and PS module production line up and humming, you're golden.  

[Cogito]

Oct 1, 2017 17:17:52 GMT -5 leonski said:

Oct 1, 2017 15:56:52 GMT -5 Cogito said:

It resulted in MUCH cheaper shipping costs for Emotiva, period.

Balanced against higher cost of switcher

I dare say that the linear power supply with it's large toroidal transformer and large bank of capacitors is considerably more expensive than the switching type.

[teaman]

I know personally I couldn't give a rat's azz about how much the amp draws at idle. I own four plasma televisions that I am sure cost more than any idling amp in my home. I still use a Technics SU-V90 integrated amp with a huge toroidal transformer that is forty years old and still sounds fantastic. I have never used or owned a SMPS amp and probably never will. I will continue to stick with what is tried and true. Big azz amps that weigh a ton and present a longevity untested as of this time. I have a pair of Class G or H Outlaw Audio mono blocks and I didn't like the sound of them. If it costs manufacturers less to ship the SMPs, then who cares as far as we the consumer.

[leonski]

Cog:
Parts count for a Linear is fairly minimal. Big Transformer? check. 30 amp bridge? commodity cheap. A brace of BIG PS electro caps?
Soft Start circuit might be a Thermister which is ALSO dirt cheap. If you want lightning surge protection tossed in? MOVs are pennies apiece.

A switcher will include several inductors and a way to generate the high frequency. You'll have a bunch of SS devices (Xenar Diodes, transistors and such) and
who know what else.
More complex design is beyond the average 'cookbooker' so you either buy ready-made, experiment a bunch or fall back to a linear.

The SMPS advantage is light weight and somewhat better efficiency.

I'd overall 'Lean' the direction of teaman. Nobody in this 'hobby' (alleged) cares much about efficiency. My plasma TV would make the AC work a LITTLE harder. Who cares?

Sorry about not liking the OutLaw M2200 Monoblocks. They generally get good reviews. At one time, I was tempted.

[AudioHTIT]

It’s amazing how little this discussion has to do with the OP. While I don’t speak for Gary, I believe he was saying that if a manufacturer is going to claim a PS is more efficient, then there should be a way to prove it, or they shouldn’t say it (I added the last part). He was only using the Gen 3 SMPS as an example, and mentioned Class D amps as another possible efficient technology. Whether SMPS or Linear or Toroidal or Laminated or Flux Capacitor is the ‘best’ has been discussed ad nasium in other threads, this one is about a measurement standard (or protocol) that could be used to evaluate efficiency or claims thereof.

No one commented on his thread for several days so I took the opportunity to piggyback a desire to know what a piece of equipment draws at idle before I buy it. I do think this is related (at least when comparing like equipment), and is useful for power budgeting purposes. I made the mistake of using ‘electric bill’, when power consumption was at the crux of my interest, it’s not just about money.

What if say you were in a capsule, and your main engines, power, and computers went down and you had to rely on alternate systems not designed to do what you needed, but if they don’t you’d die. Some really creative ... downright brilliant minds come up with a plan, but hey, you’ve gotta reboot and you can only have 1 Amp or its not going to work and ... you die. Hopefully you recognize the story as an extreme example of power budgeting (I don’t remembered what the actual power was). In the sequel they get the system rebooted but they can’t see the display because teaman speck’d out a plasma.

What if though you were building a home where you couldn’t get power, or your grid wasn’t reliable, or you just wanted to be self sufficient; and you had to use solar, wind, generators, and/or batteries. Just for round numbers we can relate to let’s say we get 20 Amps to run the whole house. Well right away we better have a good climate because we’re not getting much AC, but we could probably run a reasonable HT or music system (we’re not using Gar’s amps in A). My point is that wanting to know how much power you are using, or wanting to use only as much as you need to get what you want, isn’t unreasonable; and if you find multiple devices that meet your needs, then choosing the one that consumes the least power might also be reasonable - for some people, maybe many.

Idle power is the least amount you’ll need to to supply or you can’t even turn the unit on, and it’s the least you will consume if you do. I think it’s a useful and telling piece of information that is easy to supply, if you don’t need it don’t look at it. I might even argue that only giving a voltage under “Power Requirements” is being incomplete. It’s not about cost, other things still cost more — though PG&E (my power provider) is about to force everyone to time based billing, and raise the cost significantly from 4pm until 9pm — it’s about efficiency, waste, and independence. This isn’t Gary’s crusade it’s mine, I just think it isnt off topic to discuss it in this thread.

[pknaz]

AudioHTIT,

I get it, that makes a lot of sense. I'm a big fan of off-grid, and I hope one day I can get there (living on a sail boat). If you're considering these types of questions, look up a few posts at my power tables...I think the more important question is how much output power do you need. If I wanted to maximize power efficiency, I would ask myself these questions, listed in what I think is the order of impact:
1) How efficient are my speakers?
2) What is my output power goals?
3) What amplifier topology is most efficient? (A, A/B, D, etc.)
4) What power supply is most efficient?

[Gary Cook]

From the Emotiva blurb on the Gen 3 Amplifiers;

The new XPA Series utilizes a high-current switch mode power supply. It’s a power supply capable of delivering over 3 kVA of continuous power with absolute reliability and extremely high efficiency.

Not picking on Emotiva specifically, but this is after all the Emotiva Lounge and I don't believe that it's unreasonable to ask that they supply the data to back up the claim of extremely high efficiency. The same applies to any other amplifier manuafacturer making such a claim.  Whether or not I'm going to take advantage of that higher efficiency is irrelevant, if you make the claim then back it up with the data. 

To put it bluntly I just don't accept any inability to ship the XPA line of amplifiers due to weight. Sure that applied to the XPR range which were not available in Australia due to their heavy weight and resulting shipping difficulty. But that was never ever the case with Gen 1 or Gen 2 XPA series amps. I've had an XPA-2, XPA-3, XPA-5, XPA-100 and a pair of XPA-1L's all shipped satisfactorily half way around the world without any damage whatsoever.  It just doesn't stand up to any reasonable scrutiny. 

The claim of reduced freight is a similar distraction, the fact is the XPA Gen 3's are physically the same dimensions as XPA Gen 1's and Gen 2's, which at the very least is half of what determines the shipping costs. Using my reshipper freight rates there isn't a $30 difference between the shipping cost of a Gen 1 XPA-5 and a Gen 3 XPA-5, again that's half way around the world. With the current distributor supply here there wouldn't be any difference in freight costs as it should be coming in via shipping containers, that would be well under the weight breaks due to the physical space taken up.

There are 3 reasons given for moving to SMPS, #1 is high efficiency but there is no data supplied to support that.  #2 is shipping difficulty, which plainly didn't exist for the previous generations.  #3 is reduced freight which simply doesn't stand up to even the most basic of checks. 

Cheers
Gary

[DYohn]

Reason #4: it's cheaper to manufacture.

[leonski]

Gary,
Wasn't that it was 'too difficult' to ship but rather that weight / density of what is generally considered a fragile box directly effects handling and later damage claims.

A lighter box will get easier handling from people who handle hundreds of packages / containers per day and who grow inured to non-stop 'fragile' and 'handle with care' signs.

Damage claims cost money as would simply shipping a Brick vs a block of styrofoam. The lighter weight of an amp with SMPS is reflected in price and warranty claims. The trick is to make
such a PS as reliable as the linear it replaces.

Above, DYohn claims 'it's cheaper to manufacture'. which surprises me. Unless those pesky transformers are more $$ than I thought AND the mounting / mechanical end of things is also more robust / $$$.
Maybe a PS can be manufactured En Masse since many amps will share such and THAT saves a bundle right there. Instead of 4 different amps needing 4 different PS, you have ONE SMPS which covers 'em all
and can even be carried forward to new amp designs. Economies of Scale are not to be disregarded.

Manufacturing in this space (Home Entertainment, broadly) is quite competitive and in a sense, a 'game of inches'.

[DYohn]

Transformers are expensive. It is 1/3 the parts cost to manufacture SMPS versus a linear supply.

[KeithL]

OK..... everybody is sort of right.

A simple unregulated power supply, of the type that provides the main power to run the output stage of most Class A/B amps, is actually pretty efficient.
(I'm not going to quote specific numbers - because they vary, and because it really isn't that important.... but we're probably talking near 80%-90% for most of them.)

The big deal is that the SMPS we use in the XPA Gen3 amps is regulated, and it's still very efficient.
A regulated power supply made the old fashioned way, with a big transformer and big capacitors, would be even bigger, heavier, and more expensive... and MUCH less efficient.
The big deal about our SMPS is not that it's dramatically more efficient...
The big deal is that it performs much better, and it's lighter, AND it's also more efficient (we got significantly better performance without sacrificing efficiency).

The efficiency of a Class A/B amplifier at any given output level is limited by the Class A/B circuit topology itself.
The limit is that, at its most efficient operating point, a Class A/B amp is about 70% efficient.
A Class A/B amp CANNOT be more efficient than this.
For most "simple" Class A/B amps, this point occurs around maximum output.

At very low power levels, the idle power starts to dominate.
If you have an amplifier that idles at 50 watts, and it's putting out 1 watt, at that point it's efficiency is 2%.
(Since it's not putting out much power, it's not wasting much, but it's efficiency is still very low.)

The main catch here is that very few amps spend much of their time running at full output (it never works out that way with music).

The BIG advantage of the Class H topology is this.....
That most efficient operating point for most Class A/B amps occurs where the peaks of the signal waveforms are just below the rail voltage (that's maximum output).
With a Class H topology, because you have multiple rails, your amp can be operating at this best possible 70% efficiency at multiple power levels.
Our 300 watt amp can be operating at top efficiency at 300 watts (running on the top rail)...
But it can ALSO be operating at top efficiency at 30 watts (running on the lower rail)...
(A non-Class H amp capable of delivering 300 watts would be running at much lower efficiency at 30 watts.)
This means that the Class H amp spends more of its time operating more efficiently.
So, averaged over time, and taking into account the way music demands widely varying power levels, the Class H topology delivers much better efficiency.
(However, the exact improvement depends on lots of factors, including, literally, what music you play, and how loudly you play it.)