An XXcontroversial Way to Compare the Output of Class D Amps.

[Stub Mandrel]

Elsewhere, I have seen regular discussion of the relative power outputs of different amps.

 

The issue is that some makers use RMS power, some use intermittent RMS power, some use RMS before the onset of distortion, some use RMS with various degrees of distortion, some use the theoretical power if outputting a square wave and some use theoretical transient power measured in various ways...

 

This is particularly true of class D amps where some astonishing levels of output power get claimed.

 

However, what we are really interested in is how much power can an amplifier consistently put out over an extended period (e.g. a whole song).

 

One way is to consider the amps power consumption, as this should be based on real world figures, not hand waving. Power out cannot exceed power in for any extended period. Input power may be given in Volt Amps (VA) in which case a power correction factor of 0.65 to around 0.8 depending on the quality of the power supply will need to be applied to  derive watts.

 

I will compare two amplifiers:

 

Bugera Veyron. Claimed output power: 2000 Watts @ 4 Ohms. Rated input power: 110 Watts.

 

Orange Terror. Claimed output power: 500 Watts @ 4 ohms. Rated input power: 690 VA. (Volt-amps, multiply by a power factor of between 0.65 and 0.8 to convert to watts, so 480 to 550 watts).

 

Assuming both class D amps have similar efficiencies, the Orange is consuming 4-5 times the power of the Bugera, which is rated as being four times as powerful as the Orange.

 

Even allowing a dozen watts for the Orange's valve preamp, I smell a rat...

Edited July 22, 2022 by Stub Mandrel

[wateroftyne]

I'm assuming this is along the lines of TC's nonsense power ratings?

[agedhorse]

It's not quite so easy, though there are rules of thumb that engineers can use to determine the validity of specifications (called cross-checking or cross-verification).

 

The AC input power is required for any amp to be the current draw at 1/8-rated RMS audio power at the lowest rated nominal impedance for any amplifiers compliant with IEC 60065 or 62368 safety standards. The standards also include a provision that if an amplifier is intended to and marketed to drive more than 1/8-duty cycle, the stated AC input draw MAY be increased to account for the increased (more severe) duty cycle.

 

That said, the rated AC input power is verified by the test lab and really can't be fudged. RMS rated power can be calculated from the input power using the following method. 

 

Using an assumed efficiency of 80% for a class D/SMPS, the 110W is multiplied by the efficiency then the result is divided by the duty cycle for rated RMS audio power.

Using the 110W input example:  (110W x .80eff)/.125dutycycle = 704 watts RMS (presumably 4 ohms)

With the Orange Terror, to me it looks like they are using a more severe duty cycle because roughly (500 watts input x .8 eff)/.125 = 3200 watts. In fact, what it looks like is that the rated input is actually the maximum rated input rather than the average rated input. This would conform to a duty cycle  of 1

 

On the Subway amps, I typically use a duty cycle of between 1/4 and 1/3, therefore the rated AC input is higher than what it would be if we used 1/8-duty cycle (IIRC, I use ~400 watts rated AC input for an 800 watt RMS rated amp). This is because the amps are advertised and marketed to be able to be overdriven, and these numbers would then be accurate. 

 

 

 

[Stub Mandrel]

I live and learn. Thank you @agedhorse

[Phil Starr]

13 hours ago, agedhorse said:

It's not quite so easy, though there are rules of thumb that engineers can use to determine the validity of specifications (called cross-checking or cross-verification).

 

The AC input power is required for any amp to be the current draw at 1/8-rated RMS audio power at the lowest rated nominal impedance for any amplifiers compliant with IEC 60065 or 62368 safety standards. The standards also include a provision that if an amplifier is intended to and marketed to drive more than 1/8-duty cycle, the stated AC input draw MAY be increased to account for the increased (more severe) duty cycle.

 

That said, the rated AC input power is verified by the test lab and really can't be fudged. RMS rated power can be calculated from the input power using the following method. 

 

Using an assumed efficiency of 80% for a class D/SMPS, the 110W is multiplied by the efficiency then the result is divided by the duty cycle for rated RMS audio power.

Using the 110W input example:  (110W x .80eff)/.125dutycycle = 704 watts RMS (presumably 4 ohms)

With the Orange Terror, to me it looks like they are using a more severe duty cycle because roughly (500 watts input x .8 eff)/.125 = 3200 watts. In fact, what it looks like is that the rated input is actually the maximum rated input rather than the average rated input. This would conform to a duty cycle  of 1

 

On the Subway amps, I typically use a duty cycle of between 1/4 and 1/3, therefore the rated AC input is higher than what it would be if we used 1/8-duty cycle (IIRC, I use ~400 watts rated AC input for an 800 watt RMS rated amp). This is because the amps are advertised and marketed to be able to be overdriven, and these numbers would then be accurate. 

 

 

 

This is one of the clearest and most concise summaries I've seen. I suspect I might be quoting this for a while.

 

Interestingly the Bugera has been a subject of much debate and as a result has been repeatedly measured at around 720W depending upon the distortion levels you want to measure it at. That's entirely consistent with the figure of 705W RMS calculated above. 

 

There's a genuine dilemma for amp designers and manufacturers. You can design an amp to score well in the standard tests at 1/8 ratings or beef up the power supply, heat sinking and a few components and deliver a more rugged amp that will deliver in a wider range of conditions. In my book that makes it a better amp and likely to prove more reliable to boot but it is then marketed against amps that look better on paper. On the upside such manufacturers should build a better reputation long term.

 

For me I'd prefer a regulated market where only standardised figures are allowed to be quoted in advertisements and penalties are imposed upon companies who deliberately mislead. Even this 'simple' explanation @agedhorsehas offered is not easy for a non technical musician to follow. I'd like us as a society to encourage the good guys and discourage sharp practice. Think VW and emissions. 

[chyc]

For me one obvious tell, at least in the PA market, is where I can see active PAs with some colossal claim as to what its power rating is, and yet when the same company sells passive units the number is a tiny fraction of it (because they're worried you might *actually* attach it to a powerful amp). While it's possible, it's seriously stretching my disbelief that there isn't some marketing BS going on.

 

By no means are they the only ones, but QSC sell the 2000W K12.2 active top, but their E112 is rated for 400W continuous power. Yes yes they're not the same thing, but to me to be different by a factor of 5 is massaging measurements to the point of dishonesty

[agedhorse]

1 hour ago, chyc said:

For me one obvious tell, at least in the PA market, is where I can see active PAs with some colossal claim as to what its power rating is, and yet when the same company sells passive units the number is a tiny fraction of it (because they're worried you might *actually* attach it to a powerful amp). While it's possible, it's seriously stretching my disbelief that there isn't some marketing BS going on.

 

By no means are they the only ones, but QSC sell the 2000W K12.2 active top, but their E112 is rated for 400W continuous power. Yes yes they're not the same thing, but to me to be different by a factor of 5 is massaging measurements to the point of dishonesty

There are a couple of things going on here that the marketing departments jumped on like rabid dogs...

 

First, there is the obvious limitation of the high frequency driver which is no more than about 35-40 watts RMS above 800hz. There is dynamic mechanical power handling above this, but in practice, due to the sensitivity difference between the HF and LF sections, this doesn't even come into play. Typically the HF section is at least 10dB more sensitive than the LF section on a cabinet like this. So, even if the amp is nominally capable of delivering "1000 watts" (or whatever number), the internal processing limits this back to no more than 100 watts for "survive-ability" reasons.

 

Second, looking at the LF section, the driver is not going to be capable of more than 400 watts RMS at low frequencies, therefore in the processing, the limiting and HPF algorithms are often dynamic and increase the amount of limiting as level gets higher and frequency gets lower. This too improves "survive-ability". 

 

Since this is a biamped speaker, the easiest and most common solution is to use the same power amp design for both channels, and use the driver's impedance plus processing to manage the real world power levels delivered to the speaker. For example, the amp might be rated at 1000 watts RMS at 4 ohms, but if the drivers are 8 ohms (or even 16 ohms in the case of  the HF driver), the unprocessed power available would be 500 watts RMS to the LF and 250 - 500 watts to the HF. 

 

There's nothing wrong (and a LOT right) with managing power this way but IMO the real issue is with the way the numbers are manipulated by the marketing folks in a way that appears deceptive even if that's not the intent. Incidentally, this is a major conflict at many companies, between the design/engineering departments and marketing departments.

[Stub Mandrel]

Perhaps the most difficult issue is matching speakers and cabs when it is so hard to understand real world output figures.

 

For example, my main cab is rated '900W RMS (AES Standard)’. Not simple to grasp, and I don't know how it compares to real world figures for bass rather than "mixed programme".

 

Is there an aes standard for amplifiers?

[agedhorse]

4 hours ago, Stub Mandrel said:

Perhaps the most difficult issue is matching speakers and cabs when it is so hard to understand real world output figures.

 

For example, my main cab is rated '900W RMS (AES Standard)’. Not simple to grasp, and I don't know how it compares to real world figures for bass rather than "mixed programme".

 

Is there an aes standard for amplifiers?

Yes, but there are also other just as valid (and maybe more-so) depending on how powerful the amp might be. What's appropriate for a 100 watts amp that's going to be driven into compression and clipping might not be appropriate for a 2500 watt amp that has the capacity to destroy most speakers if driven into clipping or compression. It depends. 

 

A designer has to come up with a method that makes their intended customer happy with the results, doesn't result in premature shut-down or self-destruction and delivers adequate performanece.

[Chienmortbb]

On 19/07/2022 at 19:20, agedhorse said:

There are a couple of things going on here that the marketing departments jumped on like rabid dogs...

 

First, there is the obvious limitation of the high frequency driver which is no more than about 35-40 watts RMS above 800hz. There is dynamic mechanical power handling above this, but in practice, due to the sensitivity difference between the HF and LF sections, this doesn't even come into play. Typically the HF section is at least 10dB more sensitive than the LF section on a cabinet like this. So, even if the amp is nominally capable of delivering "1000 watts" (or whatever number), the internal processing limits this back to no more than 100 watts for "survive-ability" reasons.

 

Second, looking at the LF section, the driver is not going to be capable of more than 400 watts RMS at low frequencies, therefore in the processing, the limiting and HPF algorithms are often dynamic and increase the amount of limiting as level gets higher and frequency gets lower. This too improves "survive-ability". 

 

Since this is a biamped speaker, the easiest and most common solution is to use the same power amp design for both channels, and use the driver's impedance plus processing to manage the real world power levels delivered to the speaker. For example, the amp might be rated at 1000 watts RMS at 4 ohms, but if the drivers are 8 ohms (or even 16 ohms in the case of  the HF driver), the unprocessed power available would be 500 watts RMS to the LF and 250 - 500 watts to the HF. 

 

There's nothing wrong (and a LOT right) with managing power this way but IMO the real issue is with the way the numbers are manipulated by the marketing folks in a way that appears deceptive even if that's not the intent. Incidentally, this is a major conflict at many companies, between the design/engineering departments and marketing departments.

I well remember when at Panasonic that a certain company (Korean company beginning with S) claimed a huge brightness figure for their Plasma screens. How did they do it? By reducing the size of the white area measured. The power supply of plasma screens could not maintain full brightness over the whole screen so the smaller you could make the measured square, the brighter the picture. They reduced the square from 2.54 x 2.54cm  to 1 x 1 cm and hence the power supply could maintain a higher brightness.

 

Your calcs on the Bugera are about right.

[chyc]

6 minutes ago, Chienmortbb said:

I well remember when at Panasonic that a certain company (Korean company beginning with S) claimed a huge brightness figure for their Plasma screens. How did they do it? By reducing the size of the white area measured. The power supply of plasma screens could not maintain full brightness over the whole screen so the smaller you could make the measured square, the brighter the picture. They reduced the square from 2.54 x 2.54cm  to 1 x 1 cm and hence the power supply could maintain a higher brightness.

They're still doing it, with pretty much the same technique. This is from 2022: https://entertainment.slashdot.org/story/22/06/15/0615211/samsung-caught-cheating-in-tv-benchmarks

[dmccombe7]

This is clarifying something that's troubled me over many years.

Once again i find @agedhorse coming forward with valuable insights into this complex issue and simplifying as much as possible for non-techs.

Its nice to know how designers think about this subject.

This might explain why Mesa amps are solid bits of kit 👍

Cheers @agedhorse

 

Dave

[Chienmortbb]

On 19/07/2022 at 19:20, agedhorse said:

Second, looking at the LF section, the driver is not going to be capable of more than 400 watts RMS at low frequencies, therefore in the processing, the limiting and HPF algorithms are often dynamic and increase the amount of limiting as level gets higher and frequency gets lower. This too improves "survive-ability". 

 

Despite manufacturers claims to the contrary, there is not

 

7 hours ago, chyc said:

They're still doing it, with pretty much the same technique. This is from 2022: https://entertainment.slashdot.org/story/22/06/15/0615211/samsung-caught-cheating-in-tv-benchmarks

I never said it was Samsung ..............................but it w*s

[Chienmortbb]

On 19/07/2022 at 19:20, agedhorse said:

Second, looking at the LF section, the driver is not going to be capable of more than 400 watts RMS at low frequencies

Do you know of a 12" that can really exceed 400W RMS at low frequencies?

[agedhorse]

4 hours ago, Chienmortbb said:

Do you know of a 12" that can really exceed 400W RMS at low frequencies?

Yes, but for bass guitar especially, the mid characteristics and sensitivity are very important also. There is a tradeoff between raw power handling (including both thermal and mechanical) and sensitivity, mid extension, cost, and often there is a need for a larger enclosure size. With the number of players that would be ok with these tradeoffs being quite low, it's not really all that viable (IME anyway) for a commercial product. 

 

At this point, 400 watts RMS at low frequencies (say down to 40Hz) appears to be about the sweet spot. If you were to take such a driver and high pass the signal to 50Hz, than the same driver might be ok to 500W RMS, but that's a mighty bold assumption for a manufacturer responsible for warrantying their speaker to make with confidence that it won't come back to bite them in the butt (bum). Being wrong on such an assumption could result in extraordinarily high warranty costs.

[stewblack]

As a non technically inclined end user I have never really cared too much about the numbers. 

I understand their value to the marketing team but very early on in my career I discovered a Trace Elliot 130 watt combo was loud. Really loud and clear. Other amps with bigger numbers were not as loud. So I was lucky not to be seduced. 

I now gig with numbers varying from 50, 100, 200 and many in between up to 2000. 

The explanations above serve merely to muddy the waters even further for those like myself. That is not intended to deride the knowledge behind them, I simply can't grasp it.

[dmccombe7]

59 minutes ago, stewblack said:

As a non technically inclined end user I have never really cared too much about the numbers. 

I understand their value to the marketing team but very early on in my career I discovered a Trace Elliot 130 watt combo was loud. Really loud and clear. Other amps with bigger numbers were not as loud. So I was lucky not to be seduced. 

I now gig with numbers varying from 50, 100, 200 and many in between up to 2000. 

The explanations above serve merely to muddy the waters even further for those like myself. That is not intended to deride the knowledge behind them, I simply can't grasp it.

Understand what you mean. I was hoping for a simple jack % Jill explanation but its far more complex than i ever imagined.

I reckon designers and manufacturers are waging a "Dark Strategy" to confuse the masses................its working 

Seriously there's far more to it than i ever imagined.

All credit to the companies that are doing it right...............i just dont know who they are ? 

Dave

[Chienmortbb]

On 18/07/2022 at 23:12, wateroftyne said:

I'm assuming this is along the lines of TC's nonsense power ratings?

 

 

 TC were using a form of compression to "boost" the volume. The model numbers were designed to imply a power output and digging into the blurb, it was apparent that the model number gave TCs assessment of the perceived loudness compared to the loudness of an uncompressed amp. Of course lot of manufactures imply the power output in the model number but TC were using an extra degree of sleight of hand. It may well have worked better if they had been honest with the power rating and use the "clever trickery" as a benefit. There is nothing wrong per se with TC amps but they were caught doing a VW.

[agedhorse]

5 hours ago, dmccombe7 said:

Understand what you mean. I was hoping for a simple jack % Jill explanation but its far more complex than i ever imagined.

I reckon designers and manufacturers are waging a "Dark Strategy" to confuse the masses................its working 

Seriously there's far more to it than i ever imagined.

All credit to the companies that are doing it right...............i just dont know who they are ? 

Dave

A simplistic explanation wouldn’t be accurate. It IS complicated.

[stewblack]

7 hours ago, dmccombe7 said:

its far more complex than i ever imagined.

Yep.

[uk_lefty]

11 hours ago, stewblack said:

I simply can't grasp it.

You and me both!

 

I'd have thought the best way would be to stick them side by side in the same room and play them (with the same bass) and if one has more output than the other that's your answer. 

[chyc]

1 hour ago, uk_lefty said:

I'd have thought the best way would be to stick them side by side in the same room and play them (with the same bass) and if one has more output than the other that's your answer. 

Even then it's not simple. Different amps have different scales on their volume. Some are linear in volume (I think EA is a manufacturer that springs to mind), some are more logarithmic. Why does this matter? Well, if you have otherwise identical amps at 50% on their volume dial, then the logarithmic one will sound louder because it's actually at something like 80% of its maximum power.

 

I'd call that a marketing trick for just such a scenario as you describe, but apparently some people like the non-linearity of their volume dials. To each their own I guess.

[agedhorse]

2 hours ago, chyc said:

Even then it's not simple. Different amps have different scales on their volume. Some are linear in volume (I think EA is a manufacturer that springs to mind), some are more logarithmic. Why does this matter? Well, if you have otherwise identical amps at 50% on their volume dial, then the logarithmic one will sound louder because it's actually at something like 80% of its maximum power.

 

I'd call that a marketing trick for just such a scenario as you describe, but apparently some people like the non-linearity of their volume dials. To each their own I guess.

Marketing departments certainly like this, but apparently many players do feel that an amp's more powerful that way too. As players mature, they tend to grasp the benefit of linearizing the rotation and scaling of the range of the control though.

[dmccombe7]

6 hours ago, agedhorse said:

Marketing departments certainly like this, but apparently many players do feel that an amp's more powerful that way too. As players mature, they tend to grasp the benefit of linearizing the rotation and scaling of the range of the control though.

For me a visual indication of where my amp is running as a percentage can be helpful.

Depending on what band i'm playing with i find the volume can creep up without you truly realising it. A quick visual would let me know if i'm pushing my master volume too much and my amp is close to its max limit.

Dave

[Chienmortbb]

10 hours ago, agedhorse said:

As players mature, they tend to grasp the benefit of linearizing the rotation and scaling of the range of the control though.

 As you know human hearing, like so many things in nature is logarithmic, so we use Log or Audio scaling on pots or in DSP to mimic a linear increase in volume. I am unclear what people mean when they say Linear in relationship to Gain or Volume controls. However I fear we have gone way off topic, my apologies to all.

 

 

Edited July 23, 2022 by Chienmortbb