Class AB / Class D ratings?

[Element Zero]

LOL, the switching hash from an early Class-D amplifier smoked the audio input on my laptop computer. I thought I was being careful by turning everything down, and was just going to sneak up the volume a tiny bit to measure the response curve. I powered up the amp, and actually saw smoke coming out one of the speaker grills of the computer. Amazingly the whole computer except for its audio hardware still worked fine, and I always use an outboard audio adapter today when I'm doing testing. I also don't test power amps any more.

Think maybe you had too much hash that day

 

[Wasnex]

If you measure voltage at your 120V outlet, a meter may say 170V - that's because the meter is reading Peak Voltage. If you use a "True RMS" meter, then the voltage reads 120V
Using Ohms law
RMS: 120VW @ 8 Ohms = 1800W
Peak: 170VW @ 8 Ohms = 3612W

See correction above

Also 1800W RMS = 3612.5W peak--same exact waveform.

The red line is 120V RMS and the blue line is 169.7V peak.

 

[themickster]

Let me start by saying that I'm not an expert, anything and everything in this post should be considered incorrect and I'm asking a question about something that I recently heard that I've been thinking about and have been trying to research but I don't have the background to say there's something there or shoot it down.

That said, what I heard was that the peak power rating on a class D amp is more applicable as a rating than RMS than on a class AB amp because of the faster power ramping on a class D amp.

I don't have the engineering background to understand fully and intimately what the ratings mean. Without that, at face value it maybe makes sense that if my (well designed) class D amp can reliably hit peak power, then maybe that becomes a better rating for the amp than RMS?

I hit this looking at the rating for Trickfish Bullhead 1K which isn't listed as RMS and I'm trying to figure out what the real rating for the amp is. And then what do I really mean by "real rating". Nobody from Trickfish made the claim of peak vs RMS, don't want to start that rumor.

Talking simply, RMS is the average power output over a long time.l and peak is the highest output, even if only for a split second.

 

[MultiScaleMale]

Talking simply, RMS is the average power output over a long time.l and peak is the highest output, even if only for a split second.

I think the argument being made was that since Class D ramps up and down quickly, then peak watts is closer to how the amp feels in the real world than RMS since we are mostly playing stuff with articulation rather than sine waves.

I usually go with ~ peak watts on the amp for selecting cabs using the cab’s normal rating. I’m not under powering by don’t have to worry about blowing speakers on “peaks”.

 

[Wasnex]

the user manual states it's peak.

we can only guess, but it's generally possible that rms rating is 1/4 of the peak power.

With speaker ratings, (generally) Program is 2X the RMS rating, and Peak is 2X the Program rating. So you might see a speaker rated 200W RMS, 400W Program, 800W Peak.

RMS is sort of a long term average and Program is the max short term power the driver can handle. Both are measured using RMS volts. The idea is the driver can handle short bursts up to double the RMS power rating as long as the average power does not exceed the RMS rating.

I suggest you keep in mind that many speaker tests are only 2 hours long and the speaker is not necessarily required to meet spec after the test. So, if you want your speakers to last, don't expose them to full Program power.

Peak is stated as 2X the Program rating. But they are the same exact power level. The only difference is Program power is measured in RMS volts and Peak power is measured in peak volts. I.E. 400WRMS "Program Power" = 800Wpeak "Peak Power".

To convert RMS volts to Peak volts, multiply by sq rt of 2 (approximately 1.414).

120VRMS x (1.414) = 169.7Vpeak.

120VRMS = 169.7Vpeak

The reason we have the 4 to 1 power relationship between a speaker's RMS rating and the Peak rating is because a 6dB crest factor is used to test drivers.

Crest Factor is the ratio of the peak value to the RMS value of a waveform.

A constant sine wave has a 3dB crest factor, because the peak power is 2x the RMS power.

Normally we assume doubling the power results in +3dB. But in this case, the power level is not changing...only the way the power is measured is changing.

 

[micguy]

LOL, the switching hash from an early Class-D amplifier smoked the audio input on my laptop computer. I thought I was being careful by turning everything down, and was just going to sneak up the volume a tiny bit to measure the response curve. I powered up the amp, and actually saw smoke coming out one of the speaker grills of the computer. Amazingly the whole computer except for its audio hardware still worked fine, and I always use an outboard audio adapter today when I'm doing testing. I also don't test power amps any more.

I don't know if that was "switching hash". Most power amps can put out a lot more voltage than a computer's audio input can handle. If "switching hash" is prevalent in large amounts on the output of an amplifier, it's gonna fail for radiated emissions, and probably blow certain tweeters pretty quickly.

 

[FullumMusic]

It’s been a long time since electronics classes many many years ago, but RMS is just a function of peak:

Peak value / sqrt(2)

which roughly works out to
Peak value * .707

Doesn’t matter the class of amplifier. It’s just a circuit calculation for AC current that reflects a more realistic/real world achievable current/power similar to what would be achieved with constant direct current. (This is my pared down explanation from my fuzzy memory that could probably be explain more accurately or technically correct by someone smarter than me. There’s a lot of those people around.)

 

[seamonkey]

I think the argument being made was that since Class D ramps up and down quickly, then peak watts is closer to how the amp feels in the real world than RMS since we are mostly playing stuff with articulation rather than sine waves.

I usually go with ~ peak watts on the amp for selecting cabs using the cab’s normal rating. I’m not under powering by don’t have to worry about blowing speakers on “peaks”.

Maybe they're talking about the power supply? SMPS switches at a high frequency and the caps can recover quicker. Any class of amp can use an SMPS.

 

[Killing Floor]

A/B = Analog Bass

D = Digital

I learned this on TB

 

[agedhorse]

Talking simply, RMS is the average power output over a long time.l and peak is the highest output, even if only for a split second.

RMS is the voltage that has the equivalent DC heating effect for any waveform. For a sine wave, this is 0.707x the peak voltage of the waveform.

Peak is the voltage at the highest point of the waveform.

I think the argument being made was that since Class D ramps up and down quickly, then peak watts is closer to how the amp feels in the real world than RMS since we are mostly playing stuff with articulation rather than sine waves.

I usually go with ~ peak watts on the amp for selecting cabs using the cab’s normal rating. I’m not under powering by don’t have to worry about blowing speakers on “peaks”.

Peak power has nothing to do with the “speed” of an amp or power supply. It has to do with where the voltage is measured on the waveform.

 

[FullumMusic]

Another way to think about it is this:

AC current starts at 0, increases energy in one direction, reaches peak (+), decreases energy, hits 0, and then increases energy in the other direction, hits peak (-), decreases energy, and hits 0 again.

When you measure peak value, you’re getting the power at its highest moments in the cycle, the positive or negative peak.

But most of the time that wave isn’t at peak. It’s approaching peak or approaching zero most of the time. It’s not constantly putting out that full peak power.

RMS (Root Mean Square) is a sort of “averaging” calculation to get a number that better represents the useable power of the current taking into account that its not actually at peak most of the time. It’s analogous to the power you’d get from DC, which isn’t oscillating, which is a constant flow of current at a steady voltage.

To go further at the risk of complicating things:
- I put “averaging” in quotes because it’s not actually the average. The numeric average is Peak * 0.637 and isn’t really representative of any useful real world value.

- If you look at the shape of a sine wave (AC current) on an accurate graph. (peak value is at 1 and -1, and then mark where RMS is approx 71% up the Y axis,) you’ll notice the angle of the wave is much steeper up to the RMS line, and much shallower above RMS to the peak. If you think about a speaker reproducing this sine wave, from 0 to the RMS point the cone is moving very fast. From RMS to peak it’s slowing down. At peak it stops. Then it goes the other way, moving slowly to the RMS point then faster from RMS, down to zero, and repeat in the other direction. So if you think about the speaker moving slower when it moves between RMS and Peak, you realize it’s doing less work pushing air during that time vs between 0 and RMS. The cone is moving the most air/doing the most work between -RMS and +RMS. Ergo the RMS value is kinda the “real life” value of what the amp is doing.

-Another visual cue is to look at that same Sine Wave with the RMS line. Look at the “area under the curve.” The area between the waveshape and the 0 line. That area represents the energy the wave has. Now look at the area below the RMS line vs above the RMS line. The area below RMS is much larger than above. Another way to look at that is the waveshape between RMS and Peak is contributing much less energy to the current than the area below RMS.

All this is just a long way to say what’s been mentioned before: Peak and RMS are just measurements, inherently linked to one another. Doesn’t matter what class the amp is or how much time the circuit runs. They’re just calculations on an AC current sine wave measurements. It doesn’t matter what’s creating the electricity. It’s just what the multimeter or oscilloscope reads when measuring the current.

Marketing departments like to use Peak values because mathematically they’re bigger values. There’s also a lot of clever calculus that can be done and measurements that can be taken at different points in the circuitry for the marketing departments to get even larger numbers for the Peak Value.

(Think of it like an internal combustion engine. The horsepower it puts at the shaft is different from the HP that can be measured on the other side of the transmission, which is different than how much HP you can measure at the rear wheels on a dyno. There’s mechanical loss along the drivetrain and a HP rating at the shaft will be higher and not a lie, but you’ll never be able to put that down to the wheels. Clever peak ratings aren’t necessarily a lie but you’ll never be able to put that out at the speaker cone.)

 

[JimiLL]

I believe there has definitely been some intentional numbers games going on regarding class d amps, from a number of notable manufacturers which won’t be listed, at least not by me

 

[taketwo]

This sounds like a question about slew rate, which does impact frequency response for a given output level. Would that be the quality we're trying to distinguish between class A/B and D?

 

[agedhorse]

Another way to think about it is this:

AC current starts at 0, increases energy in one direction, reaches peak (+), decreases energy, hits 0, and then increases energy in the other direction, hits peak (-), decreases energy, and hits 0 again.

When you measure peak value, you’re getting the power at its highest moments in the cycle, the positive or negative peak.

But most of the time that wave isn’t at peak. It’s approaching peak or approaching zero most of the time. It’s not constantly putting out that full peak power.

RMS (Root Mean Square) is a sort of “averaging” calculation to get a number that better represents the useable power of the current taking into account that its not actually at peak most of the time. It’s analogous to the power you’d get from DC, which isn’t oscillating, which is a constant flow of current at a steady voltage.

To go further at the risk of complicating things:
- I put “averaging” in quotes because it’s not actually the average. The numeric average is Peak * 0.637 and isn’t really representative of any useful real world value.

- If you look at the shape of a sine wave (AC current) on an accurate graph. (peak value is at 1 and -1, and then mark where RMS is approx 71% up the Y axis,) you’ll notice the angle of the wave is much steeper up to the RMS line, and much shallower above RMS to the peak. If you think about a speaker reproducing this sine wave, from 0 to the RMS point the cone is moving very fast. From RMS to peak it’s slowing down. At peak it stops. Then it goes the other way, moving slowly to the RMS point then faster from RMS, down to zero, and repeat in the other direction. So if you think about the speaker moving slower when it moves between RMS and Peak, you realize it’s doing less work pushing air during that time vs between 0 and RMS. The cone is moving the most air/doing the most work between -RMS and +RMS. Ergo the RMS value is kinda the “real life” value of what the amp is doing.

-Another visual cue is to look at that same Sine Wave with the RMS line. Look at the “area under the curve.” The area between the waveshape and the 0 line. That area represents the energy the wave has. Now look at the area below the RMS line vs above the RMS line. The area below RMS is much larger than above. Another way to look at that is the waveshape between RMS and Peak is contributing much less energy to the current than the area below RMS.

All this is just a long way to say what’s been mentioned before: Peak and RMS are just measurements, inherently linked to one another. Doesn’t matter that class the amp is or how much time the circuit runs. They’re just calculations on an AC current sine wave measurements. It doesn’t matter what’s creating the electricity. It’s just what the multimeter or oscilloscope reads when measuring the current.

Marketing departments like to use Peak values because mathematically they’re bigger values. There’s also a lot of clever calculus that can be done and measurements that can be taken at different points in the circuitry for the marketing departments to get even larger numbers for the Peak Value.

(Think of it like an internal combustion engine. The horsepower it puts at the shaft is different from the HP that can be measured on the other side of the transmission, which is different than how much HP you can measure at the rear wheels on a dyno. There’s mechanical loss along the drivetrain and a HP rating at the shaft will be higher and not a lie, but you’ll never be able to put that down to the wheels. Clever peak ratings aren’t necessarily a lie but you’ll never be able to put that out at the speaker cone.)

Power is the area under the voltage X current curve (for a resistive load) the energy delivered to the speaker is integrated over time. Even with almost no increase in voltage or current, power is still delivered and energy is still increasing or accumulating over time.

The lesser voltage change at higher voltages in a sine wave is still increasing more quickly that you hypothesis because the power is proportional to the square of the voltage.

This sounds like a question about slew rate, which does impact frequency response for a given output level. Would that be the quality we're trying to distinguish between class A/B and D?

No, the slew rate impacts high frequency not low frequency. The slew rates of all modern amps are at least 10x higher than what’s necessary for bass guitar amplification.

 

[FullumMusic]

I hit this looking at the rating for Trickfish Bullhead 1K which isn't listed as RMS and I'm trying to figure out what the real rating for the amp is.

To give a direct answer to this…
The Trickfish Bullhead .5K specs page gives the Continuous Power specs (which is usually another term for RMS). You could probably shoot Trickfish a message and ask for the RMS values on the Bullhead 1K, ask for clarification on what the Continuous Power value means for their amp.

You can’t really just calculate RMS wattage from Peak wattage. You need to know the Voltage and Amperes.
Watts (rms) = Amps (rms) * Volts (rms)

Also, you can see on the .5K specs they have different RMS values for the same Peak value. They’re doing different THD calculations. (Total Harmonic Distortion). I think they’re getting a higher RMS wattage with the higher THD value because they’re including more of the harmonic partials in their power calculations. But don’t quote me on that, because my memory of that whole relationship is fuzzy.

I have the Bullhead .5K and love it, if that means anything to you. I have it paired with a Trickfish 1x12 w/tweeter. Haven’t had a chance to turn it up because NYC apartment life. But it sounds great even at low volumes. With my 4 string with EMG PJs direct to output (no pots, and a switch for tone open/full-rolloff), gain hits the LED without ever showing clipping, I have to keep the master volume on the 2nd or 3rd tick. My Carvin is much lower output and I need to use the +9dB input and turn up the master a bit more, but I still have to keep it below 1/4 volume otherwise it starts getting a bit too loud for neighbors.

 

[FullumMusic]

Power is the area under the voltage X current curve (for a resistive load) the energy delivered to the speaker is integrated over time. Even with almost no increase in voltage or current, power is still delivered and energy is still increasing or accumulating over time.

The lesser voltage change at higher voltages in a sine wave is still increasing more quickly that you hypothesis because the power is proportional to the square of the voltage.

Aha. Good to know.
This is why I stick to digital. I tell my computer to stay between 1 and -1 and the the converters deal with the math I’ve forgotten how to do.

 

[beans-on-toast]

When I look at an amp, I tend to classify is as loud, louder, and loudest. Modern Class D amps are often light but some Class D amps are heavy. Similarly, not all Class AB amps are heavy. Weight is important to many.

Ratings are not standardized so that can be confusing. An RMS output over a specified bandwidth and total harmonic distortion over that bandwidth is good for comparing amps but that’s not always available. Ratings are not always truthful. Output is often specified but without distortion numbers, it doesn’t allow you to compare. Acceptable distortion is subjective.

Manufacturers often obfuscate these details. I don’t blame them, they want to sell their products. Not all specs are accurate, some are based on a manufacturer’s version of the truth. Specs may not be measured at the tone settings that players use. An important point. Specs are important but are not necessarily what makes a good sounding amp. So don’t just go by the specs.

 

[Alvin2022]

I wonder why they don't use FTC ratings for watts?
Maybe something like that could drive manufactures to a more Honest and equal set of numbers and make it easier to compare different products .

Let me start by saying that I'm not an expert, anything and everything in this post should be considered incorrect and I'm asking a question about something that I recently heard that I've been thinking about and have been trying to research but I don't have the background to say there's something there or shoot it down.

That said, what I heard was that the peak power rating on a class D amp is more applicable as a rating than RMS than on a class AB amp because of the faster power ramping on a class D amp.

I don't have the engineering background to understand fully and intimately what the ratings mean. Without that, at face value it maybe makes sense that if my (well designed) class D amp can reliably hit peak power, then maybe that becomes a better rating for the amp than RMS?

I hit this looking at the rating for Trickfish Bullhead 1K which isn't listed as RMS and I'm trying to figure out what the real rating for the amp is. And then what do I really mean by "real rating". Nobody from Trickfish made the claim of peak vs RMS, don't want to start that rumor.

 

[silky smoove]

I hit this looking at the rating for Trickfish Bullhead 1K which isn't listed as RMS and I'm trying to figure out what the real rating for the amp is.

I own this amp and have spoken about its output on a number of occasions. In the absence of actual bench measurements, I summarize it like this: It's a very powerful amp with pleasant voicing that has a slight roll-off on the very top end. It's fairly "polite" sounding for lack of a better term. Regardless of actual output measurements, this voicing has brought me to the conclusion that it's comparably loud to other 800-1,000 watt heads in terms of total output, but it appears a bit quieter than comparable amps with a more aggressive upper midrange voicing such as a lot of Gallien-Krueger's offerings. If you compare it to a Markbass LM800 or Genz Benz Streamliner 900 (which I have done in both cases) you'll find that it's pretty comparable. The Streamliner overdrives more pleasantly and as such can play a bit louder, albeit differently. Stacking it up against a GK MB800 (which I also have done), the GK felt louder throughout, but that was likely due to the upper midrange voicing. It's a powerful enough amp to blow out many commercially available cabs that you would pair it with if you dimed everything. I pair it with a Trickfish TF212V and never want for volume on gigs when I need to bring an amp.

 

[agedhorse]

I wonder why they don't use FTC ratings for watts?
Maybe something like that could drive manufactures to a more Honest and equal set of numbers and make it easier to compare different products .

The FTC rating method has had to evolve over the years as amps grew in power and number of channels.

It’s more suited to amps used for reproduction where the peak to average ratio is more well defined, and the amp’s normal operation does not include overdrive or clipping.