RMS Watts

[Don Richard]

Last week a friend of mine who has a company that does sound system installations in bars, restaurants and the like asked how to calculate RMS watts. I tried to explain root mean square to him but he wasn't quite understanding the math so I looked for a writeup I could email to him. During my research I discovered that there is no such thing as "watts RMS".

It seems that you can calculate RMS volts and RMS current, but when multiplied together, what you get is average power. Now, if you look in Parts Express or just about anywhere that sells speakers and amps, you'll see the term "watts RMS" widely used. During my research I even found the advise "make sure your amp is rated in watts RMS so you know you are getting all the power you paid for".

Here is a good article written by Bob Lee of QSC on that subject, as well as amplifier rating methods and information on amplifier gain:

http://www.qscaudio.com/support/library/papers/puzzle.pdf

[Rivernuggets]

http://www.qscaudio.com/support/library/papers/puzzle.pdf

[stormin]

That was a good read. Thanks.

[russ69]

Output watts do not really tell the whole picture. The old FTC method gave the amp a good workout but it still wasn't a definitive testing method. I use stated output as more of a general class of amp rather than an absolute value. Loads vary, inputs vary, frequency varies and headroom varies. I can't say I'll pick a 200 watt amp over a 250 watt amp, it's just not enough information to make a factual decision.

If you do your own bench testing you might be able to rate the output under equal test conditions.........maybe.

Thanx, Russ

[garyrc]

Last week a friend of mine who has a company that does sound system installations in bars, restaurants and the like asked how to calculate RMS watts. I tried to explain root mean square to him but he wasn't quite understanding the math so I looked for a writeup I could email to him. During my research I discovered that there is no such thing as "watts RMS".

It seems that you can calculate RMS volts and RMS current, but when multiplied together, what you get is average power. Now, if you look in Parts Express or just about anywhere that sells speakers and amps, you'll see the term "watts RMS" widely used. During my research I even found the advise "make sure your amp is rated in watts RMS so you know you are getting all the power you paid for".

In the '50s the better manufacturers started specifying this "watts RMS" figure (even if it is illusory), and maintained it through the terrible late '60s and '70s, when the bad manufacturers were using other methods (like "Music Power," etc.) which got them much higher figures. So, wouldn't the "watts RMS" figure be a good way to compare amps, even now? I asked McIntosh what their power ratings meant on their 2,000 wt three piece monoblock, because they used a term I wasn't familiar with, and they replied "It's really watts RMS." Don't we (and some manufacturers) just turn up the amp until the waveform is barely clipped on the 'scope, then back it up so it barely isn't, than take the value in watts and multiply by .707? I have amps that are rated at 150 wts/ch. The dealer went through the above procedure, and came up with 171 watts just before clipping, so, multiplying by ,707 I assumed the "true RMS" is about 120 watts/ch

[Don Richard]

Don't we (and some manufacturers) just turn up the amp until the waveform is barely clipped on the 'scope, then back it up so it barely isn't, than take the value in watts and multiply by .707?

The problem with that method is that it results in numbers that are no less arbitrary than:

the bad manufacturers were using other methods (like "Music Power," etc.) which got them much higher figures.

The FTC rating system, which must be used when rating home stereo amplifiers, was created to eliminate differing methods of rating power so that valid comparisons could be made. The FTC procedures result in power ratings that are something like this:

"Output power is 100 watts per channel, both channels driven, from 20-20000 Hz, at no more than 0.1% total harmonic distortion"

This method requires a preconditioning period where both channels are run at 1/3 rated power for one hour prior to testing .This is a tough test that results in accurate numbers that are useable for comparitive purposes.

Home theater and autosound equipment are not required to be rated via the FTC standards, and there is much BS in their power numbers. Let the buyer beware.

[Trey Cannon]

Another word that is used is Continus power. I kind of like this one....stright forward.

Here is how we rate the Aragon amps...well when we owned them.

Specifications

design	5-ch THX Ultra2 Amplifier
power ouput 8	200 watts/ch into 8 ohms at less than 0.03% THD (20Hz-20kHz - 3 channels driven)
power output 4	300 watts/ch into 4 ohms at less than 0.05% THD
frequency response	20Hz-20kHz +0/-0.1dB 10Hz-100kHz +0/-0.5dB
tim/dim distortion	0.003%
s/n ratio	106dB A-weighted
damping factor	500 (8 ohms - 50Hz)
input sensitivity	100mV-1W / 1.42V-200W
input impedance	20k ohms
power consumption	105W @ Idle / 1200W Maximum
built from	2002

[garyrc]

Trey (and others), IYO how does "continuous" relate to RMS? For example, if my amp rated by the manufacturer at 150 watts/ch, both operating, is just below clipping (dealer mearurement) at 171 watts/ch both operating, @ 8 Ohms, and I "dry labed" 120 awtts RMS @ 8 Ohms by multiplying by .707, which figure (if any) would be close to "continuous?"

One thing that irritates me about this otherwise good manufacturer's specs, is that they rate the power into "8 or 4 Ohms" ... why not provide different figures? That's like the power plug people who say their $100 duplex outlet is "15 or 20 amps"... why not just 20, if it is appropriate for a 20 amp circuit without compromise? If the sales depts -- or whoever -- are going to do this to us, I wish they would at least spell out exactly what they mean.

[djk]

"One thing that irritates me about this otherwise good manufacturer's
specs, is that they rate the power into "8 or 4 Ohms" ... why not
provide different figures? That's like the power plug people who say
their $100 duplex outlet is "15 or 20 amps"... why not just 20, if it is
appropriate for a 20 amp circuit without compromise? If the sales
depts -- or whoever -- are going to do this to us, I wish they would at
least spell out exactly what they mean."

A 20A outlet will accept either 15A or 20A plugs, a 15A outlet will only accept 15A plugs. The difference is the 15A is to be wired with 14ga, the 20A is to be wired with 12ga, and the circuit protection is sized accordingly. One should never use a 20A outlet on a 15A circuit.

" Trey (and others), IYO how does "continuous" relate to RMS? For
example, if my amp rated by the manufacturer at 150 watts/ch, both
operating, is just below clipping (dealer mearurement) at 171 watts/ch
both operating, @ 8 Ohms, and I "dry labed" 120 awtts RMS @ 8 Ohms by
multiplying by .707, which figure (if any) would be close to
"continuous?""

As was explained previously, there is no such animal as 'watts RMS'. Average power is the correct term, but was disliked by advertising departments. Having said that, RMS watts is the preferred specification. It's easy to play games in the power ratings game, and in my mind the only numbers that are 'real' are what Trey is calling 'continuous power', all channels driven, into a stated impedance, over a stated bandwidth. As a supplemental measurement, open circuit maximum voltage gives an idea of power supply regulation, or peak power for transients.This last may be a confusing number, as the less regulation a power supply has, the more peak power is available, and some of the best amplifiers with regulated supplies have zero peak headroom.

Even these numbers are not enough to tell us how an amplifier will sound on music driving a real loudspeaker, listening will be required at some point to be sure. A case in point would be to compare the 200W/8R Carver M400 (cube) with a good, small amplifier. A Yamaha CA400 (20W) played almost as loud, the CA600 (30W) seemed to play as loud (with a bit of distortion), the CA800 (50W) seemed comparable, and the CA1000 (70W) seemed to sound much better. Another case in point would be to compare the 75W/8R Threshold S150. The Threshold sounded fantastic on the Klipschorns, very musical. However, it distorted (in a very bad way) at the point of clipping, so it didn't sound as loud as a 20W NAD receiver.

In the above situations, how would having accurate information help choose the best amplifier?

[DrWho]

Although I agree that there is no such thing as "RMS Watts", I don't think it's completely a misnomer because you don't directly measure the power coming out of the amplifier. In most cases, you measure the RMS voltage being delivered into a known load and then calculate the power. So saying "RMS watts" is really a misleading way of saying that an RMS voltage was used for the power calculation.

Also, RMS technically means "root mean square" so you could technically take the root mean square of the power, and you're going to end up with the same number as the "average power"...

[DrWho]

In the above situations, how would having accurate information help choose the best amplifier?

Well I believe in a causal universe, so I've no doubt there is a set of metrics that would describe the audible differences between the amplifiers. The problem is that those metrics would be too complicated for the general public...which at the end of the day puts you back into the realm of listening to the amplifier instead.

[djk]

My point exactly.

----------------------------------------------------------

john curl

"You people worry too much about superficial things."

"Listening is more important than measuring."

"Absolute THD is not very important to me, (because) it is usually below what most references would consider audible."

"I would prefer a transformer to an IC chip. I once designed out (removed from the design) the best IC chip that I could find.... when it failed critical listening tests."

[DrWho]

"I would prefer a transformer to an IC chip. I once designed out (removed from the design) the best IC chip that I could find.... when it failed critical listening tests."

This is off-topic, but how do you get around the 60Hz pickup from lights in the room and early LF roll-off at strong drive levels? I love the elegance of galvonic isolation, but I've never seen it implemented without tradeoffs.

[Don Richard]

In most cases, you measure the RMS voltage being delivered into a known load and then calculate the power. So saying "RMS watts" is really a misleading way of saying that an RMS voltage was used for the power calculation.

You can measure the RMS voltage of, say, a 20 ms burst and then calculate power. The reading derived in that manner will be higher than if you used a continuous voltage reading. Unscrupulous manufacturers can then say, "XXX watts RMS", and be truthful in what they say whilst misleading the consumer into thinking their amp is more powerful than it actually is.

[garyrc]

I don't know too much about electronics, so I have a couple more questions.

So ... if the continuous power of an amp, all channels operating, after the specified 1/3 power preconditioning time, is measured just below clipping, and it also comes in at distortion that is at or below whatever the current FTC standard is, wouldn't the continuous power be a higher figure than a "RMS watts" figure (even though fictional) that is mathematically derived using all of the same conditions (e.g., not fudging by using tone bursts, etc.)?

I guess I'm just trying to see if continuous ratings are less conservative than properly & similarly derived, but imaginary, RMS watts figures, i.e., would an a amplifier rated by a reputable company (like McIntosh or you-name-them) as "120 watts RMS continuous" be rated at about 170 watts continuous? Another way of putting this might be "Would an amp rated by a good manufacturer at 120 watts RMS in 1975, now be rated as about 170 watts continuous? I'm still hung up on the 0.707 thing.

The other question is whether modern, good, solid state, amps behave differently if the continuous signal used to measure power covers the whole 20 -20K bandwidth, or just 1K... I'm asking because I think I recall John Atkinson using 1K to determing the amp power.

[Trey Cannon]

One thing for sure is that if you use a factor of 5 for your headroom, the power number becomes a little less of an issue.

FYI: for speakers what we do is run them with an IEC standard (I think 266A) signal using RMS Voltage to set the test.

RF-62 IEC standard in to amp, out of amp to speaker at 24.5 Vrms with speaker as load. Ran for 8 hr. to confirm. To find the point of failure, we ease up 25 watts every 2 hr. If it dies at 125 W we back it down to 100 W and run the 8hr test. If a few pass, that is where it is ratted.

For me, I wish we would quit using Watts and Ohms in the spec's.

How about using Volts and Min Z. Then with a little math you can find most anything else you want.

Over seas, you have to post Min Z on every speaker and test the power with that number.

[djk]

"This is off-topic, but how do you get around the 60Hz pickup from lights in the room and early LF roll-off at strong drive levels? I love the elegance of galvonic isolation, but I've never seen it implemented without tradeoffs. "

Jensen and Cinemag make transformers with fantastic shielding and response -3dB at 0.25hz, and handle +19dB at 20hz, +23dB at 30hz.

http://www.jensen-transformers.com/datashts/pi2xx.pdf

[DrWho]

And how much does it cost? I've used these guys before and they don't solve the issue of nearfield magnetic disturbances cropping up above the noisefloor....

[Ski Bum]

I don't know too much about electronics, so I have a couple more questions.

So ... if the continuous power of an amp, all channels operating, after the specified 1/3 power preconditioning time, is measured just below clipping, and it also comes in at distortion that is at or below whatever the current FTC standard is, wouldn't the continuous power be a higher figure than a "RMS watts" figure (even though fictional) that is mathematically derived using all of the same conditions (e.g., not fudging by using tone bursts, etc.)?

I guess I'm just trying to see if continuous ratings are less conservative than properly & similarly derived, but imaginary, RMS watts figures, i.e., would an a amplifier rated by a reputable company (like McIntosh or you-name-them) as "120 watts RMS continuous" be rated at about 170 watts continuous? Another way of putting this might be "Would an amp rated by a good manufacturer at 120 watts RMS in 1975, now be rated as about 170 watts continuous? I'm still hung up on the 0.707 thing.

The other question is whether modern, good, solid state, amps behave differently if the continuous signal used to measure power covers the whole 20 -20K bandwidth, or just 1K... I'm asking because I think I recall John Atkinson using 1K to determing the amp power.

Regarding your C272, remember that's some variation on class G/H with dual supply rails, which kind of confuses matters compared to an amp with a conventional power supply. 'Continuous power' reflects the lower power rail's limits, which is 150w, or 171w at slightly more forgiving tolerances. The high power supply rail provides the added headroom, drawn on when the signal demands it, but is only available for fairly brief periods, so a continuous signal will default back to the lower supply as soon as the higher rail is exhausted. But it should be able to roast along at 150w all day long. With a variable signal (music), it's capable of 300+ or however many unclipped watts. It makes understanding NAD specs a little strange, but they do at least spec conservatively. It's also why your NAD rated at 150w 'continuous' sounds as powerful a conventional amp of 250w or more RMS power.

[Arkytype]

Don,

To quote Yogi Berra, It's like deja vu all over again! The RMS vs. continuous power topic came up in this forum in November of 2004.

http://community.klipsch.com/forums/t/47910.aspx?PageIndex=1

Rather than re-write my brilliant contribution back then, here's a replay plus additional info to make your head swim:

"There is no such quantity as "RMS power". You won't find it in any textbook or in the National Electrical Code. You will, however, find it in the sales brochures of most amplifier manufacturers.

McIntosh Laboratories has used the correct power rating terminology of their amplifiers for many years. It is: "Continuous average sine wave power".

While we should measure the voltage across the load with a true RMS voltmeter or the current through the load with a true RMS ammeter, that doesn't mean the resulting power quantity computed using Ohm's law is "RMS watts".

Assuming the source is a sine wave, you are measuring the average power delivered by the amplifier. This is equivalent to the DC heating power delivered to a specified load resistor.

I suspect the RMS power term came into use as a result of the FTC's misguided attempts to standardize power amplifier meaasurements in 1978. Back then, an honest 30 watt amplifier could be rated at some inflated value using the now-infamous Instantaneous Peak Power rating."

Flash forward to today. Since DC circuits preceeded alternating current circuits, it was easy enough to compute the power dissipated in a load using Ohm's Law; Power=DC voltage squared / load resistance in ohms. Now it mattered not how complex the resistive network was; series, parallel, series+ parallel. One could calculate the Thevenin equivalent and end up with a circuit containing a single DC voltage source and single load resistor. Back when I was studying EE, one of our profs would delight in filling the chalkboard (what's that?) with a complex series/parallel circuit full of resistors. Our slide rules (what's that?) were kept busy computing the single equivalent resistor and voltage source.

PWK told me when he was in "cow college", he would be assigned to go thru a complex circuit to determine the polarity across a particular resistor. He would determine the polarity and then reverse the test leads at the last second before making the measurement! He said more often than not this was successful. But I digress...

The key to the power arguments proffered to date is "equivalent DC heating power". Once alternating current circuits were being designed, the engineer had to know how much power a particular resistor might dissipate. Rather than go off on another tangent concerning AC meters (average, RMS, peak, etc), suffice it to say that measuring the RMS value of an alternating waveform corresponds to the effective (DC) heating value. Using Ohm's law for AC circuits; Power=RMS voltage squared/load resistance in ohms.

A (theoretically perfect) DC power supply delivering 28.3 volts will cause 100 watts of power to be dissipated across an 8 ohm load. A (theoretically perfect) audio power amplifier delivering a 1 kHz sinewave @ 28.3 V RMS will cause 100 watts of power to be dissipated across the same 8 ohm load. The only correct wording to classify the AC power dissipated is to state that the power amplifier is delivering 100 watts of continuous average sinewave power. Of course, we assume the AC waveform is not clipping and the non-inductive load resistors are capable of dissipated several times the actual power.

Here's a link that approaches this overarching issue from the perspective of watt-seconds and joules.

http://www.eznec.com/Amateur/RMS_Power.pdf

My hair hurts...

Lee