Return to quoting peak wattages?

[Nick Kay]

The short term peak output power of an amp isn't limited by the power available from the wall, because the amp power supply capacitors store power. Large enough capacitors can store a hundred or even a thousand times the power available out of the wall, the catch is that they can only deliver that power for very short periods before being drained. Unfortunately a few milliseconds is all it takes to toast a voice coil.

Bill, I know my math is fuzzy and I'm ignoring all sorts of factors, but the idea of an amplifier delivering 10dB crests on top of its rated RMS output sounds pretty fishy to me. You can overbuild power supplies all you want, but until I see an amp with properly measured and published specs that can do crests greater than 6dB over RMS, I'm going to remain skeptical.

 

[Mark Reccord]

Explain to me in your own words how an amplifier with a given output voltage limit V (i.e., typically the hard clipping limit of the power amp) driving a speaker with series resistance R can deliver more current than V/R.

Use for your explanation any waveform that a musical instrument is likely to produce.

Bill, I know my math is fuzzy and I'm ignoring all sorts of factors, but the idea of an amplifier delivering 10dB crests on top of its rated RMS output sounds pretty fishy to me. You can overbuild power supplies all you want, but until I see an amp with properly measured and published specs that can do crests greater than 6dB over RMS, I'm going to remain skeptical.

I know this wasn't directed at me but here's an explanation that can account for it:

For the 200W/2000W example (let's assume 8 Ohms) you'd spec the transformer at 200VA but spec the output voltage at idle to be 88VAC, instead of the 40VAC typically needed for 200W/8Ohms. Then put in a lot of filter caps and overspec your output devices. So under transient conditions, the output voltage could swing to 125V, which is indeed 2kW at 8Ohms. Of course, the amp can only sustain that for a very short period of time before the supply voltage sags to its much lower continuous power capability. So that factors in the RMS-peak (i.e. rail voltage) difference that every amp can do on a transient basis plus the dynamic headroom offered by setting your power supply up that way.

I'm not sure it isn't a good way to design an amp that will be required to deal with dynamic input signals.

 

[fdeck]

I once again refer you to post #127 and Bass Gear Mag. Mrs FitzMaurice and Gallien obviously explain it better than I could.
As for waveforms, they're almost irrelevant here. We're talking about attack/sustain ratio so an ADSR representation of a bass note would be much more useful and accurate.

"Waveform" simply refers to input voltage as a function of time... of your choosing to prove your point.

 

[fdeck]

I know this wasn't directed at me but here's an explanation that can account for it:

For the 200W/2000W example (let's assume 8 Ohms) you'd spec the transformer at 200VA but spec the output voltage at idle to be 88VAC, instead of the 40VAC typically needed for 200W/8Ohms. Then put in a lot of filter caps and overspec your output devices. So under transient conditions, the output voltage could swing to 125V, which is indeed 2kW at 8Ohms. Of course, the amp can only sustain that for a very short period of time before the supply voltage sags to its much lower continuous power capability.

What's the mechanism that causes that voltage sag to occur? Unless it is some specialized limiting circuit, all I can think of is the series resistance in the transformer windings, i.e., the tranny heats up until it burns out. And if it is a Class-AB design, it would be horribly inefficient.

Other possibilities avail themselves if you are talking about switchmode design, but I think they require an active limiter of some sort.

So that factors in the RMS-peak (i.e. rail voltage) difference that every amp can do on a transient basis plus the dynamic headroom offered by setting your power supply up that way.

I'm not sure it isn't a good way to design an amp that will be required to deal with dynamic input signals.

It's not. We'd be talking about blown power transformers instead of blown speakers here. A power transformer doesn't happily lower its output voltage when the going gets tough.

 

[Micky-D]

I personally prefer to know the output voltage capability of an amplifier at the terminals than anything else. Knowing that my choice of amplifier can do, say, 91.4volts at the terminals allows me to easily match devices. No current manufacturers to my knowledge continue this tradition, but measuring the amplifier's output capability yourself is not difficult.

But in all honesty, use your ears. After all, you are generating sound. I used to run D.A.S Aero's off a pair of EV P3000's. Incredible power, enough to destroy the loudspeakers in a flash, but the headroom was fantastic and by using your ears, a properly gain matched and impedance matched system and some common sense, nothing ever went wrong.

Those same Aero's off amplifiers with a 2000w Peak rating....as the engineer pushed it harder and harder to get the same clean transients of the EV's, more and more of the magic smoke came out. It's hard to put the smoke back in the cabinets, you know.

 

[akaTRENT]

simply +1

 

[Bob Lee (QSC)]

I caught the HINT by Tom Bowlus. I mentioned this a couple weeks ago in another thread Tom, what needs to happen is a few of the Techno Heads need to get together and actually WRITE a method of testing that allows TOOB and SS amplifiers to equally be compared.

Then say your testing staff puts the gear through an exact testing ritual.

The sad thing here is everybody likes to think RMS is this great measurement, it is not it is a calculated number .707 times PEAK. It is the heating factor of AC power compared to DC. Meaning .707 Volts of DC has the same heating capability as 1 Volt of AC.

It truly has no meaning in Audio amplification. But people here call it CLEAN POWER.

Clean power is the rating the amplifier is given in "watts peak" with a frequency response say "20Hz to 20KHz" for example at a certain distortion limit say "<.05%"

That is a real rating. See the difference between that and 300 Watts RMS which actually means nothing as it is a calculated number at the BEST FREQUENCY and BEST POWER OUTPUT with no distortion spec?

As I continue here, the amp power should be compared at a certain temperature rise over a certain time period. Most BIG PA amps are rated over 2 hours a typical engagement; with certain crests factors. This is not easy to measure stuff. I think computer simulations and automated test rig is in order.

You wanna see how this works? AMPEG SVT-4PRO 1600 watts peak. A measured number, in the amps ratings. 1200 Watts RMS a calculated number.

1600 times .707 = 1131.2 Watts. OH lookie here somebody padded a spec. Shame on them.

BOB

Technically, there is no such thing as "RMS power." AC values are expressed in RMS voltage and RMS current, but the product of them is average power. "RMS" is a method of measuring AC voltage or current in terms equivalent to DC. 0.707 V DC does not equal 1 V AC, but 0.707 V RMS. With a constant sinusoidal signal, the RMS voltage is 0.707 × the peak voltage, so a 0.707 V sine wave would have peaks of 1 V.

Continuous average power would be measured with a continuous sine wave signal. The RMS voltage would be measured into a known load resistance, and the power would be calculated using P = E^2 / R.

Amp power ratings involve running the amp's output up to the point where it barely starts clipping (this is defined by the THD figure stated in the power measurement), and then measuring the RMS voltage into the load. Power measurements using the FTC Amplifier Rule require a preconditioning procedure that takes at least 65 minutes. EIA power specs require no preconditioning.

I don't know of any reputable power amp manufacturer that bothers with "peak" power but hey, in tough times some people break in surprising ways.

 

[fdeck]

Continuous average is good enough for me, if it is spec'd with a relatively low distortion figure. From there, I can figure out the peak output voltage, and can make an educated guess about the additional reserve for transient peaks, that probably gets me within 1-2 dB of reality on mainstream solid state amps.

I would be doubtful that any mainstream amp has a magical reserve of extra power that audibly exceeds what I would estimate from rules of thumb.

 

[rbonner]

Technically, there is no such thing as "RMS power." AC values are expressed in RMS voltage and RMS current, but the product of them is average power. "RMS" is a method of measuring AC voltage or current in terms equivalent to DC. 0.707 V DC does not equal 1 V AC, but 0.707 V RMS. With a constant sinusoidal signal, the RMS voltage is 0.707 × the peak voltage, so a 0.707 V sine wave would have peaks of 1 V.

Continuous average power would be measured with a continuous sine wave signal. The RMS voltage would be measured into a known load resistance, and the power would be calculated using P = E^2 / R.

Amp power ratings involve running the amp's output up to the point where it barely starts clipping (this is defined by the THD figure stated in the power measurement), and then measuring the RMS voltage into the load. Power measurements using the FTC Amplifier Rule require a preconditioning procedure that takes at least 65 minutes. EIA power specs require no preconditioning.

I don't know of any reputable power amp manufacturer that bothers with "peak" power but hey, in tough times some people break in surprising ways.

I agree totally with what you are saying compared to reading it with a meter or reading it with a SCOPE. You also know how they do it in the sound industry to rate gear, which I currently don't.

The big definition is located here: [Invalid or Expired Link Removed]

Music is a complex waveform. I really don't want to discuss the whole thing this deeply, but it is a mess.

Thanks for bringing that up Bob.

BOB