RMS, Peak and PMPO: Restating Advertised Watts at One Test Condition
Only the continuous rating is measured against a defined test; program is conventionally twice continuous, peak is four times, and PMPO has no standard at all and typically runs 10 to 15 times the real figure. Match an amplifier at 1.0 to 1.6 times the speaker's continuous rating, because a clipping small amplifier destroys more tweeters than a large clean one ever will.
By the Speakers Guru Editorial Team
Of the three wattage figures printed on speaker boxes, one is measured and two are calculated from it. Continuous power, sometimes labelled RMS, comes from a real test: a shaped noise signal applied for a defined period until the driver either survives or does not. Program is conventionally twice that. Peak is conventionally twice program again. PMPO is not defined by anything and routinely lands at 10 to 15 times the continuous figure.
So a box shouting 1000 W is often a 75 to 100 W speaker, and the amplifier it wants delivers something between 100 and 160 W continuous. That ratio, 1.0 to 1.6 times, is the whole matching rule. It exists because an underpowered amplifier driven into clipping is the most reliable tweeter killer in audio.
Three numbers, one measurement
Continuous power handling comes from standardised noise tests. The consumer versions run band-limited pink noise with a 6 dB crest factor for a defined duration, and pro driver testing follows a similar shape over a longer window. The signal is deliberately unmusical because the point is thermal: how much heat can the voice coil shed before the adhesive lets go or the former deforms.
Program power is a convention, not a measurement. Doubling continuous is supposed to approximate what a speaker survives on real music, where the average power is far below the peaks. Nobody tests it. It is arithmetic applied to a tested number.
Peak is arithmetic applied to arithmetic. And PMPO, peak music power output, is the number you get when a marketing department multiplies until it likes the result. There is no test condition attached to it, which is precisely why it appears in the largest type on the carton.
The multipliers, and the Spec Drift Index
Restating every claimed figure at one reference condition is what the Spec Drift Index does. Define the reference as continuous power measured with band-limited noise at rated distortion, then divide the headline number on the box by the restated figure. That ratio is the drift.
Continuous (restated) Program (2x) Peak (4x) Typical max power printed Typical PMPO printed Spec Drift Index 25 W 50 W 100 W 100-150 W 250-400 W 4x to 16x 50 W 100 W 200 W 200-300 W 500-800 W 4x to 16x 75 W 150 W 300 W 300-400 W 750-1200 W 4x to 16x 90 W 180 W 360 W 350-400 W 900-1000 W up to 11x 100 W 200 W 400 W 400-600 W 1000-1500 W 4x to 15x 150 W 300 W 600 W 600-800 W 1500-2000 W 4x to 13x 200 W 400 W 800 W 800-1000 W 2000-3000 W 4x to 15x 300 W 600 W 1200 W 1200-1500 W 3000-4500 W 4x to 15x Find the headline figure on your box in the fourth or fifth column and read left to the restated continuous number. That left-hand figure is the one to take into the amplifier matching table below. The worked example is worth doing slowly. A coaxial pair advertised at 1000 W with a 90 W continuous rating buried in the small print has a Spec Drift Index of 11.1. It wants an amplifier producing 90 to 145 W continuous per channel. Bolt it to a 1000 W monoblock and you have bought a very expensive way to melt a voice coil.
The matching window: 1.0 to 1.6 times
Ratios below are amplifier continuous output divided by speaker continuous handling. Under about 0.7 you are relying on the amplifier never clipping, which nobody manages. Over about 2.5 you are relying on your own restraint on the volume knob, which some people manage.
Speaker continuous 25 W amp 50 W amp 75 W amp 100 W amp 150 W amp 200 W amp 300 W amp 25 W 1.0x ideal 2.0x thermal risk 3.0x excessive 4.0x excessive 6.0x excessive 8.0x excessive 12x excessive 50 W 0.5x clipping risk 1.0x ideal 1.5x ideal 2.0x thermal risk 3.0x excessive 4.0x excessive 6.0x excessive 75 W 0.33x clipping risk 0.67x clipping risk 1.0x ideal 1.33x ideal 2.0x thermal risk 2.7x excessive 4.0x excessive 100 W 0.25x clipping risk 0.5x clipping risk 0.75x marginal 1.0x ideal 1.5x ideal 2.0x thermal risk 3.0x excessive 150 W 0.17x clipping risk 0.33x clipping risk 0.5x clipping risk 0.67x clipping risk 1.0x ideal 1.33x ideal 2.0x thermal risk 200 W 0.13x clipping risk 0.25x clipping risk 0.38x clipping risk 0.5x clipping risk 0.75x marginal 1.0x ideal 1.5x ideal 300 W 0.08x clipping risk 0.17x clipping risk 0.25x clipping risk 0.33x clipping risk 0.5x clipping risk 0.67x clipping risk 1.0x ideal Use the restated continuous figures for both sides, never the peak or PMPO numbers. Cells marked thermal risk are usable if the amplifier gain is set properly; cells marked clipping risk are not fixable by being careful. Underpowering kills tweeters
Here is the mechanism. When an amplifier runs out of rail voltage the tops of the waveform flatten. A square wave carries far more average power than the sine it started as, and the harmonics it generates are all high frequency, so they land in the tweeter. A crossover cannot help, because the distortion products genuinely are high frequency content.
The result is that a 40 W amplifier driven hard into a 100 W speaker delivers more heat into the tweeter voice coil than a clean 120 W amplifier ever would. Every repair shop has a bin of tweeters that died this way, and almost none of them died from an amplifier that was too big.
Which is why the sensible error is upward. Choose the larger amplifier, set the gain properly, and stop turning up when the sound hardens.
Crest factor: your 100 W amp is delivering 4 W
Music has a crest factor, the ratio between instantaneous peaks and long-term average. It is large, and it means the average power an amplifier delivers is a small fraction of its rating even when the peaks are hitting the rails.
Signal Typical crest factor Average power from a 100 W amp at peak clipping What that means for the driver Sine test tone 3 dB 50 W The harshest thermal test you can run by accident Band-limited pink noise (test standard) 6 dB 25 W The signal continuous ratings are measured with Heavily compressed pop and EDM 8-10 dB 10-16 W Sustained thermal load; the genre that cooks coils Rock and mixed pop 12-14 dB 4-6 W Comfortable for any correctly matched pair Film soundtrack 15-18 dB 1.6-3.2 W Huge instantaneous demand, tiny average Unprocessed acoustic and orchestral 18-20 dB 1.0-1.6 W Headroom matters far more than power handling Match the amplifier to the peaks and the speaker to the average. This table explains why a 100 W speaker on a 150 W amp survives a whole evening of orchestral material and dies in twenty minutes of compressed dance music at the same volume setting. Reading a sheet that hides the continuous figure
Some manufacturers make you work for it. A few habits get you to the real number faster.
- Look for the words continuous, RMS, nominal or AES anywhere in the specification block. That figure is the one to use.
- If only max power appears, divide by 4 for a working estimate and expect to be within about 25%.
- If only PMPO appears, divide by 10 to 15 and treat the result as a guess. A product that only quotes PMPO is telling you something about itself.
- Check whether power handling is quoted per speaker or per pair. Per-pair figures are common on coaxial car speakers and double the apparent number for free.
- Cross-check against the magnet and voice coil. A 1 inch voice coil on a 20 oz magnet is not a 300 W driver whatever the box says; genuine high continuous ratings come with 1.5 to 3 inch coils and vented pole pieces.
- For a passive PA cabinet, compare the continuous figure against the driver complement. A single 12 inch woofer and a compression driver is a 250 to 400 W continuous box, not a 1000 W one.
Feeding the Headroom Calculator
Inputs are the speaker's restated continuous rating, the amplifier's continuous output into the load you have actually wired, and the dominant content type. Output is the headroom ratio, a verdict, and the Spec Drift Index for whatever headline figure you enter alongside it.
Enter the amplifier's output into your real load, not its best-case number. A receiver rated 100 W into 8 ohms with two channels driven is delivering 60 W with five channels running, and that changes which row of the matching table you are in.