Skip to main content
Topic: Is power amp output constant? (Read 1280 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Is power amp output constant?

Hi everyone,

Excuse me for the potentially stupid question but I have to ask. Before I go on... I want to say the answer to my question is NO but here goes:

Is the output wattage of a power amplifier constant regardless of volume/input levels?

In context, I have a set of 6ohm speakers, rated at 250w continuous, 500 peak. The amp I have attached to them is a QSC GX5, which provides 500w cont. into 8ohms each side. So lets say something like 600w cont. into each side.

Now this obviously exceeds the speakers rating by miles. However I am running the QSC GX5 volume at 50%. This is fed by a Cambridge 851D pre amp which I can never turn the volume up full on, normal listening will vary between -30dB and -15dB on the Cambridge display. At the -15dB position, my meter reads 97dB from a couple of meters away, so its loud.

The long version of my question is:

Because I have the power amp volume set to 50% and am limiting even further via pre amp volume control, the amplifier is not providing anywhere near its rated wattage regardless of the volume settings it is?

In other words, I am not killing my speakers am I? They are fused I suppose.

Thanks
Tom
Hi-Fi: Audio Technica AT-LP5 Turntable | Cambridge Audio CXC CD Transport | 851N DAC/Streamer | 851W Pre-amp | 840W Power-amp | Cerwin Vega XLS215 Speakers

Re: Is power amp output constant?

Reply #1
No. Your speakers have an efficiency specified in dB/watt. Lower the dB and you use less watts.

Re: Is power amp output constant?

Reply #2
No, it is not constant. It is not a light bulb.  ;D

The more power you feed into the speakers, the louder they will be.

The wattage label on the amp tells you, how much power the amp can handle at maximum without overheat or clipping.
The wattage Label on the speakers tells you, how much power they can consume without damages.
And forget these peak values. These are nothing but marketing lies. Only take the RMS values into account.

These values are just references that help you to select a good combination of amp and speakers. Even if the components fit together perfectly, never crank up the volume knob to the upper end.

There are some cases where a 50 Watt RMS amp will blow a 100 Watt RMS speaker (very high sine waves, heavy clipping, DC-offset...)

You actually hear, if a speaker is in danger. As long as the music is sounding fine and without distortion, you usually are safe.  ;D

If you notice signs of overload, turn down the volume suddenly.

Most speakers die by clipping amps  (exceeding the max. power of a too weak amp), faulty amps or by mechanical overloading the speaker cones (amp is too strong, leads to audible distortion and loud popping noises too.).

Really important: You must not hook up speakers with lower impedance (ohms) to an amp that requires higher impedance speakers. In this case, treat missleading "4-8 ohms"-labelled speakers as 4 ohms. That can destroy the amp (and the speakers too)
- I abandoned this account since I didn't find a way to delete it -

Re: Is power amp output constant?

Reply #3
Right!  It depends on how loud you are listening....  Turn the volume down to zero and you've got zero-watts.  ;)   Turn it down to near-silence and you've got nearly zero-watts...  

Typically, you're safe "at home" unless a teenager or a drunk person is in charge of the volume control.   Or, if you set-up your speakers in a dance hall you might over-drive them without realizing it.   Or, with a high-power amplifier you could potentially damage something if you pull an RCA plug half-way out and you get that loud "buzz".

Quote
They are fused I suppose.
Usually not...  Fuses have resistance so it would affect speaker performance and the nature of music (or "program material") makes it tricky.  Music typically has a peak-to-average power ratio of about 10:1, so if you're hitting 100W on the peaks you're averaging about 10W.  And, it's the (short-term) average that will burn-out the speakers.  

Further complicating things is that with a 2 or 3-way speaker the energy is divided-up and most of the energy is in the lower & mid-lower frequencies.

Quote
I have a set of 6ohm speakers, rated at 250w continuous, 500 peak.
I'm pretty sure you speaker can't handle continuous 250W test-tones...   And, tweeters can't handle as much power a woofers so it's not hard to fry a tweeter with high-frequency test tones.      Different manufacturers use different standards, but if you trust the manufacturer's specs, a 250W speaker should be  safe with a 250W amplifier, assuming normal program material and assuming you're not driving the amplifier into clipping (distortion). 

JBL has a paper that says it's safe to use an amplifier with twice the speaker's power rating.   (That assumes honest IEC power ratings for the speaker.)   But for applications like guitar amplifiers where the amplifier is likely to be driven into distortion, they flip that around and recommend a speaker with twice the power rating of the amp.


Re: Is power amp output constant?

Reply #4
Thanks all,

To add some specifics the speakers I am talking about are Cerwin Vega XLS215's. Yes we all know the reputation CV speakers have, but they are not my only system and when I saw a set available in the UK, I had to grab them. I have to say I am impressed, curiosity made me want a set... (and that 10 years ago when I was 18, I think they looked cool as hell).

I am not one of those people that makes woofer excursion videos for youtube either. But I am a die hard rock/metal fan, well and a bit of various other genres.

So I guess we can say that even though my amp is more than capable than exceeding these speakers RMS rating, because I am being sensible with the volumes, I don't have a problem as the wattage is safe.
Hi-Fi: Audio Technica AT-LP5 Turntable | Cambridge Audio CXC CD Transport | 851N DAC/Streamer | 851W Pre-amp | 840W Power-amp | Cerwin Vega XLS215 Speakers

Re: Is power amp output constant?

Reply #5
No, it is not constant. It is not a light bulb.  ;D

The more power you feed into the speakers, the louder they will be.
The more power you feed into a light bulb the brighter it will be.  A speaker is exactly like a lightbulb, except the power company supplies a (usually) stable and fixed voltage, resulting in stable and fixed current/power/brightness.
The wattage label on the amp tells you, how much power the amp can handle at maximum without overheat or clipping.
Amplifier specs state the amount of power an amplifier can deliver to a particular with a specified amount of distortion.
The wattage Label on the speakers tells you, how much power they can consume without damages.
And forget these peak values. These are nothing but marketing lies. Only take the RMS values into account.
Speaker maximum power figures are usually maximum continuous RMS power.  Some professional speakers even define what type of signal at what power is safe.  Speaker maximum is always an RMS vs time figure, even if they don't state what time is.  Speakers failing because of high power heating damage have a quite variable RMS power vs time value, whereas speakers that fail because of over-excursion may fail quite quickly.

These values are just references that help you to select a good combination of amp and speakers. Even if the components fit together perfectly, never crank up the volume knob to the upper end.
Knob position doesn't define maximum power.  Maximum power is achieved by the combination of applied signal and system gain, including everything in the chain.  The volume knob varies gain.
There are some cases where a 50 Watt RMS amp will blow a 100 Watt RMS speaker (very high sine waves, heavy clipping, DC-offset...)
BIG misconceptions here!  Speakers are damaged in two ways: RMS power heating and damaging the voice coil, and applied power causing over-excursion of the speaker voicecoil to the point where it collides (and possibly damages itself) with some other structure of the speaker, usually the magnet.  You can apply that RMS power by clipping an amp, or not clipping an amp, the RMS power applied is essentially the same up to a point, then increasing the amount of clipping produces less RMS power than the equivalent unclipped signal.  A clipped amp does not produced a DC offset, and speakers are not blown by clipping more than they would be by the equivalent unclipped waveform.  Part of the misconception that clipped audio blows speakers is the assumption that clipped audio contains much more HF energy because of clipping.  This too is false, as the harmonics created by clipping audio do not make up a significant portion of the total energy spectrum, and in fact do not affect the total HF RMS unless taken to the extremes of clipping at extreme HF, but even then, the difference as compared to a non-clipped energy spectrum is quite minimal. 
You actually hear, if a speaker is in danger. As long as the music is sounding fine and without distortion, you usually are safe.  ;D
The onset of clipping, even hard clipping, is usually not heard until the clipping threshold has been past by more than a few dB, where it becomes obvious, program material dependent of course.
If you notice signs of overload, turn down the volume suddenly.

Most speakers die by clipping amps  (exceeding the max. power of a too weak amp),
False.  See above.
faulty amps or by mechanical overloading the speaker cones (amp is too strong, leads to audible distortion and loud popping noises too.).
Yes, and that "mechanical overloading" is the result of applying excessive power (clipped or not).
Really important: You must not hook up speakers with lower impedance (ohms) to an amp that requires higher impedance speakers. In this case, treat missleading "4-8 ohms"-labelled speakers as 4 ohms. That can destroy the amp (and the speakers too)
Pretty much nonsense.  No speaker presents a real 8 ohm or 4 ohm load, they all have variable impedance vs frequency.  There are plenty of 8 ohm speakers that dip below 4 ohms at some frequency.  Speaker impedance figures are "nominal" only.  The fact that a speaker's impedance dips below 8 or 4 ohms will not destroy an amp or a speaker, especially if the amp is operated within its rated power range.  Nearly every amp today will protect itself against overload long before it is damaged, and as stated above, RMS power burns out speakers, clipped or not.

Re: Is power amp output constant?

Reply #6
JBL has a paper that says it's safe to use an amplifier with twice the speaker's power rating.   (That assumes honest IEC power ratings for the speaker.)   But for applications like guitar amplifiers where the amplifier is likely to be driven into distortion, they flip that around and recommend a speaker with twice the power rating of the amp.
The problem with drawing that conclusion from the paper is that it it doesn't follow how RMS energy increases above the clipping threshold. 

Here's a set of graphs of actual RMS power (measured in various ways) of a contemporary tune (mastered with typical loudness war processing) as play level is advanced, the resulting power clipped vs unclipped.  You can clearly see that applying the same gain change without clipping results in higher RMS power applied to the load.  Higher RMS power causes heat damage.



The unclipped graphs would represent amplifiers with significantly higher output power (the chart ends +10dB above clipping threshold). 

However, this analysis also assumes a well-behaved clipped amplifier that doesn't freak out and start oscillating when it's clipped.  HF oscillations cannot be predicted simply by analyzing the RMS value of clipped waveforms.

Re: Is power amp output constant?

Reply #7
Quote
A speaker is exactly like a lightbulb, except the power company supplies a (usually) stable and fixed voltage, resulting in stable and fixed current/power/brightness.
That "except" is, what a loudspeaker differs from a light bulb. Or are you listening to 50Hz-hum? And since when light bulbs have impedances (with "light bulb" I mean the traditional ones with a glowing wire inside)?

Quote
BIG misconceptions here!  Speakers are damaged in two ways: RMS power heating and damaging the voice coil, and applied power causing over-excursion of the speaker voicecoil to the point where it collides (and possibly damages itself) with some other structure of the speaker, usually the magnet.

...and where IS the misonception here? The the three kinds of "dangerous" sounds (these were just examples. I skipped the one with the over-excursion in lower frequencies) I mentioned actually lead to that what you say here even the amp is much weaker than the speaker. So it is not wrong I think (the tweeter is only 10-20% of the overall RMS wattage and gets killed easily by strong high frequencies).

Quote
Pretty much nonsense.  No speaker presents a real 8 ohm or 4 ohm load, they all have variable impedance vs frequency...
What do you think the impedance label on a speaker is for, if not to find a suitable amp that can handle the speakers and to prevent that non-technical customer hooks up speakers that the amp cannot handle well? I don't think an average customer could read an accurate impedance per frequency chart.

Quote
Knob position doesn't define maximum power.  Maximum power is achieved by the combination of applied signal and system gain, including everything in the chain.  The volume knob varies gain.
Why are you telling that to me? That was the reason, why I've said to not to turn up the volume to maximum.

Sorry... I didn't write my posting in a petty scientific detailed way. I just answered tomstephens89's question with that posting.  ;)
- I abandoned this account since I didn't find a way to delete it -

Re: Is power amp output constant?

Reply #8
Quote
A speaker is exactly like a lightbulb, except the power company supplies a (usually) stable and fixed voltage, resulting in stable and fixed current/power/brightness.
That "except" is, what a loudspeaker differs from a light bulb. Or are you listening to 50Hz-hum? And since when light bulbs have impedances (with "light bulb" I mean the traditional ones with a glowing wire inside)?
They both respond by heating resulting from RMS power.  Both "have impedances", both will end up glowing with enough power applied, and both will burn out with too much power applied, without regard to frequency. 
Quote
BIG misconceptions here!  Speakers are damaged in two ways: RMS power heating and damaging the voice coil, and applied power causing over-excursion of the speaker voicecoil to the point where it collides (and possibly damages itself) with some other structure of the speaker, usually the magnet.

...and where IS the misonception here? The the three kinds of "dangerous" sounds (these were just examples. I skipped the one with the over-excursion in lower frequencies) I mentioned actually lead to that what you say here even the amp is much weaker than the speaker. So it is not wrong I think (the tweeter is only 10-20% of the overall RMS wattage and gets killed easily by strong high frequencies).
The misconception is that a clipped signal somehow burns out a speaker more easily than an equivalent RMS power non-clipped audio signal.  Not true. Tweeters have 10-20% of the total RMS capacity because music spectrum in their range is also 10-20% of the total RMS energy. 
Quote
Pretty much nonsense.  No speaker presents a real 8 ohm or 4 ohm load, they all have variable impedance vs frequency...
What do you think the impedance label on a speaker is for, if not to find a suitable amp that can handle the speakers and to prevent that non-technical customer hooks up speakers that the amp cannot handle well? I don't think an average customer could read an accurate impedance per frequency chart.
It doesn't matter.  For any given application where both speaker and amp are used for their designed and intended purpose, they match, by default, design, and intent.  You can't damage an amp by attaching the "wrong" speaker.  Even attaching more than one speaker won't damage an amp, though it may run hot and shut down under thermal protection, but not at average living-room listening levels.  I'm sure there may be some cheap amp designed without any protection at all, but that's not typical.
Quote
Knob position doesn't define maximum power.  Maximum power is achieved by the combination of applied signal and system gain, including everything in the chain.  The volume knob varies gain.
Why are you telling that to me? That was the reason, why I've said to not to turn up the volume to maximum.

Sorry... I didn't write my posting in a petty scientific detailed way. I just answered tomstephens89's question with that posting.  ;)

You imply the knob position is responsible for total power output.  It's a part of it, but certainly not the whole story.   Sometimes lack of detail creates inaccuracy.

Re: Is power amp output constant?

Reply #9
Ok thanks all for the detailed information, even if it does look like Ive started a war.

So in laymens terms, even though my amp exceeds the speakers power handling rating. This doesn't matter since I understand this and will never be turning all the volume knobs up enough to get anywhere near the amps max rating.
Hi-Fi: Audio Technica AT-LP5 Turntable | Cambridge Audio CXC CD Transport | 851N DAC/Streamer | 851W Pre-amp | 840W Power-amp | Cerwin Vega XLS215 Speakers

Re: Is power amp output constant?

Reply #10
Ok thanks all for the detailed information, even if it does look like Ive started a war.

So in laymens terms, even though my amp exceeds the speakers power handling rating. This doesn't matter since I understand this and will never be turning all the volume knobs up enough to get anywhere near the amps max rating.

Pretty much that.

The most important component relating to speaker reliability is the nut whose fingers operate the volume control. :-)

One key piece of advice - if you turn things up to the point where things start sounding bad, back the volume off! Generally speakers and amps sound bad before things start getting hurt.

Re: Is power amp output constant?

Reply #11
In other words, I am not killing my speakers am I? They are fused I suppose.

You have a point (if I interpret you correctly) that for an arbitrary combination of pre/power then an "11 o'clock" volume knob setting might not tell you much about how loud you are playing - but you are not really pushing limits here.

What you measure is 97 dB at a couple of meters' distance.  Sounds loud, but is not that much power really, and consider your setup:
A "semipro" fan-cooled amplifier with some circuitry designed to protect itself and the amp without shutting off;
A set of fairly high efficiency "top of the line party speakers" (at least traditionally, designed to handle even more badass volumes than C-V!'s economy line).


But I am a die hard rock/metal fan, well and a bit of various other genres.
Hope they sound better than the C-V!'s that people had when I grew up. Impressive sound on action movies, awful on metal. But now C-V! is bought by Gibson, one could hope that they do not ruin electric guitars.
Memento: this is Hydrogenaudio. Do not assume good faith.

Re: Is power amp output constant?

Reply #12
Quote
The misconception is that a clipped signal somehow burns out a speaker more easily than an equivalent RMS power non-clipped audio signal.  Not true.
It's true, because:
Quote
Tweeters have 10-20% of the total RMS capacity because music spectrum in their range is also 10-20% of the total RMS energy. 

Please think about, what a clipping amp actually does to the spectrum and why it brings the tweeters in danger. That are basics. :P

- I abandoned this account since I didn't find a way to delete it -

Re: Is power amp output constant?

Reply #13
Quote
The misconception is that a clipped signal somehow burns out a speaker more easily than an equivalent RMS power non-clipped audio signal.  Not true.
It's true, because:
Quote
Tweeters have 10-20% of the total RMS capacity because music spectrum in their range is also 10-20% of the total RMS energy. 

Please think about, what a clipping amp actually does to the spectrum and why it brings the tweeters in danger. That are basics. :P
Here's some actual analysis.  The track analyzed was initially an mp3, which worked out well because the 15kHz LPF lets us see the actual spectrum added by clipping above 15kHz.  Note the spectrum curves are each clipped progressively harder, up to a completely intolerable amount where peaks exceed hard clipping by 10dB. 

Other things to note: the track is a contemporary piece with loudness-war processing normalized to just shy of 0dBFS with very little DR.  That means when it's clipped, lots of energy is clipped for lots of time, which is very different from a higher DR piece clipping briefly.   That makes this more of a "worst case" test signal. Note that the overall spectrum doesn't change with additional clipping, and that the only visible result occurs in the 15-20kHz range where the track audio was filtered.  Then, of course, note the level of those clipping artifacts about 15kHz.

The bottom red trace is unclipped, each trace above is an additional 2dB above the onset of hard clipping.



Re: Is power amp output constant?

Reply #14
In other words, I am not killing my speakers am I? They are fused I suppose.

You have a point (if I interpret you correctly) that for an arbitrary combination of pre/power then an "11 o'clock" volume knob setting might not tell you much about how loud you are playing - but you are not really pushing limits here.

What you measure is 97 dB at a couple of meters' distance.  Sounds loud, but is not that much power really, and consider your setup:
A "semipro" fan-cooled amplifier with some circuitry designed to protect itself and the amp without shutting off;
A set of fairly high efficiency "top of the line party speakers" (at least traditionally, designed to handle even more badass volumes than C-V!'s economy line).


But I am a die hard rock/metal fan, well and a bit of various other genres.
Hope they sound better than the C-V!'s that people had when I grew up. Impressive sound on action movies, awful on metal. But now C-V! is bought by Gibson, one could hope that they do not ruin electric guitars.

I think most people would be suprised by the more modern CV's, especially the XLS series. These XLS215 for example... I wasn't expecting them to sound as good as they do. I was half expecting flabby bass and boomy bottom end, however this is not the case at all. Music is expressed with a great 'fullness', bass is tight and responsive. In my opinion they make for some very fun listening.

Hi-Fi: Audio Technica AT-LP5 Turntable | Cambridge Audio CXC CD Transport | 851N DAC/Streamer | 851W Pre-amp | 840W Power-amp | Cerwin Vega XLS215 Speakers

Re: Is power amp output constant?

Reply #15
I think most people would be suprised by the more modern CV's, especially the XLS series. These XLS215 for example... I wasn't expecting them to sound as good as they do. I was half expecting flabby bass and boomy bottom end, however this is not the case at all. Music is expressed with a great 'fullness', bass is tight and responsive. In my opinion they make for some very fun listening.

It was the midrange I really hated. (Well there was one exception, a revised version of the "1215", named after its two woofers - but I later learned that it was an oddball joint venture with some other manufacturer and had a different midrange element from Peerless.)
And it wasn't the bottom end that was boomy really, it was - again, on their economy range - an octave above.

This is a bottom-end test, though not for dynamics: https://tmfdc.bandcamp.com/album/anthropomorphic
Use first 39:50 to 40:00 or around 49:00 to set the volume.  Then to work: from 2:30 and three minutes on (well more, but the 40 Hz drone does not change much for a while). Here is where one could worry over amps and speakers even if it does not sound too loud. (There was a reason for those subsonic filter buttons of the past.)
Memento: this is Hydrogenaudio. Do not assume good faith.

Re: Is power amp output constant?

Reply #16
@dc2bluelight:
I'm not talking about digital source clipping, which is indeed harmless in most cases (including yours at 48kHz). Electric guitar amps and many audio effects do similar things. ::)

I'm talking about clipping occuring inside an overdriven amp where no nyquist frequency, no lowpass and no dc filters will prevent the dangerous overtones and the dc components from passing the voice coils.. :)

Imagine,  there is no upper frequency limit in your spectogram. Or redo your experiment in a extremely high sample rate and compare again. The range above 15 kHz is the interesting one here. Is it still below 20% of overall power after applying clipping? ::)

If you still don't believe me, please read or watch one of the explainations here: http://lmgtfy.com/?q=clipping+dangerous
- I abandoned this account since I didn't find a way to delete it -

Re: Is power amp output constant?

Reply #17
@dc2bluelight:
I'm not talking about digital source clipping, which is indeed harmless in most cases (including yours). Electric guitar amps and many audio effects do similar things. ::)

I'm talking about clipping occuring inside an overdriven amp where no nyquist frequency, no lowpass and no dc filters will prevent the dangerous overtones and the dc components from passing the voice coils.. :)

Imagine,  there is no upper frequency limit in your spectogram. Or redo your experiment in a extremely high sample rate and compare again. The range above 15 kHz is the interesting one here. Is it still below 20% of overall power after applying clipping? ::)

If you still don't believe me, please read or watch one of the explainations here: http://lmgtfy.com/?q=clipping+dangerous

You haven't linked to a specific video so I can't respond in specific.  However, the first google hit links to a video of a guy driving an amp into clipping and measuring the resulting power increase.  What he doesn't show (and my initial graph does) is how that same drive, unclipped, produces even more power than a clipped wave.

I'll post the full analysis up to 48kHz later...but you won't find speaker-burning energy above 20kHz due to clipping.  To understand that, look at the total spectrum graph and not where the bulk of the energy is.

Re: Is power amp output constant?

Reply #18
@dc2bluelight:


If you still don't believe me, please read or watch one of the explainations here: http://lmgtfy.com/?q=clipping+dangerous


Wonder of wonders!

The first video shows that if you turn up the volume and get clipping, the signal contains more power

The second video beats around the bush for quite a bit longer to show something similar.

Yawn!


Re: Is power amp output constant?

Reply #19
Here's another progressive clipping spectrum set.  This original was a 24/48 file, this time resampled to 192kHz.  As before, the bottom red trace is the original spectrum unclipped, then each successive trace above is made with the original driven above hard clipping in +2dB steps. 


Again, notice, no basic spectrum change during progressive clipping, and though some additional ultrasonics are generated, they are at worst -55dB below the highest point in the audible spectrum.  For reference, if this "amplifier" were clipping hard at 100W the highest ultrasonic energy is less than 1mW. 

Now, of course, amplifiers may not clip this hard.  I haven't bothered to simulate that yet, but it would simply result in lower levels of ultrasonic content.

The (almost universal) assumption that clipping generates large amounts of high frequency energy is based on assumption that HF energy is high to begin with, which if you look at the spectral energy distribution, you can see, it is not.  No real surprise, most energy in music is below 1kHz.  So when you clip to the point where it sounds bad you're really not clipping HF energy anyway...at all.   Second, there's a universal assumption that clipping amplifiers generate square waves.  This is absolutely not true, even when pushed 10dB past their clipping point. It's still audio, and a lot of it is not clipped.  But as a sanity check, here's the spectrum of a 440Hz square wave at -10dBFS:

Note the level of harmonics above 1kHz, or 3-5kHz that would hit a tweeter: the first harmonic above 3kHz is -15dB below the fundamental.  Nothing to blow a tweeter there.  But, as I said, clipping an amp does not produce square waves, it produces clipped audio waves...not the same thing at all.

The "square wave" assumption leads to the assumption that DC is created by the resulting wave form.  Again, not true.  Clipped audio still AC, even a square wave is AC, and if perfectly symmetrical, contains no DC component.  In decades past it was possible for amplifiers to clip asymmetrically, but those days are pretty much gone now.  Even in the past, it was rare.  (In the clipped audio simulation above, there was no DC block used, and yet the DC offset after 10dB of clipping was analyzed at 0% of total peak. 

So, no square waves, no DC, no additional HF, so what about clipping blows speakers?  It's the additional RMS power.  What would a non-clipping amplifier driven to the same level produce? Even more RMS power, and even more blown speakers.

The one anomaly that is not analyzed here is an unstable amplifier that oscillates when over-driven. That's real, but also rare.  If an amp oscillates when clipped into a reactive load (like an old Crown DC300!), then yes, the resulting ultrasonics might be pretty darn hot and toast a tweeter.  Amps today, even the cheapies, are generally quite stable, thankfully. So it's a non-issue today.

Let me make a sort of pre-emptive statement here about clipping amps and active crossovers.  The analysis here is for a full-bandwidth amp followed by a passive crossover.  What happens if  you use an active crossover and clip the tweeter amp? You blow the tweeter, right?  Possibly, but consider that in the energy distribution shown here, if we crossed at 3kHz, the energy above 3kHz is at least 15dB below the energy below 3kHz.  So if you needed a 100W woofer amp, a 10W tweeter amp would give you about 3X as much reserve power as you'd need.  So you aren't going to clip an HF amp after an active crossover, even though you might still blow a tweeter with too much RMS  power.



Re: Is power amp output constant?

Reply #20


So, no square waves, no DC, no additional HF, so what about clipping blows speakers?  It's the additional RMS power.  What would a non-clipping amplifier driven to the same level produce? Even more RMS power, and even more blown speakers.

The one anomaly that is not analyzed here is an unstable amplifier that oscillates when over-driven. That's real, but also rare.  If an amp oscillates when clipped into a reactive load (like an old Crown DC300!), then yes, the resulting ultrasonics might be pretty darn hot and toast a tweeter.  Amps today, even the cheapies, are generally quite stable, thankfully. So it's a non-issue today.

Let me make a sort of pre-emptive statement here about clipping amps and active crossovers.  The analysis here is for a full-bandwidth amp followed by a passive crossover.  What happens if you use an active crossover and clip the tweeter amp? You blow the tweeter, right?  Possibly, but consider that in the energy distribution shown here, if we crossed at 3kHz, the energy above 3kHz is at least 15dB below the energy below 3kHz.  So if you needed a 100W woofer amp, a 10W tweeter amp would give you about 3X as much reserve power as you'd need.  So you aren't going to clip an HF amp after an active crossover, even though you might still blow a tweeter with too much RMS  power.

First let me praise you for going the extra mile with an arrogant, poorly informed attacker. Your original form of this evidence was more than good enough and all of his complaints suggested true belief in the usual audiophile myths.

Your points are relevant and true.

For example, the old EV T35 tweeter had a power rating of just 5 watts but also had 109dB/W sensitivity. It was commonly used with woofers with sensitivities of 15-20 dB less. To make up the difference, there was a resistive attenuator in series with it that provided the required attenuation and also had the effect of significantly magnifying the power handling capacity of the tweeter from the amplifier viewpoint. It looked to the amp like a tweeter with up to 500 watt power handling capacity.

The modern active speaker with full DSP support pretty well moves the discussion into a different universe - a universe without clipping. It is now pretty common to monitor the power amp for clipping, and simply turn down the gain when even just tiny amounts of clipping are discovered.

 

Re: Is power amp output constant?

Reply #21
I was thinking of posting a few spectra of an actual clipping amp, but it wouldn't look much different.  And then we'd get into specific amps, specific tweeters, and blah blah...  The general principles work just fine.  And the myth marches on anyway. It's one of the older ones.  I actually think it dates back to the Crown DC300 days.  Crown made an active crossover and hifi speakers back then, and someone recommended a higher power amp for the tweeter, a lower for the others (I think 3 way?), but it was likely because those amps didn't just clip, they became oscillators with the reactive load. Hence the myth was born.  Glad the 1970s are over!

 
SimplePortal 1.0.0 RC1 © 2008-2018