I have never understood!

[DavidEdwinAston]

Hi Guys
Apparently the power output of my mono amps is, 170W into 8 ohms (<0.1% THD at 1kHz).
Does this mean that it, and indeed all amplifiers, will output less, or, in fact more, at all the other Herz levels, which it can achieve?
Sorry for the stupid question.

 

[Waxx]

No, it means 170w in 8ohms fullrange (any frequency within the range it serves), and at that volume the measured THD at 1kHz (test signa) is smaller than 0.1%. They did not measure the distortion at other frequencies (or did not publis it) but the power will be there.

 

[Keith_W]

Let's look at Amir's Benchmark AHB2 review, specifically at this graph. This is how to read it.

Horizontal scale: power output in Watts at 8 Ohm. Easy enough to understand, the more we go to the right, the more power the amplifier makes.
Vertical scale: THD+N Ratio. This one needs more explanation. It is the ratio of THD (Total Harmonic Distortion) and Noise to the output signal. It can be expressed in percentage or dB. You can switch between both by using a THD Calculator. Note there are two expressions, THD and Noise.

In a typical amplifier, noise remains constant. This is why the THD+N ratio appears to decrease as amplifier output goes up. It's not the noise that is going down, but the power that is going up.

At some point, the amplifier can't take it any more ... it can't supply enough current, and it starts to clip. When it does this, THD shoots through the roof. You can see it happening here at about 100W. In fact you can see the THD+N climbing very slowly from about 15W to 100W, at which point it takes off like a rocket.

You can gain more information if you probe the amp with different frequencies. Most amplifiers, including this one, can't deliver the same maximum power output before the amp starts to clip. Here you can see the distortion rising at different rates from 2W up, and some frequencies clipping earlier than others. What you want to see is all the frequencies bunched up together.

Maybe Amir has tested a hilariously bad amplifier. I don't know, I don't usually read amplifier reviews, I don't find them particularly interesting.

 

[mhardy6647]

Maybe Amir has tested a hilariously bad amplifier. I don't know, I don't usually read amplifier reviews, I don't find them particularly interesting.

Why, yes, as a matter of fact, he's tested a few...

For example:

Yaqin MC-84L Tube Stereo Amplifier Review

This is a review and detailed measurements of the Yaqin MC-84L Stereo Tube amplifier. It is on kind loan from a member and costs US $600 on Amazon. The picture doesn't do justice to the nice design of this amplifier. The transformer case in the back nicely contrasts with the chrome parts. The...

www.audiosciencereview.com

note the comparators in the borrowed graphic below.

 

[kemmler3D]

Good answers in this thread already. As you can see from the graphs shared, the amount of THD (distortion) changes depending on the frequency.

The amount of power it can output also tends to vary a little bit by frequency, but as @Waxx mentions it tends to be more consistent.

So basically when you see these specs, you can take the power number as it is, but for the THD number, you can usually assume it gets a bit worse at different frequencies.

 

[Soniclife]

Hi Guys
Apparently the power output of my mono amps is, 170W into 8 ohms (<0.1% THD at 1kHz).
Does this mean that it, and indeed all amplifiers, will output less, or, in fact more, at all the other Herz levels, which it can achieve?
Sorry for the stupid question.

You would need to see power testing of your amp to know if it can output full power at all frequencies, the spec only tells you what it tells you.

What is your amp.

 

[Soniclife]

The amount of power it can output also tends to vary a little bit by frequency, but as @Waxx mentions it tends to be more consistent.

Some class D amps struggle with high power and high frequencies, but not usually in a way that would be an audible issue.

 

[DVDdoug]

Apparently the power output of my mono amps is, 170W into 8 ohms (<0.1% THD at 1kHz).

I tend to skeptical of that too! It seems like MOST manufacturer's "fudge" the specs. If you really need to know, you need independent measurements.

 

[DavidEdwinAston]

Thank you all for your responses. It is Quad Platinum. Following on from the 909 series, it did seem to be their best effort at solid state. Oddly, I don't see it listed now, in their development history. Following on in the manual, it states, THD 0.006% (120W at 1 kHz). <0.1% (100W at 20 Hz - 20 kHz).

 

[mhardy6647]

Following on in the manual, it states, THD 0.006% (120W at 1 kHz). <0.1% (100W at 20 Hz - 20 kHz).

Well, there you go! At moderately "hi-fi" standards for THD, it's a 100 watt broadband ("full range") amplifier.
Is that bad?
Not really -- 170 watts only amounts to 2.3 dB more power than 100 watts. Not quite negligible, but not far from it, either.
and 120 watts is only about 0.8 dB more than 100 watts (which isn't much).
All else being equal -- to get some idea of whether that's true, more information is needed (e.g., a similar set of 'specifications' into a four-ohm load vs. an eight ohm load, and, of course, graphs of THD (or THD plus noise) vs. power at different frequencies as @Keith_W illustrated with a typical good amplifier.

 

[Chrispy]

No, it means 170w in 8ohms fullrange (any frequency within the range it serves), and at that volume the measured THD at 1kHz (test signa) is smaller than 0.1%. They did not measure the distortion at other frequencies (or did not publis it) but the power will be there.

How does a 1khz rating equate to full range?

 

[kemmler3D]

How does a 1khz rating equate to full range?

<0.1% (100W at 20 Hz - 20 kHz).

 

[Keith_W]

Since we have a few experts on ASR following this thread, maybe i'll ask this question. The reason I don't bother with amplifier reviews is because it doesn't really tell you what happens when an amplifier is connected to a speaker. A speaker has a variable impedance load, meaning the current demand is different at different freqs. We also don't know how much current will be demanded, since that depends on speaker sensitivity and preference for volume output. So I do the same stupid thing I always did when I was a subjectivist - borrow an amp, plug it in, turn up the volume, listen for clipping. At some point the volume will stop increasing and I will hear a lot of nastiness at high freqs. If that subjective volume is too low, I need a more powerful amplifier.

My question: is there a way to monitor amplifier output when it is connected to a speaker and playing at different volumes if your amp doesn't have a VU meter?

 

[MaxwellsEq]

The reason I don't bother with amplifier reviews is because it doesn't really tell you what happens when an amplifier is connected to a speaker

We do if we know enough data about the amplifier, especially how it behaves with more than one value of resistive load and how it behaves with a reactive load. Capacitance and inductance force the current and voltage to be out of phase with eachother so more current and voltage are needed than expected. If voltage swing and current capability (heat) are sufficient at multiple resistance and reactance and the output impedance of the amplifier is very low and non-reactive, the amplifier will sound the same regardless of speaker.

You are right about speaker sensitivity and efficiency, since these have significant impacts on how much power is needed.

 

[Scytales]

We do if we know enough data about the amplifier, especially how it behaves with more than one value of resistive load and how it behaves with a reactive load. Capacitance and inductance force the current and voltage to be out of phase with eachother so more current and voltage are needed than expected. If voltage swing and current capability (heat) are sufficient at multiple resistance and reactance and the output impedance of the amplifier is very low and non-reactive, the amplifier will sound the same regardless of speaker.

You are right about speaker sensitivity and efficiency, since these have significant impacts on how much power is needed.

Amir also tests amplifier on reactive loads since he has get an Audiograph load test. He is not the only one. The Germans of TestLab have been doing that for over 30 year for the Stereoplay or Audio magazines. The Italians at Audioreview have been using their own test fixture and signal (a four sines test tone) for the same purpose for over 30 years also. Peter Aczel used to use the Audiograph hardware for the now defunct magazine The Audio Critic. In France, renown test engineer Etienne Lémery (R.I.P), notorious torturer of amplifiers, especially pro ones, had devised his own test hardware and signal in order to measure amplifiers on reactive loads for publication in the magazines Sono Mag (pro market) and Le Haut-Parleur (one of the oldest and longest standing French electronic magazine, now defunct) from 1990.

There are most probably other sources of measurements of the same kind that I am no aware of.

There is a great deal of knowledge that have been assembled throughout decades about the issue of driving reactive loads.

 

[mhardy6647]

Since we have a few experts on ASR following this thread, maybe i'll ask this question. The reason I don't bother with amplifier reviews is because it doesn't really tell you what happens when an amplifier is connected to a speaker. A speaker has a variable impedance load, meaning the current demand is different at different freqs. We also don't know how much current will be demanded, since that depends on speaker sensitivity and preference for volume output. So I do the same stupid thing I always did when I was a subjectivist - borrow an amp, plug it in, turn up the volume, listen for clipping. At some point the volume will stop increasing and I will hear a lot of nastiness at high freqs. If that subjective volume is too low, I need a more powerful amplifier.

My question: is there a way to monitor amplifier output when it is connected to a speaker and playing at different volumes if your amp doesn't have a VU meter?

Well... you can use a good, old VOM or VTVM if you wish.


Neither a VU meter nor a proper meter's ballistics will tell you what you want to know, though.

Semi-off-topic, I've got to say, I never have trouble with audible clipping. Ever.

 

[Keith_W]

Well... you can use a good, old VOM or VTVM if you wish.

Umm, what is a VOM and VTVM?

Is there a better way, like somehow make a network of resistors, plug it into the mic input of my interface, and use REW to measure?

Semi-off-topic, I've got to say, I never have trouble with audible clipping. Ever.

You must be lucky. My very first amp (NAD 3020i) clipped all the time. The speakers were too small for the room, and I had a preference for loud volume.

My current solid state amps for the woofers also go into protection mode when I play too loud.

 

[Old_School_Brad]

Umm, what is a VOM and VTVM?

Is there a better way, like somehow make a network of resistors, plug it into the mic input of my interface, and use REW to measure?



You must be lucky. My very first amp (NAD 3020i) clipped all the time. The speakers were too small for the room, and I had a preference for loud volume.

My current solid state amps for the woofers also go into protection mode when I play too loud.

The classic case of "horses for courses" -different use cases lead to different outcomes and requirements.

I’ve experienced amplifier clipping as well, including with high-powered units. And each time it happened, it prompted me to upgrade.

 

[mhardy6647]

Volt-Ohm Meter and Vacuum Tube VoltMeter, respectively. The latter is a more sensitive and higher-impedance variant of the latter. A DMM (ahem, Digital MultiMeter ) mimics a VTVM, but lacks the visual continuity of a swinging analog meter.
A VU meter is (ahem, should be), in essence, a carefully calibrated AC voltmeter.
Volume Units are a well defined thing from the days of sending remote audio feeds for radio broadcast via telephone lines (thus "line level" ). The ballistics (rise and decay times) for a VU meter were also defined, so that the meter readout gave a practical and useful indicator for distortion (overmodulation).

An oscilloscope will show clipping very definitively, but if and only if the input signal is something interpretable, e.g., a single frequency sine wave. Sine waves ain't music, though (nor, of course, vice versa).

EDIT: Here's a fairly nice old Hewlett-Packard (true) RMS-reading AC voltmeter that lives in my basement.


Note the line-level decibel definition

 

[DavidEdwinAston]

I am thanking everyone who responds to the thread. However, I am so far out of my comfort zone that I can only assume you are all replying in English?