WATTS: Derate for 1khz vs music. Derate for Class A vs D.

[Sokel]

What is your take on the SA video? I think he saying the Fosi V3 gets to much higher watt output when fed a 1khz tone vs pink noise.

I think Erin is saying something similar - what is your take on Erins discussion of 1khz tone vs multitone test signal?

I think currently Amir only uses multitone to test THD. Can one determine multitone power output from that?
If not, can he test output power with multi tone and with pink noise?

I think Erin tested output power with multitone and with 1khz, but I am not sure.

Have a look at this thread:

Power amplifier tests with respect to FTC: 16 CFR Part 432 (July 5, 2024) requirements on output power claims

Some of us feel the need of honest specifications of amplifier output power according to acknowledged documents like FTC regulations and IEC standards. The reason is to get comparable data, protect potential customers and draw attention to false claims of some manufacturers. I have decided to...

www.audiosciencereview.com

 

[voodooless]

And we do this 1) with a 100khz input tone. And 2) we do it again changing the input to pink noise. I am saying the output in case 2 will often be less, and sometimes way less.

100kHz? How about 1kHz

This indicates that there should not be a lot of issues reproducing music with enough power.

That doesn't mean there are no amps that have issues with power delivery, for instance, the recent Topping B100, but the issues here are of thermal nature. And this is also why @pma (among others) would like to see more advanced (and standardized) and longer-lasting power tests:

Power amplifier tests with respect to FTC: 16 CFR Part 432 (July 5, 2024) requirements on output power claims

Some of us feel the need of honest specifications of amplifier output power according to acknowledged documents like FTC regulations and IEC standards. The reason is to get comparable data, protect potential customers and draw attention to false claims of some manufacturers. I have decided to...

www.audiosciencereview.com

What Erin saw with the ZA3 is a similar issue, but this time, it's more likely a power supply issue. In both cases, this is due to the amp needing to deliver sustained power for a longer period. This has zero to do with the amplifier class, as the OP suggested. But these budget amps do sometimes have issues with sustained power delivery, simply because they are cheap, and concessions had to be made. With music though, the Fosi at least should do quite well. The B100, far less so. For sure, this should be highlighted more in reviews.

The concept of "derating" however is way too reductive and eventually doesn't tell you a whole lot. Better just to provide all the data, then we can see what is going on.

 

[SSS]

What is your take on the SA video? I think he saying the Fosi V3 gets to much higher watt output when fed a 1khz tone vs pink noise.

I think Erin is saying something similar - what is your take on Erins discussion of 1khz tone vs multitone test signal?

I think currently Amir only uses multitone to test THD. Can one determine multitone power output from that?
If not, can he test output power with multi tone and with pink noise?

If amir can do it, this will be great for comparison of power amplifiers. Basic question is what definition of power measurment is good to be a somekind of standard. I think that the move to class-D amps with switchemode power supplies changed the behaviour compared to the old ones. Even in the old days some amps could deliver full power only for a certain time span due to thermal issues. Question stays what is the measurement method to be a valid one regarding music listening? Almost nobody listens to single pure sine waves.

 

[xrqp]

The FTC test still only sends a single tone at a time (and it is swept to cover the range of 20 to 20k). But what if the signal was pink noise, or 32 multitones all at one time?

Voodoo, yes, I meant 1khz. And the graph is good because I think all 32 tones are sent at once, together. But it is only for 5 watt. If they did that test at the max watts of the amp (at 1% THD watts), then what would be the watts?
Amir also graphs power vs THD for various frequencies done one at a time. What if he did it for pink noise and multitone - just add two more lines?

SSS, I feel like you are the only one who gets what I am saying. When an amp ( a cheap class D amp in this case) that can do 100 watt with a 1khz input, can also only do 0.5 watts with a pink noise input, does it not raise ones interest in testing power with pink or multitone?

 

[voodooless]

Voodoo, yes, I meant 1khz. And the graph is good because I think all 32 tones are sent at once, together. But it is only for 5 watt. If they did that test at the max watts of the amp (at 1% THD watts) and with pink noise, then what would be the watts?

That depends on the duration of the test. With the tests that @amirm does, probably it would still show good power. Lengthen the test, and numbers will most likely fall for that particular amp.

How does an amp that can do 100 watt with a 1khz input also do only 0.5 watts with a pink noise input?

It most likely doesn't. That test is unclear, unrepeatable, and unscientific.

 

[SSS]

Viewing the German DIN 45500 for measuring power amplifiers from the late 1960ies it says that the stereo amp should work for 10 minutes delivering 2 x 6 Watt. Of course this is low power compared to nowadays. But there is the time span defined. What time span would be appropriate for today in relation to music listening, anybody has an idea ?

 

[SIY]

I think currently Amir only uses multitone to test THD.

No.

I didn't watch the video (I deal with enough idiocy in my day job), but superposition and crest factor are the things you need to understand when doing amplifier measurements.

 

[xrqp]

That test is unclear, unrepeatable, and unscientific.

I think it is repeatable etc. But you have to listen to the whole video, and possibly more than one. He uses a CD or DVD to send the test tones at 2V and turns up the amps, and measures the db from the speakers. On some test he uses a 4 ohm resistor instead of speakers. Can you support your statement?

 

[xrqp]

I deal with enough idiocy

Are you referring to his presentation style, or the real findings he presents?

I don't know anything about crest factor and superposition. But superposition sounds like it might be sending a signal with many frequencies at once superimposed together - probably not that. I don't recall Amirs reports mentioning those.

When I said Amir only uses multitone to test THD, I meant that he does not use multitone to test max power at 1% THD.

 

[SIY]

Are you referring to his presentation style, or the real findings he presents?

You can be both truthful and dishonest with enough lack of understanding, whether yours, the audience's, or both.

 

[voodooless]

He uses a CD or DVD to send the test tones at 2V and turns up the amps, and measures the db from the speakers

You measure gain, not maximum output... it's a pointless exercise. And a really convoluted way of measuring it as well.

 

[SIY]

I don't know anything about crest factor and superposition.

Start there if you want to understand measurement.

 

[xrqp]

He measures the power coming out of the speaker with mic, or a sound meter, and uses the speaker sensitivity to calc the power going into the speaker. I guess we are talking past each other so I will try not to get argumentative going forward.

 

[xrqp]

SIY, do you think that some of these cheap class D amps are not much affected by the test tone of 1khz vs pink noise?

 

[SIY]

SIY, do you think that some of these cheap class D amps are not much affected by the test tone of 1khz vs pink noise?

With the caveat that I have only tested two or three of the cheapies, if the measurement is done correctly, there's not a significant effect. "Measurement done correctly" is the key. If you don't understand the basics of signals, your measurement will not mean much.

 

[xrqp]

Did you test power with a multitone input and compare it to 1khz input?

 

[SIY]

Did you test power with a multitone input and compare it to 1khz input?

Just to see if you spent a bit of effort learning the basics about these measurements, if a 1kHz sine wave shows clipping at 2.83V into 8R, what's the power?

Now, if you have a two tone IMD test, what should the voltage level of the tones be at clipping? What's the power?

Now if you have a 32 tone multitone, what should the voltage level of the tones be at clipping? What's the power?

 

[NTK]

Did you test power with a multitone input and compare it to 1khz input?

You cannot compare power measurements using signals with (very) different crest factors. That's the basic problem with the interpretations of Erin's and SA's multitone/pink noise tests.

The max/peak voltages required to reach the same average power are different (as reflected in the crest factors and why crest factor is an important consideration in power measurements).

 

[SSS]

You cannot compare power measurements using signals with (very) different crest factors. That's the basic problem with the interpretations of Erin's and SA's multitone/pink noise tests.

The max/peak voltages required to reach the same average power are different (as reflected in the crest factors and why crest factor is an important consideration in power measurements).

Absolutely right. With multitone clipping can occur due to the fact that the higher tones ride on the lower ones. So at clipping point the lower tone peak can not reach the clipping limit the same time if ever. How this reflects in the average wattage hope somebody knows the exact math for this. On the other hand it makes no sense to me having for instance a 10 kHz tone with full wattage since the tweeter of a loudspeaker will be destroyed. And most music has a spectrum with declining sound level with rising frequency. Therefore I doubt that a multitone measurement of power amplifiers at full throttle makes sense. At lower levels of course valid to see the IMD.

 

[Old_School_Brad]

You cannot compare power measurements using signals with (very) different crest factors. That's the basic problem with the interpretations of Erin's and SA's multitone/pink noise tests.

The max/peak voltages required to reach the same average power are different (as reflected in the crest factors and why crest factor is an important consideration in power measurements).

Doesn't a multitone signal actually mimic the crest factor of dynamic music? It seems to me that this provides an advantage for power measurements. However, the downside is that its evenly distributed signals don't accurately represent real music. What test signal would be a better representation? Pink noise has an even lower crest factor, and sine waves have an even lower one, moving even further away from "realism."
I don't think that a single signal test would be adequate, and comparing them one-to-one is likely to cause confusion.