'Headroom' is a measure of the badness of an amplifier. The bigger the number, the worse the amplifier.

[Zaphod]

Edit: New thread from posts excised from the Fosi amp thread.

Fosi Audio ZA3 Amplifier Review

This is a review and detailed measurements of the Fosi Audio ZA3 stereo amplifier with balanced XLR inputs. It was sent to me by the company and costs US $149.99 with 48 volt power supply. The amplifier has really dressed up from its earlier origins of common enclosures in budget class...

www.audiosciencereview.com

To cite you: "Nice bit of headroom."

You make that sound like a good thing. A large headroom figure is merely measure of how bad the power supply is.

Oh and the frequency response is shockingly bad. Still, it is a cheap, attractive amp. It won't be going in my system. My amplifier has no measurable headroom. Maximum power is 220 Watts continuous. Peak power is 220 Watts. Which is how it should be. Of course, a 5kVA power transformer and 310,000uF of filter capacitance makes for a very stiff power supply.

 

[Bruce Morgen]

You make that sound like a good thing. A large headroom figure is merely measure of how bad the power supply is.

On what planet is that even remotely true? The peak power demands of most music are relatively brief, and "headroom" simply allows an amplifier of relatively small size, continuous power rating, and thermal dissipation capability to handle such momentary peaks with aplomb and without an audible increase in distortion. That's "a good thing" for those of us who don't own stock in the local electric utility paying us dividends to offset the operating cost of a power supply sourcing literally thousands of watts in order for "headroom" not to be an issue. In my world that's extreme overkill for the use case of a home audio system, but of course YMMV -- and apparently does.

 

[Zaphod]

On what planet is that even remotely true? The peak power demands of most music are relatively brief, and "headroom" simply allows an amplifier of relatively small size, continuous power rating, and thermal dissipation capability to handle such momentary peaks with aplomb and without an audible increase in distortion. That's "a good thing" for those of us who don't own stock in the local electric utility paying us dividends to offset the operating cost of a power supply sourcing literally thousands of watts in order for "headroom" not to be an issue. In my world that's extreme overkill for the use case of a home audio system, but of course YMMV -- and apparently does.

On this planet.

OK, I'll bite. Tell me how long the "momentary peaks" are that you speak of? Give me a number.

You can't because every piece of music is different. Some music will exhibit a peak to average ratio of as little as 10dB and other music might be as high as 40dB. Then, of course, is the duration of the those peaks. I don't recall the timing, but the low organ note in my copy of Also Sprach Zarathustra (Zebin Mehta version) lasts around 10 seconds. Maybe. And. before you ask: Yes, my speakers will reproduce those notes. Yes, the windows rattle. A bit.

Yeah, I get that it is a cheap 'n cheerful amp. That's fine. I was simply challenging the commonly assumed notion that big headroom figures are good. They're not. They are an indicator of how BAD an amplifier is.

Here's the problem with the 'headroom' nonsense:

I used to sell an NAD amplifier (the NAD 2200) which possessed ENORMOUS headroom. Around 6dB. That was great, until I began receiving speakers from owners that were VERY badly damaged. I popped the amp on my bench and began measuring. Sure enough: The 100 Watts rated NAD could deliver around 400 Watts for short periods. However, when the amp had exhausted it's headroom, the clipping became utterly horrendous. It soon became obvious what was happening. I ceased selling that model and any other that used such a silly system.

Think of it this way: If we take a high end amp from Krell, Mark Levinson, Boulder, D'Agostino, or one of several other manufacturers, you will quickly find that their continuous power rating is IDENTICAL (or very, VERY close) to their peak power rating. It's all a function of power supply stiffness. Crappy power supplies in an amplifier allow for large headroom figures. Some manufacturers use a regulated power supply for the output stages. They will have headroom figures that are tiny.

 

[Deleted member 48726]

On this planet.

OK, I'll bite. Tell me how long the "momentary peaks" are that you speak of? Give me a number.

You can't because every piece of music is different. Some music will exhibit a peak to average ratio of as little as 10dB and other music might be as high as 40dB. Then, of course, is the duration of the those peaks. I don't recall the timing, but the low organ note in my copy of Also Sprach Zarathustra (Zebin Mehta version) lasts around 10 seconds. Maybe. And. before you ask: Yes, my speakers will reproduce those notes.

Yeah, I get that it is a cheap 'n cheerful amp. That's fine. I was simply challenging the commonly assumed notion that big headroom figures are good. They're not. They are an indicator of how BAD an amplifier is.

Here's the problem with the 'headroom' nonsense:

I used to sell an NAD amplifier (the NAD 2200) which possessed ENORMOUS headroom. Around 6dB. That was great, until I began receiving speakers from owners that were VERY badly damaged. I popped the amp on my bench and began measuring. Sure enough: The 100 Watts rated NAD could deliver around 400 Watts for short periods. However, when the amp had exhausted it's headroom, the clipping became utterly horrendous. It soon became obvious what was happening. I ceased selling that model and any other that used such a silly system.

Think of it this way: If we take a high end amp from Krell, Mark Levinson, Boulder, D'Agostino, or one of several other manufacturers, you will quickly find that their continuous power rating is IDENTICAL (or very, VERY close) to their peak power rating. It's all a function of power supply stiffness. Crappy power supplies in an amplifier allow for large headroom figures. Some manufacturers use a regulated power supply for the output stages. They will have headroom figures that are tiny.

In your examples with NAD the customers surely should have bought more powerful amplifiers to begin with. It's not the manufacturers fault that people mismatch their components.

The design with non-regulated power supplies with large capacitors are among the most widespread and provide great results from the compromise between much larger, heavier and more expensive designs as music is dynamic. Yamaha has successfully used this design in slightly different iterations for many decades. Proton also comes to mind. Older Onkyo as well. All well regarded.

 

[Zaphod]

In your examples with NAD the customers surely should have bought more powerful amplifiers to begin with. It's not the manufacturers fault that people mismatch their components.

The design with non-regulated power supplies with large capacitors are among the most widespread and provide great results from the compromise between much larger, heavier and more expensive designs as music is dynamic. Yamaha has successfully used this design in slightly different iterations for many decades. Proton also comes to mind. Older Onkyo as well. All well regarded.

I agree that the NAD buyers should have spent a great deal more money on 400 Watt continuously rated amplifiers. However, it is highly likely that their budgets did not stretch that far.

My fundamental point remains: Large headroom figures simply tell us how bad an amplifier is. It is not a measure of goodness.

BTW: Of the manufacturers you mentioned, I would point out that Proton (the company was called Fulet back then) was contracted by NAD to build amplifiers. When NAD and Fulet ceased their relationship, the company released a range of products that were electronically identical to the equivalent NAD ones. So identical that the service manual for the NAD could be used with the Proton products. Which of course means that the early Proton amps had exactly the same problems as the NAD (model 2200) that I outlined earlier.

 

[Zaphod]

It was good that you pointed that out. You can test how high up in frequency you can hear and then take it from that, how to approach this amplifier's FR that is.
I think I can hear up to 15 kHz, but it's been a few years since I tested. Probably 14kHz now.

Okay,
+/-0.5dB at 15kHz and +1dB at 20kHz is at 8 Ohms, but if the speaker presents higher Ohms, what happens? Even more deviation from FR? I do not know.

The opposite is more likely the case. Higher impedance loads = better FR. Lower impedance loads = worse FR. Pretty normal with cheap Class D amps, due the type of output filter used.
[EDIT] Oops. I just looked more closely at the FR graphs. The 8 Ohm one is terrible. The amp is a fail. Clearly, the output filtering has been very poorly implemented. Even a very inexpensive Class A/B amp will exhibit a far flatter FR than this thing.

 

[Bruce Morgen]

OK, I'll bite. Tell me how long the "momentary peaks" are that you speak of? Give me a number.

It's very rare for a demanding musical peak to last more than a single-digit number of seconds, with the vast majority of such events much shorter than that.

 

[amirm]

[EDIT] Oops. I just looked more closely at the FR graphs. The 8 Ohm one is terrible. The amp is a fail. Clearly, the output filtering has been very poorly implemented. Even a very inexpensive Class A/B amp will exhibit a far flatter FR than this thing.

There is nothing terrible about it. Class D amplifiers create out of band noise (switching). That needs to be filters so a high power one is included much like you have in a speaker crossover. The issue is that the filter interacts with the load (speaker) and its response changes. For a long time, this was the case with every class D amplifier. Then we had likes of Hypex and brought the filter into the amplifier feedback loop, thereby negating its load dependency. This has to be carefully done as to avoid oscillations. The IC uses here for amplification has that capability so it is a matter of implementation and any extra cost.

 

[Zaphod]

It's very rare for a demanding musical peak to last more than a single-digit number of seconds, with the vast majority of such events much shorter than that.

The CEA-2006/490A standards call for a 480mSec 'rest' period between 20mSec bursts.

IOW: Not even remotely useful in the real world. Like I said: My Also Sprach Zarathustra recording has several seconds of low organ bass at very high level.

I will say again: Headroom measurements tell use how BAD and amplifier (mostly power supply performance) is. It is not a measure of goodness.

It is just a feeble attempt by crappy amplifier manufacturers to drag us back to the bad old days of the 1970s, when 'peak power' was all the rage in budget equipment and ghetto blasters.

 

[Zaphod]

There is nothing terrible about it. Class D amplifiers create out of band noise (switching). That needs to be filters so a high power one is included much like you have in a speaker crossover. The issue is that the filter interacts with the load (speaker) and its response changes. For a long time, this was the case with every class D amplifier. Then we had likes of Hypex and brought the filter into the amplifier feedback loop, thereby negating its load dependency. This has to be carefully done as to avoid oscillations. The IC uses here for amplification has that capability so it is a matter of implementation and any extra cost.

Correct. The Hypex products are very well implemented, as can be seen by their load tolerance. The reviewed amplifier is OK for the money, but is can hardly be regarded as a high fidelity product. The phase shift issues will be noted well into the audible region.

Look, it's a decent little amp, delivering lots of power, for not much money. That's as far as it goes, IMO.

And, since I have your attention, I will say once more: Lots of headroom indicates a poor quality amplifier (notably power supply). It is not an indicator of quality.

 

[amirm]

And, since I have your attention, I will say once more: Lots of headroom indicates a poor quality amplifier (notably power supply). It is not an indicator of quality.

It is. Take two amps. One has 100 watts average power and burst power of 100 watts. The second also has average power of 100 watts but able to have peak power of 150 watts. Are you saying the second amplifier is worse than the first?

 

[Zaphod]

It is. Take two amps. One has 100 watts average power and burst power of 100 watts. The second also has average power of 100 watts but able to have peak power of 150 watts. Are you saying the second amplifier is worse than the first?

All things being equal (which they usually not), yes. The second amplifier has a poor quality power supply (probably). The effect of the poor quality power supply will be exacerbated by low(er) impedance loads.

Let me put it another way:

Which amp will sound better (Assuming identical topology)?
1) A 100 Watt amplifier that employs a (say) 1kVA power transformer and 100,000uF of filter capacitance?

Or:

2) A 100 Watt amplifier that employs a 400 VA power transformer and 10,000uF of filter capacitance?

Which amplifier exhibits biggest headroom figure?

Of course, it is amplifier #2. It's headroom figure will be higher than amplifier #1. Yet it is clearly a poorer quality amplifier.

Therefore: High(er) amplifier headroom figures are an indicator of how bad an amplifier is.

Trust me on this: I have performed the above experiment under single blind conditions on myself. The results are not subtle when listening under demanding conditions. That said, my experiments were confined to Class A/B amplifiers.

 

[Deleted member 48726]

All things being equal (which they usually not), yes. The second amplifier has a poor quality power supply (probably). The effect of the poor quality power supply will be exacerbated by low(er) impedance loads.

Let me put it another way:

Which amp will sound better (Assuming identical topology)?
1) A 100 Watt amplifier that employs a (say) 1kVA power transformer and 100,000uF of filter capacitance?

Or:

2) A 100 Watt amplifier that employs a 400 VA power transformer and 10,000uF of filter capacitance?

Which amplifier exhibits biggest headroom figure?

Of course, it is amplifier #2. It's headroom figure will be higher than amplifier #1. Yet it is clearly a poorer quality amplifier.

Therefore: High(er) amplifier headroom figures are an indicator of how bad an amplifier is.

Trust me on this: I have performed the above experiment under single blind conditions on myself. The results are not subtle when listening under demanding conditions. That said, my experiments were confined to Class A/B amplifiers.

Can you take this repetition posts of yours and start a thread about dynamic power or "bad" power supplies?
Then you can repeat your laughable claim as much as you like and I'll happily participate in shooting them down there..

 

[IAtaman]

All things being equal (which they usually not), yes. The second amplifier has a poor quality power supply (probably). The effect of the poor quality power supply will be exacerbated by low(er) impedance loads.

Let me put it another way:

Which amp will sound better (Assuming identical topology)?
1) A 100 Watt amplifier that employs a (say) 1kVA power transformer and 100,000uF of filter capacitance?

Or:

2) A 100 Watt amplifier that employs a 400 VA power transformer and 10,000uF of filter capacitance?

Which amplifier exhibits biggest headroom figure?

Of course, it is amplifier #2. It's headroom figure will be higher than amplifier #1. Yet it is clearly a poorer quality amplifier.

Therefore: High(er) amplifier headroom figures are an indicator of how bad an amplifier is.

Trust me on this: I have performed the above experiment under single blind conditions on myself. The results are not subtle when listening under demanding conditions. That said, my experiments were confined to Class A/B amplifiers.

I don't get it. All other parameters being equal, including rail voltage, why would an amplifier that has a bigger power supply would have less headroom and sound worse?

 

[Zaphod]

I don't get it. All other parameters being equal, including rail voltage, why would an amplifier that has a bigger power supply would have less headroom and sound worse?

It won't. Perhaps my words were a little unclear. #2 can never sound better than #1.

Therefore, all other things being equal, an amplifier with a higher headroom figure will always sound worse than one with a lower headroom figure.

IOW: 'Headroom' is a measure of the badness of an amplifier. The bigger the number, the worse the amplifier.

 

[Deleted member 48726]

It won't. Perhaps my words were a little unclear. #2 can never sound better than #1.

Therefore, all other things being equal, an amplifier with a higher headroom figure will always sound worse than one with a lower headroom figure.

IOW: 'Headroom' is a measure of the badness of an amplifier. The bigger the number, the worse the amplifier.

Utter nonsense.

 

[IAtaman]

It won't. Perhaps my words were a little unclear. #2 can never sound better than #1.

Therefore, all other things being equal, an amplifier with a higher headroom figure will always sound worse than one with a lower headroom figure.

IOW: 'Headroom' is a measure of the badness of an amplifier. The bigger the number, the worse the amplifier.

I am sorry, I do not follow.

1. A 100 Watt amplifier that employs a (say) 1kVA power transformer and 100,000uF of filter capacitance
2. A 100 Watt amplifier that employs a 400 VA power transformer and 10,000uF of filter capacitance

In the example you have given, all else being equal, which one has more "headroom" and why would it sound worse?

Maybe a better starting point : what do you mean by headroom?

 

[amirm]

1) A 100 Watt amplifier that employs a (say) 1kVA power transformer and 100,000uF of filter capacitance?

Or:

2) A 100 Watt amplifier that employs a 400 VA power transformer and 10,000uF of filter capacitance?

Which amplifier exhibits biggest headroom figure?

Of course, it is amplifier #2. It's headroom figure will be higher than amplifier #1. Yet it is clearly a poorer quality amplifier.

That's nonsense. No one designs a 100 watt amp and then put a 1 KVA power supply in it. Ignoring that, you have no idea of the level of headroom an amplifier has in those circumstances. Most likely amp #1 would have the same headroom as limited by the amplifier design as #2. And since amp 1 would cost a lot more, that is the bad option.

The right analogy is the same power supply with one having 10,000 uf filter cap and the other, 50,000. The latter would provide more headroom with some incremental cost that is cheaper than having same headroom but with more average power.

 

[Bob101]

All things being equal (which they usually not), yes. The second amplifier has a poor quality power supply (probably). The effect of the poor quality power supply will be exacerbated by low(er) impedance loads.

Let me put it another way:

Which amp will sound better (Assuming identical topology)?
1) A 100 Watt amplifier that employs a (say) 1kVA power transformer and 100,000uF of filter capacitance?

Or:

2) A 100 Watt amplifier that employs a 400 VA power transformer and 10,000uF of filter capacitance?

Which amplifier exhibits biggest headroom figure?

Of course, it is amplifier #2. It's headroom figure will be higher than amplifier #1. Yet it is clearly a poorer quality amplifier.

Therefore: High(er) amplifier headroom figures are an indicator of how bad an amplifier is.

Trust me on this: I have performed the above experiment under single blind conditions on myself. The results are not subtle when listening under demanding conditions. That said, my experiments were confined to Class A/B amplifiers.

I'm sure in your head all that is true. Bless you and enjoy the music.

 

[Zaphod]

Utter nonsense.

Damn. You really nailed me to wall with your carefully reasoned technical response.

NOT!