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

[Lambda]

Then let us perform a mind experiment.

You can not even be bothered to define what exactly would be "better"
And how this "better" would measure.

You seam to think more power is better?

Then how and why is it better to have less "Headroom"?
Sure an amp with more continues Power can be called "better" but this is not the same as saying Headroom is bad.

You seam to say "just make the power supply bigger" but why?
This costs money and resources that can be spend on other components.

Engineering is an compromise:
'Anyone can build a bridge that stands, but it takes an engineer to build a bridge that barely stands.'

 

[Zaphod]

You can not even be bothered to define what exactly would be "better"
And how this "better" would measure.

You seam to think more power is better?

Then how and why is it better to have less "Headroom"?
Sure an amp with more continues Power can be called "better" but this is not the same as saying Headroom is bad.

You seam to say "just make the power supply bigger" but why?
This costs money and resources that can be spend on other components.

Engineering is an compromise:
'Anyone can build a bridge that stands, but it takes an engineer to build a bridge that barely stands.'

No. I accept that a STIFFER power supply is always better than a 'soft' power supply. All other things being equal.

I never mentioned cost. Not once. I accept that various compromises may be made in order to reduce costs.

 

[Short38]

Why in hell is this thread active?

 

[ahofer]

So far I'm mostly interested in the single blind test.

I've never had the urge to put Also Sprach on at volume.

 

[Keith_W]

Zaphod? Is that you, Trevor? From Queensland? Services ME amplifiers? Guys, if this is who I think is posting, he is well known in Australian hi-fi circles. He has a proven track record as an objectivist going back 30 years to the days of USENet.

I am prepared to entertain your thoughts on this, but please post measurements to prove your case ... otherwise nobody will believe you.

 

[Chrispy]

Or is just the audio gear isolation that Australia suffers from?

 

[kemmler3D]

I just wanted to mention that 20ms with ~500ms in between is a perfectly good standard for peaks in the sense of transients, with IMO good correspondence to actual music. 20ms peaks with half a second in between would cover a bass drum going well above average while playing a beat at 120bpm.

Sustained notes, then would correspond to... sustained power, I guess.

 

[Deleted member 48726]

Okay. @Zaphod -Why are these bad amplifiers?

Rated to 1 ohm load dynamically and have plenty dynamic power. -Like almost all Yamaha amplifiers of older designs.
They are large soft power supplies with large capacitors. Optimized for cost.
This way of cost optimization is luckily also benefitting both power consumption and actual real world use cases of dynamic(!) music.

Now you'd probably argue that a stiff PS amp. would be better. No. It wouldn't. You'd just convert those dynamic values to continuous.
And assuming you'd only need those power figures in musical peaks the Yamahas' performance would be equal to the imaginary 2.5 kW monster amp.

Unless you got measurements to share that show compression or audible distortion when amplifiers with soft PS operate within their "headroom envelope" I'm sticking to my opinion.

 

[Punter]

"Headroom" is a somewhat domestic expression when applied to amplifiers. As I understand it, it's not actually a term used in relation to the amplifier circuitry but more specifically to the power supply. For any given load situation, within the design parameters of the amplifier, the power supply should be able to maintain a stable output voltage. Any failure of the power supply to "keep up" with the required load would see the voltage of the supply "sagging". This phenomenon would present itself when the energy stored in the smoothing/reservoir capacitors is depleted. This could happen due to a big transient or sustained volume that exceeds the power supplies ability to source the required current.

The sonic effect of this situation would be "clipping" where the peaks of the music signal would be squared off due to insufficient power supply voltage/current. The critical component involved in preventing clipping is of course the capacitors as they act like a battery which is being charged by the rectifier. Capacitors can supply huge amounts of current due to their relatively low internal resistance and of course the fact that electrical energy is stored in a capacitor as actual electric potential rather than a chemical reaction.
The original assertion that "the more headroom an amplifier has, the worse the amplifier" is a complete non-sequitur as headroom is something built into the power supply and has little to do with the actual amplifier circuit. It's telling that amplifiers designed for real work, in live sound systems, generally have massive transformers and massive capacitors, "headroom" is what any PA engineer wants in spades so as to ensure the cleanest sound possible. It's also a desirable quality in a domestic HiFi amplifier for the same reasons.

 

[solderdude]

The stiffer the power supply, the lower the headroom figures.

Depends on the current delivery capacity, regulation, amplifier circuit (internal resistance) as well as over-current protection.
That may be logical to you but generalizing such a statement will bring resistance to such generalizations.
I am quite sure you understand amp topology, power supplies, loads etc.
A high resistance in the source is not good, in the load it is far better.
I think you are often misunderstood because of the short, or lacking in depth, explanations and generalization of this topic.


One can build a very stiff power supply (regulated) and as long as the max. current capability is not exceeded, the amp circuit has a very low internal resistance near max. voltage output swing there will be small headroom numbers.
One can also massively over dimension a non regulated power supply and connect it to the same amplifier and get the same result.
Design choices often made for financial reasons (profit and/or budget range of the intended consumer group).

In this case the amplifier is what determines the sound quality (near clipping) and not so much the power supply.
The moment some form of current limiting kicks in (regulated power supply, non regulated sagging or amp limiting) you will see increased headroom numbers and non doubling output power at halving load impedance.
At that point one has reached the limits of the amplifier (as a whole).
From that point on sound quality (how the device handles that, i.e. what the waveform looks like) will differ substantially depending on design choices.

When this point is reached regularly one should simply buy a more powerful amplifier. It is that simple.
When you clip your amps often buy something else.
When you never reach that point there is no need for that and the stiffness of the power supply and max continuous power is all irrelevant.

 

[HarmonicTHD]

Then let us perform a mind experiment.

The ZA3 is supplied with a 48 Volt 5 Amp power supply. The headroom figures are as in the review.

Let's assume that the ZA3 is supplied with a 48 Volt, 10 Amp power supply.

Both power supplies are SMPS and fully regulated.

What is likely to happen to the continuous power output?

What is likely to happen to the headroom figure?

Why don’t you come back to the original question eg the question posed by @amirm here:

Post in thread ''Headroom' is a measure of the badness of an amplifier. The bigger the number, the worse the amplifier.'
https://audiosciencereview.com/foru...er-the-worse-the-amplifier.52176/post-1881288

Meaning which amp is better? One with 100w avg and 100w peak vs one with 100w avg and 150w peak?

… and leave aside for a moment how the PS is constructed.

 

[EJ3]

On this planet.

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

I'll answer this first with some measurements (of the "momentary peaks" from one of the NAD 2200's that you think that you found a clipping problem with):
Dynamic power measurements were somewhat complicated by the thermal feedback system of the amplifier. As it heated up the maximum power decreased, so we were never able to repeat our measurements exactly on successive attempts. Nonetheless, the results were very impressive. With the standard 20-millisecond test signal, the maximum output was 450 watts into 8 ohms (for a dynamic headroom of 6.5 dB), 685 watts into 4 ohms, and 870 watts into 2 ohms. We repeated these measurements with 4-ohm loads and with longer-duration tone bursts.

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.

And now to answer why these amps were not clipping (unless there was something else actually wrong with them):
The output of 700 watts did not change for bursts in the 20- to 50-ms range, but it decreased slightly to 570 watts with 100- and 200-ms bursts and to 285 watts with bursts of 300 ms and longer. (The output did not drop significantly with a 1-second burst, and, in fact, it remained at 240 watts in a “continuous” measurement lasting perhaps 30 seconds.) We repeated this test (in part) using the bridged mode, reading outputs of 1,100 watts for 20 milliseconds, 800 watts for 50 milliseconds, and 700 watts for 100 milliseconds.

Based on these tests on a properly working NAD 2200, there is no clipping after the "momentary peak" runs out of time, just a lowering of the output to a sustainable level.

Ahh, yes: current capability:
The output transistors of the NAD 2200 amplifier are high-powered, fast-switching devices capable of delivering some 60 amperes of peak current for brief periods.

 

[KSTR]

What a weird thread...

ALL amplifiers, starting from early class-A/B tube amps, that don't have a fully regulated supply for the output stage a) always have supply sag under heavy load and thus increased, larger than nominal, peak output power for short periods, and b) have a nominal power that scales with mains voltage.

This leads to naturally inconsistent and scattering power figures, please note that power scales with voltage squared that's why it easy to have large differences. 20% sag gives 1.5x more peak vs nominal power. Also keep in mind that class-A/AB transistor outputs usually have saturation voltages (the output cannot go all the way up/down to the rails) which makes the situation typically more severe for smaller output voltage (lower power amps). Mains tolerance of +-10% again give at least a 1.5x difference in nominal power.

Fully regulated supply rails are feasible with switched-mode power supplies, either ones directly fed from the mains as well as others regulating the output voltage of a standard transformer-based supply. But they are seldom used for good reasons. They make the product much more complex and expensive by offering no apparent benefit, except a fixed and guaranteed power rating for the first type even with low supply (say, 120Vac US product used at 90Vac or 100Vac in Japan).

The OP's argumentation is totally moot. He just moans about the natural behavior of almost all amps.

As for the sound, it is possible that an amp's sound degrades measurably when the supply rails are bouncing up and down and that indicates bad design (unstable operating points, only little "power supply rejection" as it is called, technically).

 

[Sokel]

An empty (discharged) capacitor looks very much like a short circuit at the instant of power on. That's the inrush current.

Many, many (dare I say most) SMPS regulators have built in soft start features to prevent the regulator from having issues at start up. The designers of the regulators take into account inrush.

In my previous post I didn't mean to imply that there is not some limit to the amount of bulk capacitance that can be hung on the output of a SMPS. I'm certain there likely is. What I did say, is that SMPS just don't need the tens of thousands of uF to prevent voltage sag.

We're literally talking about time periods changes on the order of 1000 times to 10,000 times smaller when comparing linear and SMPS. In a very simplistic view, the output capacitance can shrink by those same ratios and still provide the same voltage sag performance.

What's with the new options of adding capacitance in the fashion MicroAudio offers or described in the below post (No7 option) ?
I think I saw that again couple of times if one wants to drive full range and down low.

Buckeye Amps 2024 News and Upcoming Products

Happy New Year to everyone! With a new year comes new products, revisions, and overall updates designed to keep Buckeye Amps going in the right direction for the audio community. 1) One of the biggest updates you can check out right now is our brand new website! It is still the same domain...

www.audiosciencereview.com

 

[Zaphod]

Zaphod? Is that you, Trevor? From Queensland? Services ME amplifiers? Guys, if this is who I think is posting, he is well known in Australian hi-fi circles. He has a proven track record as an objectivist going back 30 years to the days of USENet.

I am prepared to entertain your thoughts on this, but please post measurements to prove your case ... otherwise nobody will believe you.

Yep and I'm in NSW. Measurements? When I get time. And yes, I agree that measurements are vital. A thought experiment should be sufficient though. I'll fire up the QA403 sometime and see about those measurements.

 

[Zaphod]

Okay. @Zaphod -Why are these bad amplifiers?

Rated to 1 ohm load dynamically and have plenty dynamic power. -Like almost all Yamaha amplifiers of older designs.
They are large soft power supplies with large capacitors. Optimized for cost.
This way of cost optimization is luckily also benefitting both power consumption and actual real world use cases of dynamic(!) music.

Now you'd probably argue that a stiff PS amp. would be better. No. It wouldn't. You'd just convert those dynamic values to continuous.
And assuming you'd only need those power figures in musical peaks the Yamahas' performance would be equal to the imaginary 2.5 kW monster amp.

Unless you got measurements to share that show compression or audible distortion when amplifiers with soft PS operate within their "headroom envelope" I'm sticking to my opinion.

View attachment 349339
View attachment 349337
View attachment 349338

How long can the amps maintain their 'dynamic power'?

Long enough to play the organ notes on my Strauss recording?

What happens if I turn the volume up to just peak the music on one of these amps and it runs out of 'dynamic power' to maintain that peak?

 

[Lambda]

I accept that a STIFFER power supply is always better than a 'soft' power supply.

You still have not defined "better"
More energy efficient. more cost efficient, more space efficient, more weight efficient...

Whats the point if you can't even define how its better

 

[antcollinet]

I am arguing from a position of logic.

A stiff power supply in an otherwise identical amplifier to one with a 'soft' power supply will always measure and probably sound better.

How does a stiff power supply universally correlate with headroom?

How about headroom as a result of short to medium term thermal performance?

Are you simply saying an amp rated at 100W but with 50W of headroom is worse than a 150W amp? Then no argument.

Or are you saying when comparing 2 150W continuous amps where only one has headroom up to 300W, that the one with additional headroom is worse?


You need to:

1 - Define what you mean by headroom.
2 - Define your definition of "better" as has been pointed out above.
3 - As part of 2, define "than what": What are you comparing with?
4 - Describe why you seem to think headroom is only related to power supply "stiffness" - a term you've also not defined, but your posts suggest you think it is one or both of output impedance or current rating.


@Keith_W above seems to be suggesting you might be technically competent. Unfortunately that competence doesn't seem to extend to making your case in a reasoned written argument. Based on your posts here, any competence is (to this reader at least) masked.

 

[antcollinet]

That was not my intention.

You were very successful at it though.

 

[Ordin Aryguy]

What's with the new options of adding capacitance in the fashion MicroAudio offers or described in the below post (No7 option) ?
I think I saw that again couple of times if one wants to drive full range and down low.

Buckeye Amps 2024 News and Upcoming Products

Happy New Year to everyone! With a new year comes new products, revisions, and overall updates designed to keep Buckeye Amps going in the right direction for the audio community. 1) One of the biggest updates you can check out right now is our brand new website! It is still the same domain...

www.audiosciencereview.com

Where the additional capacitance is placed depends on what problem the additional capacitance is supposed to address. Is the voltage at the output terminals of the supply drooping, or is it the voltage at the input to the amplifier?

It's my suspicion that lots of people are just adding additional capacitance without understanding the "why's and where's" of what they are actually doing. While it might seem like adding the capacitance anywhere between the power supply and the load (the amplifier in this case) is the same, it's really not. What most people neglect to consider is the resistance/impedance of the wiring itself.

Real world application? If I think there's voltage droop during transients causing some audible effects, the very first place I'd put additional capacitance would be right on the input terminals of the amplifier. That addresses the resistance/impedance drop of the wiring, which is more likely to create problems as long as the power supply itself is sufficiently sized to begin with.

As for the "down low" comment, the lower frequencies have longer time periods. In this case "down low" would mean longer periods of higher current draw. There are also capacitors internal to the amplifier itself that are meant to maintain a constant voltage. Did the engineer put in enough capacitance to ensure that the voltage to the supply will never drop when driving low frequencies at the maximum output of the amplifier? Who really knows, only bench testing with an oscilloscope will tell the real story.

In manufacturing capacitors, particularly larger electrolytic capacitors, are expensive and every engineer is burdened with the COG's (Cost of Goods) challenge.