Controversy regarding Bridged Amplifier power output

[audioguru]

I used the term "RMS Watts" because it is real power and is half the "peak" or "maximum" power number usually stated.
The Peak Power number is exactly double the real power number.

Many amplifier and speaker manufacturers lie and state "music power" which is just momentary power but they usually don't say the duration of the very short test pulse and you don't know the duration of the high power pulses in music.

Usually the power number is stated without saying the frequencies (it should be for all frequencies from 20Hz to 20kHz for an amplifier) and usually doesn't say the amount of distortion which should be as low as when you listen. Many power ratings are made with square-waves at the output of an amplifier because a square-wave has double the power of a sine-wave and the amplifier is much cooler. The distortion is 50% because the added harmonics have the same amount of power as the fundamental sine-wave.

 

[rogs]

I understand your thinking here --it is the usual reason given, to explain the term 'RMS Watts' in the audio context.
It's still not true though --for the reasons I mentioned above, there are no such things as 'RMS Watts'.
Still, not really that important. This is after all a hobby forum, not an academic institution!!

With regard to the thread title, I have recently had the chance to experiment with some class D audio amplifier modules, from Hypex in the Netherlands. Their UCD modules have very low output impedances (in the order of 0.1 Ohms) and can operate with efficiencies of over 90%.
So in a bridge mode --providing the PSU can support the power -- they come pretty close to the 'x4' power output, unlike most A/B amps, where the failure to swing 'rail to rail' at virtually 0 ohms output impedance is not an option!

But when you consider that most audio sources -particularly speech -have a duty cycle power requirement of something like 25% or less, then in many applications this relentless 'quest' for high power outputs is often unnecessary.

If you're looking at running a disco, for hours on end with a major 'sub-bass' componenet though, then of course everything changes.........

 

[audioguru]

A modern class-AB amplifier has an output impedance of from 0.016 ohms to 0.04 ohms. If the power supply has enough power, the amplifiers can pass the increased current and are heatsinked for the heating then two class-AB amplifiers can be bridged for almost 4 times the output power of 1 amplifier.

Very few amplifiers are used to listen to people speaking (the Queen or the Pope?). They are used to play music and today's music is loud.

 

[rogs]

Very few amplifiers are used to listen to people speaking

Maybe not in your world. In my field, emergency public address systems, the amplifiers are used almost exclusively for 'speech only' reproduction.
And in large systems, over specifying amplifier power can be expensive!

Those output impedance figures are very impressive for linear class A/B amplifiers. I've only seen those kind of figures for switching (class D) amps. Do you have any specific models I might refer to? It's always useful to find suitable potential 'rail to rail' candidates, when power supply voltages are a limiting factor!

 

[Nigel Goodwin]

I'm intrigued to read that even after all these years, the term 'RMS Watts' still appears in these audio threads?
'RMS VA' I understand, but as a Watt is a derived unit of power, measured specifically as a rate of energy conversion of 1 joule per second, I fail to see how the term 'RMS' 'music power' - or anything else - can be added.

I tend to think that perhaps this guy is possibly right? Hi Fi Writer - Meaningless RMS power

But maybe I'm missing something?
'RMS Watts' certainly seems to have been used for years and years in audio circles - and continues to be??

It continues to be because everyone understands it - and it's simple to measure (or calculate).

 

[Nigel Goodwin]

Maybe not in your world. In my field, emergency public address systems, the amplifiers are used almost exclusively for 'speech only' reproduction.
And in large systems, over specifying amplifier power can be expensive!

That's why you use 100V line, and add the speaker wattages up

As you say, it's still a fairly common application, if a bit specialist.

Those output impedance figures are very impressive for linear class A/B amplifiers. I've only seen those kind of figures for switching (class D) amps. Do you have any specific models I might refer to? It's always useful to find suitable potential 'rail to rail' candidates, when power supply voltages are a limiting factor!

It's VERY rare for the output impedance values to be specified - if you're lucky they may specify damping factor, the higher the damping factor, the lower the output impedance.

However, any 'modern' transistor amp has extremely low output impedance, but I don't see as it has anything to do 'rail to rail' or low supply voltages?.

 

[rogs]

It's VERY rare for the output impedance values to be specified - if you're lucky they may specify damping factor, the higher the damping factor, the lower the output impedance.

However, any 'modern' transistor amp has extremely low output impedance, but I don't see as it has anything to do 'rail to rail' or low supply voltages?.

As you say, my application is a bit specialised, but when power supply voltages and capacities are limited, by things like battery support requirements, the closer to 'rail to rail' you can swing, the more efficient the amps can be.
The 0.02 Ohm output impedance quoted for this unit for example : https://www.electro-tech-online.com/custompdfs/2011/01/UcD180ST_datasheet.pdf allows for efficient bridging (and minimum heatsinking!).

Problem with linear amps is getting anything like that kind of efficiency, even with low output impedances. Even with the more complex class G '3 rail' configurations you can only hope for something like 72 -75% at best.

Still always nice to know of alternative devices though. Hence my question to the previous poster about specific items.

 

[audioguru]

Crown made professional class-AB amplifiers for many years. They are used in live theaters, stadiums, rock concerts and awards shows.
On their class-AB amplifiers they specify a damping factor that is well over 3000 from 10Hz to 1kHz.
Then if the load is 8 ohms the output impedance is less than 8/3000= 0.0027 ohms.
Here is their article about damping factor:
https://www.electro-tech-online.com/custompdfs/2011/01/damping_factor.pdf

Low output impedance damps the resonances of speakers and keeps the output level constant when the load is changed from no power to full power. A low output impedance is created by plenty of negative feedback that makes corrections in output level and has nothing to do with efficiency.

 

[rogs]

As you say, high damping factors tend to be derived from high levels of negative feedback. So effectively 'dynamic' low impedance. Nothing to do with high effciency.
The link to the spec of the class D amp I made in an earlier post refers to a very low static output impedance, which of course can have significant effects on efficiency.

 

[audioguru]

A class-D amplifier has its output devices switch on and off. They saturate completely then turn off completely with PWM so of course they are efficient and do not waste much power by making heat.

But a linear amplifier circuit has its output devices as variable resistor in series with the power supply and the load so of course they waste power by making heat.

 

[rogs]

But a linear amplifier circuit has its output devices as variable resistor in series with the power supply and the load so of course they waste power by making heat.

So, referrring back to the original thread title then, linear amps are not likely to come near to the 'x4' bridge power output - too many heat losses from the 'variable resistor' part of the linear amp concept.
Class D might come nearer -as I suggested in my previous post?

 

[Mr RB]

"The x 4 myth

It is sometimes stated, usually on internet forums, (and previously here on WikiPedia) that operating an amplifier pair in bridge mode can give four times the power (of one of the pair).

Reference may be made to the fact that power is proportional to the square of the volltage, implying that if the output voltage is doubled – as it is in bridge mode – then the power available increases by a factor of four.

This would only true if the current available could also double – as it would if the output impedance of the amplier stages was zero. But it is not; our 4ohm output impedance amplifier pair now presents as a mono amplifier with an output impedance of 8 ohms and no more current is available from each amplifier than when working singly.

I don't know who wrote that Wiki but he didn't know much about amp chips! He seems to be ignorant of the fact that modern analogue amp chips are closed loop power op-amps with high-current low-Rds FET drivers on the output, and the output will almost exactly follow the waveform regardless of load. Effectively they are almost exactly an "output impedance of zero".

Due to that fact, a bridged amp WILL produce 4x the wattage into the same speaker, at low power levels, and at high power levels for a limited time. The problem of course is that it will exceed it's operating spec and go into thermal shutdown, but that is very different from the Wiki stance that it *won't* produce 4x the power.

It will produce exactly 4x the power from the same music input, at lower levels, and for intermittant peaks (which is most of the music), it just won't produce 4x the power at max power RMS continuously. I think the Wiki author must be one of those non-EE "experts" that argue incessantly on the "audiophile" forums...

 

[audioguru]

The guy who wrote the Wiki..... article was talking about vacuum tube amplifiers with no negative feedback so their output transformer's impedance is thousands of times higher than modern solid-state amplifiers that have plenty of negative feedback.

Bridged amplifiers produce about 3.5 times the power, not 4 times because their saturation voltage loss is a little higher when the current is doubled.
Negative feedback reduces the output impedance of an amplifier so that the output level does not change when the load current changes. The negative feedback has no effect on the max output power.

 

[Nigel Goodwin]

So, referrring back to the original thread title then, linear amps are not likely to come near to the 'x4' bridge power output - too many heat losses from the 'variable resistor' part of the linear amp concept.
Class D might come nearer -as I suggested in my previous post?

Class-D makes no difference - they also produce slighly less than double power in half the speaker impedance, and for the exact same reasons. There's nothing inherently different in class-D in that respect.

 

[carmusic]

isnt class D more noisy then other class. i know they are often use in subwoofer amplifier (high frequency noise is not a problem) but i read for a full range amplifier thay re not as good as class AB amplifier for quality of sound (this is maybe why class ab is still use in low-mid level amplifiers)

 

[audioguru]

Class-D amplifiers from older designs used a fairly low switching frequency which causes distortion at high audio frequencies. So they are used as a sub-woofer amplifier.
But modern class-D amplifiers use a very high switching frequency so that there is very little distortion at any audio frequency.

They do not produce more audible noise, maybe some test equipment detects the high switching frequency as noise.

Class-D is efficient and do not get hot but costs more. The heatsink costs less so they are used for high power.
For low-mid level amplifiers the heat for a less expensive class-AB amplifier is low so they are used to save money.

 

[rogs]

Class-D makes no difference - they also produce slighly less than double power in half the speaker impedance, and for the exact same reasons. There's nothing inherently different in class-D in that respect.

As I mentioned earlier, a major benefit of class D topology is the efficiency, which can be pretty important when power supply votlages and capacities are fixed, for whatever reason.

In recent tests, I have measured the Hypex modlues I linked to earlier, at over 90% efficient. Way above anything likely from class A/B amplifiers. but where heat dissapation and extra power are not limiting factors, then I'm sure some of the fine A/B designs out there will continue to be well liked.


carmusic - noise can be a problem with some class D designs, but the Hypex amps I mentioned earlier are full range, very high quality amplifiers, and exhibit virtually no noise from the output. The Hypex site has some interesting papers on the UCD concept by the designer Bruno Putzeys. Quite interesting reading.

And, IMHO, they do perform astonishingly well. Probably why some of the highest quality Hi fi amp manufacturers - like Channel and Bryston for example - now include UCD modules in their equipment.

They need split rails though -so no good for car audio!!

 

[Nigel Goodwin]

As I mentioned earlier, a major benefit of class D topology is the efficiency, which can be pretty important when power supply votlages and capacities are fixed, for whatever reason.

Yes, I'm perfectly aware of the advantages of Class-D - but it has none for the point in question (maintaining double power output to half load impedance).

In recent tests, I have measured the Hypex modlues I linked to earlier, at over 90% efficient. Way above anything likely from class A/B amplifiers. but where heat dissapation and extra power are not limiting factors, then I'm sure some of the fine A/B designs out there will continue to be well liked.

Smaller, lighter, more efficient, wins all round really.

carmusic - noise can be a problem with some class D designs, but the Hypex amps I mentioned earlier are full range, very high quality amplifiers, and exhibit virtually no noise from the output. The Hypex site has some interesting papers on the UCD concept by the designer Bruno Putzeys. Quite interesting reading.

And, IMHO, they do perform astonishingly well. Probably why some of the highest quality Hi fi amp manufacturers - like Channel and Bryston for example - now include UCD modules in their equipment.

They need split rails though -so no good for car audio!!

Many car amps are Class-D, probably a higher percentage of in-car than home amps?.

Most car amps run on split supplies anyway, which are generated in the PSU which is used to increase the supply voltage to giver decent power outputs. Otherwise you;'re limited to 16W in to 4 ohms.

There's also no requirement for a Class-D amp to be split supply anyway, again no more than any other amp - they mostly are, just as other amps mostly are.

Noise wise, a properly designed Class-D amp is as quiet as a properly designed linear amp.

 

[audioguru]

They need split rails though -so no good for car audio!!

No.
High power car amplifiers use a switching power supply that produces a higher voltage split rails supply.
There are many class-D car amplifiers.

You think that a class-AB amplifier "wastes" power that reduces its max output power. It doesn't.
A class-AB amplifier has exactly the same saturation voltage loss as a class-D amplifier so their max power is exactly the same.
But a class-D amplifier switches the output transistors on and off so they are cooler.

 

[rogs]

You think that a class-AB amplifier "wastes" power that reduces its max output power.

I'm not sure why you think I have that opinion? Sorry to mislead.

Certainly, a class AB amp uses a lot more power to provide a specific output, than a class D amp.

Whether that is considered 'wasted' power???