Question on efficiency if PWM amps and PS

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
The efficiency of the amplifier does not depend on the type of power supply used. Switch mode power suppies are more efficient than linear supplies, and in many cases cost less. Class D amps are fairly happy with switch mode supplied because noise from the power supply will not affect the signal as much as with say a class AB amp, generally speaking.

Switch mode power supplies are also smaller and weigh less than their linear counterpart if that is a concern in your design.
 
Noah,

Many switching PSU's use a linear supply for a front end... so it is entirely possible for a linear supply to be more efficient than a switcher. Switcher's claim to fame is reduced size and weight. In the case of a large supply, this translates into cost as well.

Yes, you will realize PWM efficiency regardless of PSU type.

:)
 
"Many switching PSU's use a linear supply for a front end..."

Do you mean many switching amps?

If not, what's a front end to a PS, and what's after it?

"Yes, you will realize PWM efficiency regardless of PSU type."

That's good to know in any case :)

I can't think of anything (xformer, rectifier, caps) in a regular PS that should be very dissipative.

Thanks
 
The general desire in a switcher is to run directly off line... this eliminates the mains transformer altogether. Because the input to a switcher PSU can be made very tolerant of ripple voltage, the size of the filter caps cam be dramatically reduced as well. All this creates a raw voltage (front end) for switcher circuitry (DC to DC converotr) to convert and regulate up, down, whatever.

As far as switching PSU's are concerned; they exhibit efficiency worse than a linear at low power output levels... Why? the lights are on, the AC is on, the computers etc... but everyone is at lunch.



;)
 
Well, to put it in retrospect; a switcher has all this stuff to run when it's idle.

For example, the first section of the power supply must draw current to run the PWM IC(s) and then the driver and output transistors which in turn energize the transformer. After the transformer, you have rectifiers, output caps, filters, etc and a feedback loop. Many switching supplies also have fans...


All of these things draw little bits of current, and in some cases, a fair bit depending on the power supply.
 
A well-designed switching power supply consumes a bit of power with no-load. This is due to small losses in the primary components, power drawn from the control circuitry and perhaps some loading inside the supply to keep the feedback loop under control in these conditions.
For example, our bigger coldamp (around 1KW) switching power supply losses around 4-5W with no load. (I speak about it only because it is the one I know better)

A comparable power linear power supply has some losses, too, and they will be well in that order, due to transformer self-heating, etc.

As you increase power consumption, usually the efficiency of the switching power supply gets better and better as compared with the linear PSU of comparable power capability.

Our experience with our switching power supply and modules is that the modules get considerably warmer at idle (and they dissipate only around 10W at +/-60V with no signal) than the associated PSU, even when the PSU feeds two modules ;-)
 
The question of efficiency is not so black and white.

Looking at full load voltage conversion only, the linear supply will usually be the winner. However, this ignores the requirement for regulation. In linear supplies, regulation is usually accomplished by a series pass element; this element trims the voltage down by wasting power. Linear supplies may also use significant amounts of R between C's for ripple reduction, these also burn power. Linear supplies must also cope with line voltage variations, +/- 10%; in order to accomplished this, the linear must be capable of full output voltage at low-line conditions. At high-line conditions, because of the fixed turns ratio, the linear supply may be create an excess of 20% output voltage. This must be eaten by the series pass regulator.

Switchers have an advantage in that voltage regulation is accomplished within the voltage conversion hardware itself. This avoids the power loss associated with series-pass regulations elements.

;)
 
Is it a requirement? I thought most conventional PS are not regulated and that the feedback loop makes it unnecessary in any case.

Sorry... but my fundamental approach is to electronics in general, and less so to the specific requirements of audio. Regulation is desirable feature for most power supplies and is also required in many audio designs.

Feedback does indeed greatly reduce the need for regulation in many, if not most, audio amps.
 
Hi Poobah,

All good points you've made except you went on the basis of the linear supply having to compete on a regulated basis with the smps.

In practice they're usually raw dc unregulated, at least where power amps are concerned, which gives them every edge towards both efficiency and robustness, also when it comes to a homebrew one-off you're never going to approach it in design cost either.

Your example also made use of the worst possible way of regulating with respect to efficiency, series pass, what about cap multipliers, shunt regulation, or ground return regulation?


From what I gather alot of the power amp SMPS' seem to be going unregulated as well, so would they not also vary with both line and load regulation to some degree?

Like you say, it is never so black and white.

Cheers,
Chris
 
Like I said Chris, I was just adressing things in general... not as they pertain to audio. The reason being that some rather non-audio blanket statements were made regarding efficiency of one versus the other. The overwhelming demand require supplies that are regulated... this is ONE reason switchers have garnered the rep for being more efficient when it not strictly the case.

I wouldn't recommend a SMPS for a DIY thing either.

As far as regulation is concerned... you either convert power or burn it. A shunt burns power. Any method that uses conduction angle modulation is really a switcher wearing linear clothes.

Regarding SMPS without feedback... I'll have to think about that. The output caps must be huge though. Sounds whacked too.

There is another and perhaps more important reason why switchers came popular... where they got their start... but it's not audio...so.


;)
 
Hi again,

Yeah, your comments were of course on the money, but least people get confused when it comes to audio I guess I thought I'd present the other side of things, at least from a power amp perspective.

I can't and won't argue that shunt regulation burns power rather than converting, but it's far less of a villain in this respect than series pass. I can't say if it can compete with the efficiency of a switcher though, but at least it can approach it?

Interesting comment you made about phase angle regulation..

We've opened a nice can of worms havent' we? :D

From what I see they do make the output caps huge, nearly as huge as in a RAW DC, at least much more capacitance than in your usual switcher, I guess to help load regulation out, and also help absorb freewheel current.

I'm not 100% on the reasoning for doing away with the feedback control loop of the supply, apparently it can interract with the audio .. possibly also through pumping? I'd have to look that up as it's been discussed before.

It's an interesting case of adapting the existing technology to audio purposes though, and active PFC seems to be the answer to make it robust against line variation.

There is another and perhaps more important reason why switchers came popular... where they got their start... but it's not audio...so.

All I can think of is reduced cost in terms of mass production (increased profit margin), and as you pointed out, increased efficiency when it comes to regulation.

Apparently the demand for cell phones (with long battery life), laptops, PDA's etc, were a big factor in the popularity of switchers, I'd also throw in high current regulators for CPU's as well.

It's because of all these that we can thank the industry for finally providing us with "decent" components for class d.

Anyway I do appreciate you having shared your views from a general electronics /non audio perspective. It does no harm to explore all aspects of a topic.
:D
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.