# Peak Power Vs Rated Power

1. ## Peak Power Vs Rated Power

I wish to understand more on power parameter needed to design the power supply of audio amplifier.

Which power parameter should we use when designing the SMPS? Is it peak power or rated power?

Lets take an example as below:

SMPS topology:
Flyback quasi-resonant type

Assume Class-D audio amplifier requirement as below:
Peak power : 250W (in short period, less than 5s)
Rated power : 50W
Speaker : 8 ohm

This is example I found on internet https://www.eetimes.com/understandin...requirements/#

Based on above example, the BTL output is 50W. This 50W is based on RMS value. Based on that, the minimum power supply voltage is 33.4V.

But if we use peak BTL output 250W, this will give minimum power supply voltage around ~70V.

Based on above calculation, I conclude as below:

If use rated power, the minimum voltage power supply is ~33.4V
If use peak power, the minimum voltage power supply is ~70.0V

Based on that, which voltage should I use for my power supply output voltage? Is it 33.4V or 70.0V?

•

2. ## Re: Peak Power Vs Rated Power

It is normal to keep the supply voltage constant and ensure you have sufficient current capability from the power supply. I think you are looking at the problem from the wrong perspective.

It is more likely that the design challenge is to make sure the voltage remains constant at peak current demand. The voltage would, nevertheless, have to be high enough to allow the amplifier to deliver full power into the required load. The problem becomes managing the PSU losses at peak demand and therefore has to be looked at in thermal terms rather than adjusting the voltage.

Brian.

•

3. ## Re: Peak Power Vs Rated Power

Like rated power, peak power also refers to undistorted sine output, Vpk = 63.4 V for 250W into 8 ohms. Not sure how you calculated the supply voltage, I presume bridge output stage with some empirical voltage margin.

The specification of a peak power rating for amplifiers can be meaningful in two regards:
1. Unregulated power supply voltage drops under load. This can be expected for classical transformer power supplies, but not SMPS. Also the time constant would be rather 50 - 100 ms than 5 sec.
2. Thermal constraints, as mentioned above.
In any case, the supply voltage must be calculated for peal power.

•

4. ## Re: Peak Power Vs Rated Power

Cheesy low end consumer audio products, back when I
last looked, often appeared to play games with power
ratings. Like, yeah, you can get rated power for long
enough to run the test, but the heat sinks inside a
minimally perforated case and no forced air are not
going to hold safe temperature for long.

Fortunately in audio applications the peak / average
ratio is low, for normal content. Single or two-tone
tests, though, could put you in a bad place.

Power supply shortcomings can turn into clipping or
into tone-mixing by PSRR@frequency. The "frequency"
of linear supply behaviors, especially when driven by
the very signals they are supposed to be supporting,
probably depends a lot on the details of both amp and
supply.

5. ## Re: Peak Power Vs Rated Power

Just curios, if the peak power is needed only for short time (example if we play music that have some peak level in certain period), will this short peak power affect the thermal?

In general, I understand the challenge. Managing peak power at different low/high line need to be deal carefully to avoid core saturation which can lead to thermal issue.

•

6. ## Re: Peak Power Vs Rated Power

The answer depends on the duration of the peak power. Often this is unpredictable of course. The only safe way to be certain the ratings are met is to design for continuous peak power but most designs work on the principle that peaks are sporadic so there is some scope for cooling between peak periods and hence less heat to dispose of. No calculation will give an exact answer, it depends on the heat generated during quiet times, heat generated at peak times and the rate at which heat is removed by a heatsink or other cooling method. For music it is probably safe to 'assume' 50% will be peaks but there would be a big difference between say soft orchestral and heavy metal music content!

Brian.

--[[ ]]--