What is power dissipation?

[geminni]

Probably i am the laughing stuff now but i have never, realy never, have atention on this term.
If i, lets say, want to make a amplifier of 100wats and it "spends" 150 wat from power suply. What is discipation? Is it that whole 150w or is it just that 50 wats that goes in heat?
I have used some calculations for heatsink and there are some squares, divides, muliplies operations but it all just made me mad

[jackinnj]

power dissipation is energy which is wasted in the form of heat -- due to the inefficiency of the amplifier --

[Nordic]

In class A, the amp draws a constant voltage from the PSU, regardless of input signal... output is only a fraction of input.

[geminni]

So it means that if 100 wats "goes" to speaker, 50 wats is that dissipation... Theese are rough numbers i just made it simple
And that 50 wat is taken in calculation for heatsink?

Another question: that 50 wats waste both transistors (if having one pair) meaning 50+50wats which is 100w or they share it, so it means that each transistor "wastes" 25wats?
In calcuation by G Randy Slone there is taken that dissipation is divided by number of output transistors-in his case there are 4 devices.
The dissipation is first calculated by power suply voltage (one rail).

[AndrewT]

Hi,
you have two different situations.
1. driving a load with a continuous high power test signal.
2. listening to unclipped music at normal and even loud levels.

These two senarios require completely different heatsink dissipation capabilities.

For 1. (high power test signal) the dissipation (in a ClassAB amplifier) is approximately 60% of the maximum output power when driving unclipped sine wave. If you plan to do this you will need VERY BIG heatsinks to prevent overheating of the output stage and driver stage.

For 2. (normal music & ClassAB on low bias setting=optimal) you can manage with MUCH smaller heatsinks, gentle use can even survive with just the casing as a sink, but if you turn the volume up the transistors will get hot very quickly.

You, as designer, have to decide how reliable you want your amp for the operational conditions you predict for your type of usage.

Very few will design the heatsinks for continuous maximum power duty, except commercial PA amps that must keep the customers dancing or similar.

[geminni]

Yes, i need a heatsink for PA amplifier. Till now i made only amplifiers maksimum of 100wats. For theese amplifier i allready had some heatsink with fans and i didn't have to calculate Rthja for heatsink. Now i want to build 2x380wats and want to be shure that it would have enoguh heatsink. It will be class AB and by calculation for one channel it , with 3 pair of IRFP450, heatsink has to be 0,21. This is done by calculation of Randy Slone.

[AndrewT]

Hi,
have you interpreted Slone correctly?

power output is stated @ 380+380W.
60% =456W=max dissipation
delta Ts-a=0.21*456=96Cdeg.
assuming Ta=25degC (a cool night in a disco- 35degC might be nearer the worst case).
Then sink temp is predicted @ 121degC.
with 12 devices each dissipating 38W delta c-s could be 10Cdeg giving Tc max ~=131degC.
derating for a a 190W 150degC device gives max Pd=29W.
three pair of devices @29W = 87W.
This amp/heatsink combination can push about 30W +30W into your load.
The sink has crippled the amplifier if you require PA duty.
Even if you could keep Tc below 50degC, three pair of p450s will not reliably push 380W into any load.

[geminni]

Well, that Slone...I don't know...