1

I have a problem with my power audio amplifier.

I am unfortunately forced (because of the specifications I have been given) to make it work at a dual voltage of +-90V.

Finding the components to do it was not difficult, I built it, I balanced the currents of the whole amplifier and now it is perfectly stable at that voltage.

The problem, however, I have in the final stage:

enter image description here

My drivers are, as you can see, KSC3503 and KSA1381, and the power transistors are 2SC5200 and 2SA1943.

The problem: the amplifier is current and thermally stable. But the power transistor power dissipation is too high. For only 10W of output in 8 Ohm I have to dissipate 24W of heat. I tried to increase the resistance of degeneration (0.1 ohm even up to 47 ohms), but the problem is not solved: what the transistors should dissipate, the resistors dissipate. Did I forget something somewhere? I should be able to deliver about 200W with this amplifier, but I can't dissipate 480W of heat. (P.S. variate R89 does not solve)

B. Dam
  • 21
  • 3
  • hm, is it really 480 W of heat? When you calculate the waste power in the moment the OUT is pulled as low as possible or as high as possible, how much is that? – Marcus Müller Aug 31 '19 at 09:04
  • Maximum efficiency is at maximum power output. What does this tell you about your assumed power dissipation of 480 watts. – Andy aka Aug 31 '19 at 09:10
  • What is the quiescent current draw? – Bruce Abbott Aug 31 '19 at 09:12
  • Nevermind about 480W of dissipation, is a number that I pulled out by making a simulator estimate. @MarcusMüller. The problem is another. – B. Dam Aug 31 '19 at 09:13
  • @BruceAbbott technically 0.9W for the entire amplifier, with the bias almost in class B – B. Dam Aug 31 '19 at 09:15
  • @B.Dam no, seriously, it's not a different problem. If your simulator says you've got 480W of waste at the maximum power point, then there's no way around that using this amplifier class and these components, UNLESS you made a mistake in your simulation. So, focus on whether that simulation result matches a hand-calculation of waste power in the maximum power point (which is what I tried to recommend in my first comment). – Marcus Müller Aug 31 '19 at 09:16
  • @MarcusMüller Okok I reviewed and made a mistake: the power consumption at 200W output is 60W for each power transistor wich means 240W. Isn't it still very high? It's however under the efficiency of a class AB technically, no? – B. Dam Aug 31 '19 at 09:21
  • Technically I didn't make any mistake, the total power dissipation is effectively 480W (of the two branches, N and P). – B. Dam Aug 31 '19 at 09:24
  • What is your bias current set to? Quite apart from the efficiency, you will want to think long and hard about the SOA of those transistors... Particularly if the load is reactive (Second breakdown is very much a thing in audio power stages). – Dan Mills Sep 01 '19 at 09:58

1 Answers1

3

Class B has a maximum theoretical efficiency of 78.5% at maximum undistorted sine wave output. However at lower power output the efficiency drops because there is more voltage across the transistors and less across the load. At very low power the quiescent power consumption also becomes significant, so a high power amp will have lower efficiency than a low power amp at low output.

Here's an example of efficiency and power dissipation vs power output for a 12W audio power amp IC (TDA7240):-

enter image description here

We see that power dissipation actually increases at lower output power, and is the same at 1.5W output as it is at 12W.

An amp that delivered 500W into an 8 ohm load at 75% efficiency would dissipate 167 Watts. It might also dissipate 167 Watts at 60W. Your 24W for 10W output sounds quite reasonable.

Bruce Abbott
  • 55,540
  • 1
  • 47
  • 89