LM338 Regulated PS

[Puffin]

I have fitted heat sinks to the 338's, but find that they remain virtually stone cold in operation. The chip heat sink gets mildly warm, but not a warm as before adding the regulation.

Have any others found this? I use very efficient speakers and so the amp is under no stress.

[Ted205]

it depends on the voltage you are putting into them, the voltage out, and the current drawn by the amp

[Nuuk]

Originally posted by Puffin
I have fitted heat sinks to the 338's, but find that they remain virtually stone cold in operation. The chip heat sink gets mildly warm, but not a warm as before adding the regulation.

Have any others found this? I use very efficient speakers and so the amp is under no stress.

Yes.

[AndrewT]

Hi,
typical music levels are average power out ~=peak power out/100 (-20dB).
if your 50W amp is peaking at 20W then the average output could be as low as 200mW. This is little more than running at quiescent current draw.

Similarly the regulator is only supplying that 200mW + quiescent current draw and that times the volts drop across the regulator accounts for the tiny temperature rise in the sink.

Try measuring the volts drop across the regulator, when no music is playing and when the music is as loud as you would normally play it.

[pacificblue]

Originally posted by Puffin
I have fitted heat sinks to the 338's, but find that they remain virtually stone cold in operation.

And now you think you wasted the money for the heatsinks? You didn't.

Originally posted by Puffin
The chip heat sink gets mildly warm, but not a warm as before adding the regulation..

That is probably due to the reduced ripple on the rails.

Originally posted by AndrewT
Hi,
typical music levels are average power out ~=peak power out/100 (-20dB).

Doesn't that rather depend on the speaker efficiency than on the amplifier's peak power? Somebody, who listens at an average of 60 dB with a speaker that has 80 dB/W/m should be using an average of 0,01 W, no matter if the amplifier can deliver 1 W, 100 W or 10.000 W.

[AndrewT]

Originally posted by pacificblue
Doesn't that rather depend on the speaker efficiency than on the amplifier's peak power? Somebody, who listens at an average of 60 dB with a speaker that has 80 dB/W/m should be using an average of 0,01 W, no matter if the amplifier can deliver 1 W, 100 W or 10.000 W.

of course, but look at it from the other direction and having already committed to a chipamp with a maximum of ~60W.
200mW into 87dB/W/m speakers gives an averge level of ~75dB @ the listening position for a two channel system @ 2.4m listening distance.
The peak level can be about 95dB and still have 30 to 40W (another ~5dB) to spare if you want to turn it up a bit more. Alternatively if ~75dB is too loud, then turn it down a bit.
Either way it explains why the regulator and the chipamp are cooler than might be expected, if one has not thought about average listening levels.

[Puffin]

Nuuk and AndrewT. Thanks for your responses. I suspected that the current draw was actually very little. I came across a project that showed how to build a very simple peak level meter. I set the resistors for 1Watt. It was surprising how much (deafening)volume there was before the LED started to register.