squeezebox power supply

[zoubour]

I tried to upgrade the power supply of my squeezebox by using a diy kit. Tested on the multimeter, it gave me 9.7V, while the squeezebox wall wart is specified at 9V.

My squeezebox worked for 3 minutes playing a piece, then went off.
Several times, I was able to restart it and play for about 3 min, but in the end it stopped working altogether.
Everytime after it stopped, the LM317 voltage controller was getting very hot.

Any idea about the cause and what to do about it?

Thanks

Z.

 

[Leeuwarden]

Perhaps the 317 needs some cooling.

 

[zoubour]

LM317 was on a heat sink

I even used two different kits, with the same results.
I don't understand what could be different for the worse between the SMPS walwart and my linear PS

 

[dsavitsk]

Maybe Logitech changed something, but my SB takes 5V.

At any rate, an overheating LM317 will shut itself down. The internal regulators might do the same, and since you are giving it .7 extra volts (or 4.7 extra), that might be enough.

 

[zoubour]

This is the duet. The wall wart outputs 9 V.

 

[pilli]

Check whether it is your supply that shuts down, possibly because of overheating protection of the 317.
What voltage does the 317 get in input?
I think the current drawn by the Squeezebox Receiver isn't huge, but still...


-

 

[VT67]

Which DIY kit are you using? Is there a heatsink on the LM317?

Regards

 

[zoubour]

There is a heatsink on both kits I used.
It's a possibility there is a shut down on the supply side, but why?
The max current from the wall wart is like 450mA. The LM316 is capable of handling much more. So the question is why does it get so hot?

 

[Bas Horneman]

So the question is why does it get so hot?

What is the voltage before the LM317?

 

[pilli]

So the question is why does it get so hot?

(Sorry, you may know all this already)

The heat comes from the power to be dissipated, and that is
(Vin - Vout) x I

If your Vin is much more than 9V (your Vout)
and your heatsink is too small, the heat cannot go away from the chip, and the chip shuts down (lucky and smart that it has thermal protection - a transistor would just melt).


Just as a gross guess:
you are at about 500mA;
those TO-220 heatsinks that are like one inch on each dimension, have maybe something like 5°C/W rating;
you said it's "very hot", and you usually cannot stand to touch above 60°C
(when the inside of the chip is quite hotter)

so let's say that this heatsink is at its limits with a raise from ambient temp of about 40°C,
which is 8W as calculated here:
40°C divided by 5°C/W = 8W

At 500mA current, you get 8W for 16V
If your voltage in-out difference is much less than 16V then you should be fine.
(This is large, you'd be feeding 25V input.)

Now if your heatsink is smaller, and has only 10°C/W rating,
then you get that hot at 4W that only leaves you 8V (still large...)


I hope I didn't make too many mistakes...

A much nicer description of the thermal situation is for example in the datasheet from National
http://www.national.com/ds/LM/LM117.pdf


Tell us what transformer you are using, and the dimensions of the heatsink, and number of fins.

-

 

[zoubour]

Thanks Pilli,

I have been pumping 24V AC in. That might be the problem with a small heatskink.

 

[Bill Fuss]

Thats 34VDC so you are dropping 34 to 9.7, thats 25v at 450ma, close to 12 watts dissapation. Too much.