How hot may resistors run ?

[Bernhard]

I have a 25W golden Dale heat sink mounted power resistor that is dissipating 10W.

The heat sink is only double the size of those small heat sink that are usually used for 78XX regulators.

It is getting really hot, I can touch it maybe for half second.
There will be some ventilation but not much.

Unfortunately the space is restricted and I can not use a larger resistor or heat sink.

Is that safe for long term ?

[AndrewT]

Hi,
the little flags attached to To126 and To220 are ~20C/W.
About double the size is about 15C/W.

assume Ta=25degC
Delta Ts-a = 10W * 15C/W = 150Cdegress. but this needs to be derated for Delta >75C
Ts ~ 130degC.
Delta Tc-s = 10W * 3C/W = 30Cdegrees.
Tc ~ 160degC
Delta Tj-c = 10W * 1c/W = 10Cdegrees.
Tj ~ 170degC = a very hot wire that is no longer it's nominal resistance.

Put your own numbers in and find a suitable value of sink Rth s-a that limits Tj to a sensible long term temperature.

Hint,
start with Rth s-a = 1C/W

[Bernhard]

Thanks, I have no choice with the heatsink.

The regulator for one protection rail of a Nikko Alpha 2000 amplifier was completely shot.

The main transistor has to stabilize from 100V down to 33V, it is mounted on the main heatsink that is not accessable without risk...

So I put the transistor with heatsink on the regulator pcb and added the Dale to take away the dissipation from the transistor.

[Magura]

If you turn your heatsinks, so the fins are vertical, and at best, elevate them 10mm from the board. That should bring you somewhat better performance.

If that's enough, well measure the temperature. Keep in mind that your resistor may fry other components in the vicinity if running at 150C.


Magura

[Bernhard]

There is a cooling fan that blows a little bit into the basement.

The question is just,
is it ok if the resistor is so hot that I can touch it not more than half second ?

That is of course not 150°C


Edit:

consulting the datasheet

Vishay Dale RH25

@ 25°C = 9W unmounted

operating temperature range up to 200 °C

[cliffforrest]

Is it OK? Probably ...

Is it wise? No!

Long term you will get problems.

In general, any component too hot to touch (that is really scientific!) is bad news: Burnt PC boards, dry joints, brittle wires etc etc.

100V to 33V really needs a switcher ....

[Bernhard]

If that was my design I would simply have used an additional transformer winding...

[Conrad Hoffman]

Good suggestions above about turning the heat sink. IMO, you're probably well within the limits, but hot things don't last as long as cool things. What's the desired service life? Industrial equipment isn't usually expected to last forever, and service is expected, so parts are sometimes run near their limits. I have test equipment that's required zero service at 50 years old, due to conservative design. I also have test equipment that was reduced in size, ran hot, and didn't last 5 years without service. I like to think in terms of decades, so if I can't touch most parts for a few seconds (about 160F), they're too hot for me.

[Bernhard]

One of the most stupid things I remember, was in a Grundig 9009 CD player, a real hot regulator IC with heatsink was mounted directly beside a DAC chip.

[apassgear]

Another alternative is to use two resistors in parallel (double the value of course) and replace the HS with a small alu plate… it should be enough.