"Silvia Janssen" <(E-Mail Removed) THIS.nl> wrote in message
news:brc8jt$sk1$(E-Mail Removed)...
> I've connected an industrial PC to an LT1085 voltage regulator which
> converts a transformer output of 27V to the 24V that is needed for the PC.
Is that a typing error? Do you mean "the _5V_ that is needed for the PC"?
The LT1085 is a 5V regulator isn't it?
> It all works fine, except when I shut down the PC with its own
> on/off-switch.
> In that case the output voltage of the LT1085 drops down to about 3 Volt
That doesn't sound right. You would not expect the voltage to _fall_ when
you _remove_ the load unless something is very wrong. When you shut down a
PC it normally results in a burst of disc activity that increases power
consumption briefly. Perhaps your regulator is overloaded and this extra is
the final straw?
> Would I need a regulator that can handle larger currents? Or is there
> another solution?
It might not be the current that's the problem but the power required could
be.
How much power does the PC need?
The data I've seen for the LT1085 says it can provides 7,5A at 5V. If your
PC is taking say more than a few Amps I hope you have a big heat sink
and/or a fan on the LT1085 !
Lets say the PC need 7A max. That's less than the 7.5A limit on the
regulator BUT....
I calculate that the power dissipation in the regulator would be...
(27V - 5V) x 7A = 154 Watts !
The LT1085 is available in various packages. These have max thermal
resistance specs of between 1 and 3 degrees per watt. (even even connected
to a perfect/ideal heat sink).
2 degrees x 154 = 300C
I think it's getting too hot.
The real problem is that the LT1085 is a linear regulator. I think if you
are taking more than a few amps you will need to look at using a switching
regulator to reduce losses in the regulator and allow a smaller practial
heat sink.
Colin
(E-Mail Removed) (delete BOX)