As fixed regulators tend to be cheaper than power resistors,
I prefer to use an other regulator in front with even better results, let=B4=
s=20
say you got 24V
and need 5V with like 100mA, so take a 12V regulator first=20
(24V-12V=3D12V*100mA=3D1.2W)
and than the 5V regulator (12V-5V=3D7V*100mA=3D0.7W) and you get off without=
=20
needing
the extra cap, which results even cheaper.
Friedel

At 09:30 p.m. 13/11/01 +0200, you wrote:
>For regulators with a large voltage drop across them the power drain of
>the regulator proper becomes a significant part of Pd at low loads. With
>devices used without a heatsink this can be the straw that breaks the
>camel's back and cause strange faults in cased devices (regulator enters
>thermal shutdown under hard to duplicate conditions - i.e. at the
>customer's location, over the bridge, down the lane, etc, on hot Sundays
>exclusively).
>
>Pdreg =3D U1 * Iddreg + (U1 - U0) * Iload
>
>7805 series regulators use little extra Iddreg to drive the output
>transistors because they are darlingtons and the base currents also flow
>to the load. With other regulators this is not so.
>
>The self-heating problem is especially visible with TO92 regulators. The
>usual fix is to add a R in series with the voltage input such that at max.
>load Uin on the regulator is higher than the required min. input voltage.
>A decoupling capacitor is required after the resistor.
>
>R =3D (U1 - Uinmin) / Iloadmax
>
>It is possible to split an arbitrary part of the power to be dissipated
>into the resistor (i.e. 50%, 80% etc) and it need not be just one
>resistor, it can be a series combo. Me, when I do this, I arrange for the
>power dissipation in each device to be equal (including in the regulator),
>so I have no hot spots on the board (with similar sized devices - e.g.
>TO220 7805 and 2W resistors). The formula for that is:
>
>Rtot =3D (U1 - Uload) / (2 * Iload)
>
>This is inefficient but nobody cares whether a couple of resistors
>dissipate 2W in a unit that draws 100 or 3000W. You can put other things
>in series if the current is reasonably constant, like dial lights, pilot
>lights, anti-dew heaters, LCD backlights, even a small fan with
>decoupling. The fan trick is neat because as the circuit draws more
>current the fan accelerates ... (assuming DC brushless fan).
>
>A 7805 TO220 is not to dissipate more than about 1.5W without a heatsink
>(at 50C ambient) and it will run very hot to the touch at that. Up to
>about 1W into a device of that size will keep it 'touchably' warm. At
>24Vin and 5 Vout that is less than 50mA total, leaving only 40mA for the
>load (acc. NS data book).
>
>With an 'equal dissipation' R the current can be doubled under the same
>conditions (R =3D 180 Ohms, 2W). Decouple the regulator input after the
>resistor with 10 uF 16V to GND. The resistor will run about as hot as the
>TO220 case (not as hot as the TO220 case without the R). This will allow
>loads up to 80mA and may save your design as is now (cut trace, connect
>resistor with wires - after packing it in heatshrink tubing and add
>decoupling capacitor on bottom of board - if you have room). Remember that
>heat does not go away, you will have to put the resistor in a place where
>it does not heat something else - like making ugly dents in the plastic
>case after a while).
>
>Peter
>
>--
>http://www.piclist.com hint: PICList Posts must start with ONE topic:
>[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics