# How much power does LM7805 dissipate?

I'm using an LM7805 in voltage regulator mode.

• Vin is 15V
• Vout is 5V (regulated)
• The attached load draws 200mA

How to calculate how much power the LM7805 would dissipate in this setup (and whether or not I need a heat sink)? Couldn't understand it from looking at the datasheet.

P.S. I'm asking because it gets quite hot, which I didn't expect. Checked the attached load, it is around 0.2A, so well within the limit of 1.5A.

• For a linear regulator, Iout is roughly equal to Iin. All voltage dropped across the pass device is dissipated as heat.
– M D
Commented Mar 30, 2019 at 10:55
• Wait, this has been asked multiple times, see here. Commented Mar 30, 2019 at 12:38
• Also there is more detailed datasheet here Commented Mar 30, 2019 at 23:37

Take a look at package thermal data section in your datasheet page two.

Depending on the package of the regulator you are using, lets say the thermal junction-ambient coefficient $$\\theta_{JA}\$$ is roughly $$\20°C/W\$$.

You've got $$\ P = V \cdot I = (15\ \text{V} - 5\ \text{V}) \cdot 0.2\ \text{A} = 2\ \text{W} \$$ dissipated as heat.

If your ambient temperature is $$\25°C\$$, then the regulator would heat up more or less into $$\65°C\$$.

It is quite hot for sure.

• 2W is about the limit for use without a heatsink. Commented Mar 30, 2019 at 11:48
• Yes, I'm very linearizing here for simplification. The ambient will also heat up which eventually also increase the regulator temperature. Also, in the NOTE 1 below the package thermal table in the datasheet OP linked there are maximum power dissipation formula and careful warning statement which would be useful for OP while learning this topic. Commented Mar 30, 2019 at 11:56
• Different manufacturers have very different thermal resistances. For TI it's 19 °C/W (--> 63 °C) for STM 50 °C/W (--> 125 °C), for Fairchild even 65 °C/W (--> 155 °C). The latter two will probably die without a heat-sink (see absolute maximum ratings). Commented May 3, 2020 at 15:40
• As far I remember, I assumed 20 as a best case for quite hot. +1 for clarifying. Commented May 4, 2020 at 12:40