In this highly inefficient circuit:


simulate this circuit – Schematic created using CircuitLab

(I forgot to insert a 1K resistor before the pot in this diagram)

I have 2 components that I use as the TRANSISTOR of this circuit. A STP55NE06 (TO-220) (which for some reason, is shortened a P55NE on the marking and fades off quite after a while), and a large TIP35C (TO-218).

The circuit works as expected, however they get really hot, where the STP55NE06 would reach about 90 degrees celcius and the TIP35C about 80 degrees celcius, due to the obvious wasteful power dissapation.

Recently, I noticed that in the TIP35C, the bottom back of the plastic looks a bit burnt. I looked in the datasheet and saw that it could handle the current and voltage of the bulb and the temperature as I measured wasn't over. Do I need a heatsink? Also I think I made an error and had the base voltage over 5V (12V in fact). Would that be the reason? The transistor is working fine as I can see from this circuit but I'm a bit worried about discolouring of the plastic.

In addition, the STP55NE06 looks not so good as well. THe bottom back of the metal looks a bit burnt (or is that corrosion) but the plastic isn't as burnt as the TIP35C.

enter image description here

Have I done anything wrong? Is this normal?

  • 2
    \$\begingroup\$ Temperature may mean die (internal) rather than case temperature - heatsink strongly advised. \$\endgroup\$
    – pjc50
    Sep 7, 2016 at 14:33
  • \$\begingroup\$ Do you have a heatsink on the transistor? \$\endgroup\$
    – Eugene Sh.
    Sep 7, 2016 at 14:33
  • \$\begingroup\$ Yes you definitely need a heatsink with that circuit. \$\endgroup\$
    – brhans
    Sep 7, 2016 at 14:47
  • \$\begingroup\$ There was no heatsink... Well not sure if I can be bothered buying and adding one. Also as a side question, the TIP35C would be superfluous for this situation, and only the STP55NE06 is good enough, right? \$\endgroup\$
    – Bradman175
    Sep 7, 2016 at 14:50
  • \$\begingroup\$ Do a thermal calculation and notice that the Rth(j-a) is in the order of 50K/W for a TO220. These things get awfully hot with just 2W. \$\endgroup\$
    – Arsenal
    Sep 7, 2016 at 14:51

1 Answer 1


Do the math. We don't know what your load is, other than it takes 20 W at 12 V. Its current at 12 V is therefore (20 W)/(12 V) = 1.67 A. That means, at least at the 12 V operating point, its effective equivalent resistance is (12 V)/(1.67 A) = 7.2 Ω.

Without knowing more about the load we can't really know what its equivalent resistance at other voltages is. However, for purpose of illustrating how to calculate power in the transistor, I'll pretend the load is a constant 7.2 Ω (very likely does not match reality).

The power dissipated by the transistor is the voltage across it times the current thru it. When the transistor is off, the current thru it is 0, so the dissipation is also 0. When the transistor is fully on, the voltage across it is very low, so the dissipation is also low. The worst case is right in the middle, when the transistor has half the supply voltage across it. That means the load has the other half across it.

With a 7.2 Ω load with half the supply voltage across it, the current is (6 V)/(7.2 Ω) = 830 mA. The power into both the load and the transistor is then (6 V)(830 mA) = 5 W.

That's a lot for what looks to be a TO-220 case to handle. Look at the TIP35 datasheet and see what the maximum allowed dissipation is without a heatsink. If I remember right, the TIP series usually specifies maximum power in 25 °C ambient air. I don't even have to look to know that is less than 5 W. With the right heat sink, a TO-220 should be able to dissipate 5 W quite well, but without a heatsink it will fry.

  • \$\begingroup\$ I should note that the TIP35C is in a SOT-93 package which is much larger. How would that be any different? \$\endgroup\$
    – Bradman175
    Sep 7, 2016 at 15:09
  • \$\begingroup\$ So would that be the reason why the plastic is melting? \$\endgroup\$
    – Bradman175
    Sep 7, 2016 at 15:13
  • \$\begingroup\$ @Bradman175 SOT-93: Rth(j-a): 35K/W compared to TO-220: Rth(j-a): 50K/W, without a heatsink, at 5W the SOT-93 will stay a whopping 75K cooler (at the junction) than the TO-220. The SOT-93 I found has a specified maximum dissipation of 3.5W at 25°C. \$\endgroup\$
    – Arsenal
    Sep 7, 2016 at 15:25
  • \$\begingroup\$ If the '12V20W-IKEA' is a halogen bulb then its resistance reduces at lower brightness. 50% current occurs at about 3V, so the transistor's maximum power dissipation will be significantly (50%?) higher than it would be with the equivalent (20W@12V) resistor! \$\endgroup\$ Sep 7, 2016 at 19:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.