I know that a rule of thumb to judge the efficiency of a traditional iron core transformer is to look at its size: the larger it is, the higher the number of windings, the larger the core, the ticker the wires... All factors that in the end lead to lower energy losses and higher efficiency. (Or maybe I should say the other way around: a requirement for a transformer to be efficient is to be large.)

But what about modern switch-mode adapters? Empirically, I can feel that small adapters heat up very easily, even with low loads, while larger adapters don't heat up much (if at all) even with high loads. This could be because larger adapters disperse heat more easily, but it could also be because they're truly wasting less energy.

Hence the question: assuming the same load, same input wattage and same output voltage, would a larger switch-mode adapter be more efficient than a smaller equivalent?

Switch-mode adapter empirical comparison

  • \$\begingroup\$ @DKNguyen: you're absolutely right. That's why I asked the question \$\endgroup\$
    – josh
    Commented Apr 5, 2020 at 18:30
  • \$\begingroup\$ Overhead power used to keep itself running favours being small. Power used to run the load favours being large. So they run counter to each other which means there is a sweet spot. Overhead power in electronics tends to be pretty low though relative to load power unless you are idling a lot. \$\endgroup\$
    – DKNguyen
    Commented Apr 5, 2020 at 18:33

1 Answer 1


Your photo comparison is fun, but:

Don't forget that "heating up" is a function of wasted power, and ability to get rid of heat, and that

  1. white is not a good color for radiative cooling, and
  2. all ambient cooling mechanism scale with surface area.

Anyways, the general rule of thumb holds:

The losses incurred in a switch-mode power supply are in two places:

  1. switching losses in the semiconductors (mostly: MOSFETs in these compact devices)
  2. losses in the transformer

For 2., the same rule of thumb as you quote applies.

For 1., generally, the lower the rds,on of a FET needs to be, the larger it has to be.

So, yeah, same, but also:

You don't have to optimize a cheap 10 W Apple-design-imitation wall wart supply as much as a 100 W notebook supply; 15 % of 10 W isn't that much. 15 % of 100 W is significant. Relevant "What If?" article.


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.