4
\$\begingroup\$

Does current have an effect on the wire it moves through? I think I heard somewhere that over time small wires in CPUs can move because of the current going through them. If it is a thing what's the technical term for it?

\$\endgroup\$
  • 3
    \$\begingroup\$ Like melting..? \$\endgroup\$ – Paul Uszak Jun 10 '17 at 12:41
  • 1
    \$\begingroup\$ As far as I understand it... The power dissipated by the heat produced affects it but not in a permanent way unless you get very hot. The magnetic field it creates in the wire could cause motion of the wire. In a straight wire I'd expect that to be tiny/nothing against the stiffness of the wire. But both of these are secondary effects, not change because of the passage of electrons through the material. \$\endgroup\$ – TonyM Jun 10 '17 at 12:46
  • 1
    \$\begingroup\$ Link to what you heard/read/imagined please. So go googling and find an article then, if the article isn't clear come back and ask a sensible question. If the answer is clear then job done, come back and delete your question. \$\endgroup\$ – Andy aka Jun 10 '17 at 13:13
14
\$\begingroup\$

Yes, at the nano-meter scale you always run into an issue titled Electromigration.

At the nano scale a conductor has to not only be rated for voltage drop but also for current density. To achieve sufficient reliability due to deterioration from electromigration.

\$\endgroup\$
  • 1
    \$\begingroup\$ Yes, I recall this phenomenon being described to me by an IC engineer at RCA in 1972. The effect, within an IC, is something like the flow of water in a river causing an "oxbow" or "meander" -- the metal comprising a turn in the wiring would migrate to the outside of the turn, until either the wire got too thin to carry the current, or until it migrated into an adjacent wire. Obviously, component failure was the result. (No doubt the phenomenon had other effects as well.) \$\endgroup\$ – Hot Licks Jun 10 '17 at 19:21
3
\$\begingroup\$

The vast majority of the time, the physical effect of a current through a wire is negligible.

High current densities can physically effect a wire in 4 very different ways:

  • The electromigration effect ( as mentioned by sstobbe ) is the gradual movement of ions along the length of a conductor due to the momentum transfer of electrons moving in a wire. At high current density, eventually enough ions of the metal wire get pushed out of where they were originally placed to leave a gap (an open circuit) in that wire, or build up enough displaced ions to bridge the gap between that wire and some adjacent wire (a short circuit). This can cause long-term reliability issues -- when people design things to run right on the edge or slightly past manufacturer's recommendations, this effect allows the device to seem to work fine for a while, but lead to early failure.

  • The resistive heating of wires, also known as Joule heating or "I squared R heating", is proportional to the square of the current through the wire. When an integrated circuit is stressed by voltages far beyond the manufacturer's recommendations, this effect often melts or vaporizes one or all of the small wires of the IC.

  • The magnetostriction effect causes some materials to change shape when magnetized. This effect is the most common cause of transformer hum. (The current through the coils of the transformer creates a strong magnetic field, which leads to magnetostriction in those wires or in the ferrite core or both).

  • In a strong magnetic field, the wire is physically shoved sideways by a force proportional to the current and to the magnetic field -- that force is sometimes called the Laplace force, but more often calculated as part of the Lorentz force (as alluded to by Uwe ). This effect also causes transformer hum.

\$\endgroup\$
-1
\$\begingroup\$

If strong currents flow in two parallel conductors, the magnetic fields cause a force between the conductors, see https://en.wikipedia.org/wiki/Amp%C3%A8re%27s_force_law. But the currents flowing through small wires in a CPU are too small for movement.

\$\endgroup\$
  • \$\begingroup\$ "But the currents flowing through small wires in a CPU are too small for movement." - Are they though? Modern CPUs can pull something like 75A. Now that current is not running through a single bond wire, but still, several hundred mA through a single bond wire wouldn't surprise me. And they're small so they need less force to move. \$\endgroup\$ – marcelm Jun 10 '17 at 12:56
  • \$\begingroup\$ The bond wires are too short for a substantial force. \$\endgroup\$ – Uwe Jun 10 '17 at 13:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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