Autore: ExtremeTech

A standard computer is a complex group of individual parts working together as a whole — RAM, some kind of data storage, a processor, and so on. When one of those integral parts breaks, the computer is rendered useless and the part must be replaced, but what if the computer could begin routing the broken part’s tasks through the bits that are still functional? Computers can’t do that just yet, but researchers have now managed to coax a microchip into doing so.

Ali Hajimiri and a group of Caltech researchers have managed to create an integrated circuit that, after taking severe damage, can reconfigure itself in such a way where it can still remain functional. The chip isn’t exactly like Terminator 2′s T-1000. It doesn’t slowly melt back together, or generate new physical parts to patch up the broken area. However, it contains a secondary processor that jumps into action when parts of the chip become compromised. So, the bulk of the chip is damaged, but the secondary processor uses a bit of quick-thinking to figure out how the chip can still perform tasks. The chip is also able to tweak itself on the fly, and can be programmed to focus more on saving energy or performance speed.

Healing chip

The team’s microchip lives its life as a power amplifier, the kind of circuit that processes signal transmissions, such as those found in mobile phones. Aside from the aforementioned secondary processor, the chip contains 100,000 transistors, as well as various sensors in order to monitor the unit’s overall health. The team tested the self-healing capabilities of the chip by blasting it with a laser, taking out around half of its transistors. It only took the microchip a handful of milliseconds to deal with the loss and move on. On top of that, the team found that a chip that wasn’t blasted by lasers was able to increase its efficiency by reducing its power consumption by half. Now like you’re undoubtedly thinking, if the secondary processor is smashed to pieces, or all of the transistors go on the fritz, the chip can’t fix itself. It won’t have any remaining parts to actually push processes through. If smashed into a fine dust with a hammer, the chip won’t be able to put itself back together. However, the chip would mainly be able to withstand the kind of damage that it would naturally receive, such as wear and tear from age, or a slight bump on the noggin.

Though the chip can’t melt into metal puddles and slide back together while John Connor stares in horror, and can only fix itself if the secondary processor and at least some of the parts remain intact, the self-monitoring and tweaking capabilities are a monumentally important addition to the unit. Not only can the chip monitor itself in order to provide the best possible performance in general, but if some of the parts do break down, the chip won’t provide some kind of jury-rigged output. It will look for and tweak itself to generate the best possible output with the remaining working bits. So, rather than a slow chip that is technically operational, you can rest easy knowing that your chip is doing the best it can, and nothing is being wasted.

According to Hajimiri, the technology behind this self-healing circuit can be applied to any kind of circuit, as the secondary processor is tucked away safely underneath the main unit. Though the system is still simple (in that it uses more hardware that could conceivably get damaged as well), it does point toward embedded repair systems for delicate tech in the future. Perhaps one day when we drop our phones, we’ll be able to pick them back up off the ground without then spending the next twenty minutes carefully testing integral functions to make sure everything is still running as efficiently as it could.

Now read: Self-healing, self-heating flash memory survives more than 100 million cycles