How can you “burn out at work”
Having dealt with processor cooling issues, we can now easily figure out the question of why Duron processors burn up if they are turned on without a heatsink.
Generally speaking, the question is somewhat strange. For a long time, everyone has been searching for effective radiators, and some still have the idea to try – or maybe the processor will work without a radiator?
Nevertheless, such a question exists. Apparently, the first “faced collectors” faced such a problem.Saving time, they quite often turn on the computer even before the assembly is completed. While the computer turns on, such a collector manages to install a heatsink on the processor and tighten all the screws. So far, this has not led to disastrous consequences. The processors withstood short-term overheating quietly. Moreover, even if the collector hesitated, did not have time to install the radiator, the processor, as a rule, did not burn, but simply “hung”. Only had to reboot. And now all of a sudden this “number” does not work. It’s a shame.
But let’s understand, why, in fact, before the processors did not burn?
They really didn’t burn. True, no one ever gave a guarantee of performance under these conditions. The processors did not burn, not because some measures were taken, but because of the imperfection of their production technology. We have already figured out that the main part of the energy is spent at the moments of switching the transistors that make up the processor. At these times, each transistor heats up. If at least one of them heats above the permissible temperature, then it will burn, and the processor as a whole will find its rightful place on the New Year tree. However, due to the imperfection of production technology, transistors quite often began to “junk” long before the temperature reached a critical value. If at least a few transistors out of millions of the same ones in the processor “get stuck” and stop switching, then the processor will stop. The remaining transistors will stop switching. But in this mode, almost no heat is generated! Without additional development costs, processor manufacturers “got a new quality.”
To them (and more often to ours), unfortunately, now this may not work. Firstly, the “packing” density of transistors on a chip has significantly increased. Secondly, the operating switching frequencies of these transistors have long been in the microwave region. Therefore, the temperature in individual regions of the crystal rises very quickly and quickly reaches a critical value. In addition, a crystal without a radiator heats up unevenly. All this leaves the processor (at least the Duron processor) no chance of surviving without a radiator. Logically, the Athlon processor should burn just as easily, but none of my friends have tried this. The Pentium III processor, apparently, has certain chances to survive, but for some reason it is not included without a heatsink. Although this is strange. The experiment is simple: take a dozen or three processors of various types and turn them on without radiators. Reliable fault tolerance statistics are ready. Maybe someone will undertake such an experiment? Only, please, at your own expense.
Do processor manufacturers know about our misfortune? It seems that they not only know, but also take certain measures. Look at the photos of modern processors.