Chasing (Semiconductor) Rainbows

As a consumer, I find this constant trend towards miniaturization actually costs me money. Back in the days of discrete components, if I could not figure out how to repair a faulty device, I could probably find someone who could. Then we reached the level where board-level replacement was the most cost-effective solution, although that meant replacing a whole lot of components that were still good (and not generally salvageable). Now we are to the point where, if something fails to perform properly, we replace the entire device. Each generation represents an increase in economic impact on the consumer...

True, miniaturisation cost money. Your response reminds me of the reality of the dilema faced by most component level repair technicians/Engineers and hobbyists as they are now endangered species.

A member of this forum indicated in a recent tread that he's had to opt for another trade since he could no longer repair radios which are now coming in miniaturised circuit elements.

cheers,

ethobil

There are two principal driving forces behind the development; one is the heating effect and the need to dissipate the energy from what is an ever greater concentration of devices packed onto the chip, the other is the inability of the current mask technology to develop ever smaller and more precise boundaries to the circuitry using conventional methods.

In the case of the former lies the move to light as the medium rather than the conventional electron source. Opto-electro technology is with us already and the natural progression is for this to move into the processor arena. Heat dissapation is then no longer an issue and the density of devices is only limited by the accuracy and ability to miniaturise the individual transmissive elements. Hence the need to change the mask preparation process but with the use of X-rays rather than light to eradiate the substrates prior to etching, this gives a much clearer boundary layer and therefore a potential for greater density of devices.

Who pays? The end user , ultimately, but as the reliabilty of thr new device architecture will be greater there should be a trade off in terms of the performance/wasteage rate and therefore benefits for all concerned.

What new processes will be needed? A whole industry of them as completely new techniques will be called for as these ideas move from the lab to the large scale production plant.

How long before these devices become available..... possibly 20 years.It all depends how much stimulus is applied by those semi-conductor industry moguls prepared to launch the next era!

Massey.