Boffins fabricate the 'most complex bendy microprocessor yet' A team of engineers from the Vienna University of Technology in Austria has created what they claim is the most complex flat and flexible microprocessor to date – using a molybdenum disulfide semiconductor. In other words, it's a demonstration of a potentially viable alternative to silicon. The keyword here is alternative, as …
I've programmed complex industrial machinery using an 8-bit (Z80) processor with 64KB memory. You could do an awful lot of useful stuff with these as long as you don't want all the flashy features and the complex displays. In fact you could offload the display to your phone or your watch.
>>In fact you could offload the display to your phone or your watch.
...which is how IoT will eventually come of age: 1) a tiny local (bendy?) processor, low electrical and compute power, but with some HW crypto blocks and sensor interfacing, loosely coupled to 2) an arbitrarily complex display/GUI device - smart phone, lappy etc.
Until then, we can all sit back and enjoy the complete f'kwit entertainment show :0
"in return for the US adding the second i in aluminium."
Years ago I read an article in The Chemical Engineer which claimed that the American spelling was due to a typo in an early brochure promoting the metal, which got perpetuated. Sorry, can't find a link right now.
I remember reading somewhere else that a new substance - I think it was molybdenum disulfide - was a potential alternative to graphene as the successor to silicon. So there may be potential to get fabrication of circuits working well enough to permit useful microprocessors to be built in this fashion soon enough!
Ah, yes, it was molybdenum disulfide, and it allowed smaller transistors (1 nm as against 5 nm for a critical feature) and operation at higher temperatures (220 degrees C) compared to silicon, and it even compared favorably to graphene in some circuit parameters. There was another thing that was preventing usable transistors from being made, but that was overcome.
A 486 has pretty much spot on 1 million transistors, using the quoted Intel figures for current/next generation densities that should fit into 0.1mm^2, or 100 to the square millimetre. *
That gives us a Doom capable t-shirt (flexible OLED display) with all the electronics lost in one of the seams. For the applications being touted, a z-80 (packing density 10k/sq. mm) would be overkill.
We don't need flexible electronics, we need a flexible interface for electronics of a size that doesn't even rate skidmark on the fly-dirt scale.
Yes MoS2 (or "Moly" as some engineers refer to it) is often used as a solid lubricant so there should be plenty of materials prop data around for (especially at highish temperatures like 200c, much better than Silicon). Plus it's (relatively) non toxic (compared to Gallium Arsenide, which contains err Arsenic).
The amazing bit is not the bit ALU microprocessor.
IBM built a "cheap" computer in the 1960's using a 1 bit ECL ALU with registers to give the actual word length. Likewise when Ferranti built their 16 bit microprocessor in the 70's in TTL the chip had a 10 000 gate allowance. Once again a 1 bit ALU with I/O registers made the impossible possible.
But a microprocessor core in 115 transistors!!! The downside is the mobility at 3 cm^-2 v^-1s^-1 (compared to 1000+ for Si). That suggests (like all previous "bendy" circuit plans) the clock frequency will be low, RF comms will be very difficult and you'll still need a chunk of Quartz for the clock (unless you go to asynchronous or clockless designs).
BTW this was done on a Silicon wafer. They say it makes no difference and the work could be done on glass or plastic substrates IE plastic film as the processes are mild.
The 2 biggest issues with the whole roll-to-roll electronics concept are the low clock speed (due to poor material properties) and the high line widths that truly high volume (printing press) type systems can achieve. IIRC that's 10micrometres. Historically there has been little pressure to narrow it but that's what you'll need to raise those clock frequencies. Again it's not the processing, it's the RF comms. Until you fix that a mobile phone you can print remains a comic book fantasy.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
