An open community 
of Macintosh users,
for Macintosh users.

FineTunedMac Dashboard widget now available! Download Here

Topic Options
#20741 - 02/20/12 08:29 AM Single-atom transistor
grelber Offline


Registered: 08/05/09
Loc: North of 49th ||

Top
#20742 - 02/20/12 09:30 AM Re: Single-atom transistor [Re: grelber]
Hal Itosis Offline


Registered: 09/03/09
Loc: 10.6.8 (build 10K549)
Has the word "transistor" somehow become limited to meaning a discrete switch... whether basic single-bit binary (on/off), or multi-bit (several states) ???

Because i remember when transistors were used for analog signals, in order to control higher power circuitry... i.e., an amplifier. wink

edit/ they should call a switch a switch.


Edited by Hal Itosis (02/20/12 09:33 AM)

Top
#20751 - 02/21/12 12:17 AM Re: Single-atom transistor [Re: Hal Itosis]
grelber Offline


Registered: 08/05/09
Loc: North of 49th ||
Originally Posted By: Hal Itosis
Has the word "transistor" somehow become limited to meaning a discrete switch... whether basic single-bit binary (on/off), or multi-bit (several states) ???
Because i remember when transistors were used for analog signals, in order to control higher power circuitry... i.e., an amplifier. wink
edit/ they should call a switch a switch.

According to numerous sources, transistor is defined as:
A semiconductor device with three connections, capable of amplification in addition to rectification.
Transistors are commonly used as electronic switches, both for high-power applications such as switched-mode power supplies and for low-power applications such as logic gates.
It has always been considered a switch: see http://www.techterms.com/definition/transistor

Top
#20754 - 02/21/12 02:07 AM Re: Single-atom transistor [Re: Hal Itosis]
tacit Offline


Registered: 08/03/09
Loc: Portland, Oregon, USA
Yep, a transistor has always been used as both a switch and an analog amplifier. The transistors used in digital circuitry have a very fast response and a discrete, rather than linear, switching curve.
_________________________
Photo gallery, all about me, and more: www.xeromag.com/franklin.html

Top
#20767 - 02/21/12 07:16 AM Re: Single-atom transistor [Re: grelber]
Hal Itosis Offline


Registered: 09/03/09
Loc: 10.6.8 (build 10K549)
Originally Posted By: grelber
According to numerous sources, transistor is defined as:
A semiconductor device with three connections, capable of amplification in addition to rectification.

"capable of amplification"

And are these single-atom jobbers capable of amplification? (in the linear/analog sense).

I think (i'm guessing) that they are switch-only... and their application will be strictly digital.

But it's not important. I would just prefer a different semantic.

Top
#20783 - 02/22/12 03:06 AM Re: Single-atom transistor [Re: Hal Itosis]
tacit Offline


Registered: 08/03/09
Loc: Portland, Oregon, USA
My understanding is that the single-atom transistors is either on or off, no in-between; it can't be used for analog signal amplification.

Though it may be that conventional analog transistors work the same way. An analog transistor can, if I remember the physics rightly (and I might not be; apologies if I get this wrong) you can think of an analog transistor as a collection of billions of digital transistors. Normally, there are junctions between parts of the semiconductor which are doped to be negatively charged (having an excess of electrons) and parts doped to be positively charged (having an excess of "holes" into which an electron can fit). Electrons can't move across the junction and the transistor is off.

When you apply a small current to the transistor's base, some of the holes are filled, and there is a path for a small number of electrons to move across the junction, so it acts like it has a high resistance. Apply a larger current to the base and more of the holes are filled, so more electrons can move, and the resistance goes down. But each "hole" is either filled or not filled, and when filled provides a path for electron movement, so each individual "hole" acts like a transistor that's either completely on or completely off. The more of these you switch on, the more current can flow, and the lower the transistor's effective resistance becomes.

It's been quite a while since my last physics class, so I might be wrong on the details, but I believe that's basically how analog transistors work.

In any event, the semantic is the same; the word "transistor" literally means a resistor whose resistance can be changed by the addition of a signaling current. The difference between an analog and a switching transistor is simply in the response curve of the resistance, that's all--a switching transistor has a steep response curve and an analog transistor has a shallow response curve.
_________________________
Photo gallery, all about me, and more: www.xeromag.com/franklin.html

Top
#20789 - 02/22/12 09:01 AM Re: Single-atom transistor [Re: tacit]
Virtual1 Offline


Registered: 08/04/09
Loc: Iowa
Originally Posted By: tacit
My understanding is that the single-atom transistors is either on or off, no in-between; it can't be used for analog signal amplification.


Amplification does not require linearity in the output. (that's why some amplifiers are referred to specifically as "linear amplifiers", because those are linear) A relay is an excellent example of an amplifier with a non-linear/digital output. When a small current is passed across its coil, the armature moves and closes a contact that can pass a large amount of current. This is current amplification, where once a small threshold of current input is reached, results in a large change in output current. (from none to some fixed large amount) Basic computers like ENIAC were made with relays and also tubes that were also working in analog (not linear) mode.

This is important in computing because you need two basic functions for computation:
1) a single input must be able to trigger more than one output in the same manner (a different interpretation of "amplification", ten people all chanting your words together instead of one person yelling them out much louder)
2) there must be a way to compare ("mix") at least two inputs to arrive at a variable output. one example is an "and gate", that produces a 1 when BOTH inputs are 1, otherwise producing 0.

neither of these require linear amplification. thus transistors used in computing do not require linear amplification.

But then I suppose I'm straying a little off-topic.
_________________________
I work for the Department of Redundancy Department

Top
#20794 - 02/22/12 09:53 AM Re: Single-atom transistor [Re: Virtual1]
Hal Itosis Offline


Registered: 09/03/09
Loc: 10.6.8 (build 10K549)
Actually, your relay-as-an-amplifier example is great. Point well taken.

Given the minute scale however, i will probably still think of this new technology in terms of a "single atom register" rather than transistor (a term which i will always associate with a fundamentally linear device, that can also be configured to become a switch, by employing positive feedback).

Top
#20795 - 02/22/12 10:11 AM Re: Single-atom transistor [Re: tacit]
Hal Itosis Offline


Registered: 09/03/09
Loc: 10.6.8 (build 10K549)
Originally Posted By: tacit
My understanding is that the single-atom transistors is either on or off, no in-between; it can't be used for analog signal amplification.

Though it may be that conventional analog transistors work the same way.

True enough. When it come to finding the "line" between analog and digital, we essentially wind up at a point where there is no perfectly linear/analog world. Go deep enough and it all boils down to discrete levels of something, and we're back into quanta (how much). Planck's constant is the LSB (least significant bit) of action in the universe. And my vague memory of studying transistors recalls that Boltzmann's constant was mentioned quite often.

So yes, perfect linearity does not exist. But for human perception it doesn't need to be perfect. Our hearing and seeing are also based on sampling (and limited bandwidths), as our brains process the information. E.g., 30 frames per second or above appears linear... and i think the decibel scale was based on what our ears can detect. I.e., changes in level less than one decibel go almost unnoticed.



Edited by Hal Itosis (02/22/12 10:21 AM)

Top

Moderator:  alternaut, cyn