How many transistors in a ...

And to Rick C's point, you can divide the current between capacitors in any ratio you like.

Cheers

Phil Hobbs

I was thinking along the lines that those that were doing "op-ampish" things were linear, but if there's any that are on or off hard they'd count as "digital".

AAAAAAAAAAAAAAAAAGH!

Is that a satisfactory answer? I think if they're inside a chip I'll just count them as one, for sanity's sake.

For a standard flyback power supply:

In on/off mode: 1 power FET Linear mode: none

If you include the SMPS controller the number is hard to say

The same goes for the feedback regulation performed by the TL431 (just check the data sheet for the internal schematics of that part)

The same story for the half bridge design for a LCD backlight. Here you have 2 FETs, but again the addition of controller IC

Some newer micro controller is actually running in border linear mode to reduce power consumption. I forgot the name of the company. Sub threshold technology

Cheers

Klaus

I thought of that, in trying to figure out where ECL fit into the mix. Because, on the one hand, it's logic. But on the other, the transistors never saturate.

"Would work fine if the thing actually saturated" works for me, except I think there's 4-level logic inside of at least some FLASH, which complicates things.

Um, time to display (!) my ignorance again. How many transistors in a TFT monitor like the one in my laptop I'm looking at right now? Are they operating linear or digital?

Mark L. Fergerson

And yet we can count electrons flowing.

But you can't have 1.5 electrons flowing in a condutor. As the electrons move they are propelled one at a time across whatever medium you choose they are integer charges.

The only way to make current flow a continuous variable is for time to be continuous and we aren't sure it is.

I was trying to count transistors whether they are discrete or buried in ICs. Just counting the discrete transistors isn't what I had in mind at all.

Huh? Of course you can. It's a tiny current, but you can have any fraction you want. What you can't do is have 1.5 e's move over a barrier like in a pn diode... there you have shot noise. In a current from a battery and resistor there is no shot noise.

(It's a subtle point and the exact details are involved.)

George H. As the

Ok, please show me a circuit that transfers 1.5 electrons from one place to another.

But only if they are in a digital circuit? You said to count all the transistors on a chip as one.

You are confusing current and charge. You can't have fractional charges, true. Current is *moving* charge. You can move charge arbitrarily slowly. You *don't* get a tick out of your measuring device every time the amount of charge moved happens to coincide with an integer times q0. Current in metal conductors is *not* individual electrons moving one-by-one, but rather a large number of them moving collectively in a common direction on average.

Read up on "shot noise".

Jeroen Belleman

If the charges are moving by they are moving in discrete quantities. I don't care that your equipment is not capable of measuring that. But current is inherently a digital measurement.

You're visualizing the situation incorrectly. Current density is charge density times velocity. Velocity is not quantized, so neither is current.

Cheers

Phil Hobbs

On 18/02/17 00:00, rickman wrote: [snip!]

Aargh!

Jeroen Belleman

Just a question, what is the point of knowing the answer to your question?

I know you always have good grasp of most you are talking about, but this one reminds me of some PhD dinner conversation and nothing that has any remotely practical impact

Cheers

Klaus

Electrons are discrete only when you want to see them discrete. Once you are down to single electrons you should use quantum mechanics and wave function is continuous. Due to continuous nature of wave function at least theoreticaly quantum computer can do more than digital computer.

Concerning "look closely enough": observation disturbs quantum system, so when you look at it it may malfunction and behave in digital way. When you stop looking it will work in continuous way.

You don't have to invoke quantum mechanics in order to count electrons. Check out Millikan's oil drop experiment.

No, it's not about equipment. It's about quantum mechanics; in a metal, there's a SEA of electrons, you cannot localize one. Think of an atom: the electrons are in orbitals, probably at some distance from the center, but never at one spot. In a conduction band of a metal, the orbitals are the size of the metal object. They're not at one spot, so you can't count how many pass through your meter.