How transistor really works, an alternate viewpoint | Hacker News Reader

How transistor really works, an alternate viewpoint

amasci.com

25 pointsdfox14 comments

On the one hand, I really like this article in general. It's a pretty clear explanation of how transistors work physically, without requiring too much knowledge of semiconductor physics. Though I don't think he introduced the idea of holes very well.

On the other hand: the whole article is written in a style that suggests this is some kind of major revelation to someone with a background in electrical engineering. I really find that hard to believe: is this really never explained in EE courses? I'm a physics major turned materials scientist, and this is really well-understood in both of those communities. HN EE's, did you all never get a course in solid state physics or electronic materials?

Which isn't an insult: I can understand that this kind of understanding might not be essential to circuit design, though I might think it would help. But if this is seriously not something they teach EE's, you're missing out. Because solid state physics is some seriously cool shit.

For an electrical engineer transistor is a black box 99.9% of the time. It's enogh to know it's interface.

I think it varies by program. I was a CS major, but at the school I went to, EE was actually not that far away from us. I tend to think of circuits as sort of ground-zero of most EE programs, and then some programs push more downwards in the abstraction hierarchy, covering solid-state physics and similar, while others (like ours) push more up the abstraction hierarchy, covering more CEish topics like chip design, pipelining, even up to embedded-systems programming (there was also a very strong focus on signal processing). Ideally I guess you'd want to know everything, from fundamental physics to CS, but a 4-year program sort of has to pick some subset.

Although borisk is correct that engineers mostly work at higher layer of abstraction, this absolutely was taught when I was at engineering school. What's more, there was a detailed discussion of transistor geometry, which is fairly critical but elided completely in this explanation, presumably because it's hard to draw ASCII art pictures of it.

I was in a CompE program a year ago. I didn't take semiconductors but I'm pretty sure it was required of the EEs at some point.

EE is a huge field. One of my recently-graduated friends is going to grad school to be an analog engineer, and this kind of thing is his bread and butter. Other EE's not necessarily.

This isn't a major revelation, at least to me (trained as a CompE with EE tendencies), but I will agree that I've never heard a really good, intuitive explanation of how a bipolar junction transistor works, despite modeling and working with them as current-controlled current amplifiers, as the author mentions. This especially obvious in contrast with MOSFETs, which are IMO easy to understand intuitively.

For an interesting experience, ask a group of EEs if bipolars are inherently current-controlled devices. This tends to create long-running flamewars (similar effect to asking pilots how airfoils create lift.) The ones who know that bipolars are current-controlled will nearly come to blows with the ones who know that they're voltage controlled. Very strange, but not unexpected when considering the poor jobs textbooks do with the non-math intuitive explanations. See sci.electronics.design for multiple examples of such fights, also wikipedia discussion history.

Interesting stuff, some more info: http://en.wikipedia.org/wiki/History_of_the_transistor

Not much of a difference - what does it matter where in the circuit you start tracing current? Why use analogies to water - I dont find it helpful.

Good to distingush between Current and Charge. I likewise like to distinguish between money and value - which is money over time.

-1, semantics crank. "Current does not flow."

Eh, I dunno. He did get really pedantic in that part, but you do run into a lot of circuit analysis that seems to forget that there are actual charged particles moving around in there. And that is really helpful to remember when you want to understand what's going on in detail.

I'm a big fan of physical insight, but you simply cannot redefine "current" as "that which does not flow and by implication is stationary".

Better would be to clarify current as charge in the state of motion. Then you can have your semantic cake and eat it too.

> cannot simply redefine "current"

Instead, authors of educational articles aimed at the public should stick to proper usage because it avoids misconceptions. "Flow of charge" is proper, "flow of current" is not. Unknown to most readers; I'm also using a further technique to reinforce certain concepts: always avoiding jargon (e.g. don't hide "charge flow" behind the word "current," instead don't define current at all. Remove the word entirely, and replace it with "charge flow.") Note that the intended audience here is little kids and grandmothers. In years of working with electricity misconceptions among the general public, I find that "flow of current" is one of the top learning barriers preventing easy grasp of electricity concepts. It connects to many other mistaken ideas.

Example: "batteries send out current," versus "batteries cause the conductors' charges to flow."

Reply on news.ycombinator.com