How Transistors Work

Status
Not open for further replies.

LvW

Apr 12, 2014
604
Joined
Apr 12, 2014
Messages
604
Oh - I forget to mention that the main question was "How transistors work" and not "What is the most simple method for designing a BJT stage".

PS: Sorry, but I have to stop the conversation at this point because of a jouney starting tomorrow morning.
 

chopnhack

Apr 28, 2014
1,576
Joined
Apr 28, 2014
Messages
1,576
Oh - I forget to mention that the main question was "How transistors work" and not "What is the most simple method for designing a BJT stage".

PS: Sorry, but I have to stop the conversation at this point because of a jouney starting tomorrow morning.

Safe travels, I look forward to your continued input - all of your inputs! It's certainly a bit on the deep end for me, but I love to learn and perhaps in time all of it will make sense. For now, the equations are still foreign.
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
It will soon became clear young skywalker. In about 10 years or so, dont give up its worth it.
Adam
 

chopnhack

Apr 28, 2014
1,576
Joined
Apr 28, 2014
Messages
1,576
It will soon became clear young skywalker. In about 10 years or so, dont give up its worth it.
Adam
I have no intention :D I enjoy electronics and have since I was young, now that I have my own, I think its important that they get involved as well. If it takes me 10 years, that will be fine, they will be teenagers by then, but I should think along the way they will get a darn good education!
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
Good I will help you all I can. It is nice to have someone who doesnt think they will learn everything in five minutes. But if you have any burning questions you are really struggling with then message me if you dont want to post it I know some simple questions can seem embarising to ask. I was once in your shoes and I know how confusing it can get. If I dont know the answer I know someone that will.
Adam
 

chopnhack

Apr 28, 2014
1,576
Joined
Apr 28, 2014
Messages
1,576
Good I will help you all I can. It is nice to have someone who doesnt think they will learn everything in five minutes. But if you have any burning questions you are really struggling with then message me if you dont want to post it I know some simple questions can seem embarising to ask. I was once in your shoes and I know how confusing it can get. If I dont know the answer I know someone that will.
Adam
I really appreciate that Adam :D I didn't realize how many great resources for learning electronics there are on the web, including this site with its great members. I will do my due diligence first before asking, but expect many questions!
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Safe travels, I look forward to your continued input - all of your inputs! It's certainly a bit on the deep end for me, but I love to learn and perhaps in time all of it will make sense. For now, the equations are still foreign.

Let's do it without equations then :) And since you've seen the voltage vs current argument it seems fair to explain transistors from both perspectives.

You're probably familiar with diodes.
  1. The naive explanation is that they allow current to flow in one direction and not in the other.
  2. A better description tells you that there is a small voltage drop when current flows in one direction, and a far greater one on the other direction -- often larger than the voltages you have present in your system.
  3. An even better description tells you that there is a function which approximates very closely the current that will flow if a certain forward voltage is applied across the diode.
Note that in the last case we swap from talking about things being about the current to being about the voltage. (2) tells you that there is an approx 0.65V voltage drop. (3) explains how that forward voltage is actually related to the current that will flow through the diode. It's not a simple step.

When we look at transistors, there is a similar breakdown into three explanations of the base-emitter behaviour (note that a transistor is far more complex, so this just a description of a single small part of transistor behaviour:
  1. The base current controls the collector current
  2. After the base is more than about 0.6V more positive than the emitter (for NPN) the base current increases, and this is correlated with an increase in collector current until saturation is achieved
  3. The voltage across the base/emitter junction controls the collector current. A linear increase in base voltage results in an exponential increase in collector current (again until saturation)
For a deep understanding of transistors you need to understand the descriptions to at lease the third level of abstraction (I've broken it down somewhat arbitrarily, you won't find people talking about "the third level of abstraction"). If you want to understand more, you need to look at the physics of the device. Each level of abstraction provides an explanation for the former and introduces more that you have to accept until you go deeper again.

The question we've been debating is which level of understanding is appropriate. For an EE you probably need at least the third given here, but you will also likely be introduced later to the physics which will explain this level -- it is an understanding that is perhaps the minimum to understand the actual design of the device (a transistor) itself.

My opinion differs from LvW (probably only very slightly) in that I suggest that for someone who wants to solve a practical problem now (and typically a switching problem) that an understanding at the second level above is often sufficient. I believe this because in the absence of some actual understanding, the explanation given at (3) invites people to apply a voltage to the base of a transistor.

If a person were to go forward and actually design more complex circuits, they may run into the limitations of this approach and an understanding of (3) might be worthwhile. This may become an issue if they have a voltage source of around 0.75V and need to turn on a transistor. 0.75 is way higher than 0.65V, right? Well (like the issue with voltage regulators and dropout voltages vs current) we now need to consider the rise in Vbe that is correlated with a certain base current.

LvW and I perhaps disagree on what voltage vs current control mean, but both of us agree (I think) that in many cases we will use in our calculations some form of β=Ic/Ib (where β defines the ratio of collector to base current under some range of conditions). After that, it is typical to determine some resistance values which will have voltages across them is the calculated Ic, Ib, or Ie (depending on the topology). The calculations will then typically use voltages such as Vbe, Ve, and Vc along with Ib (but the latter typically defined in terms of voltages and resistances) in various calculations to model the behaviour of the circuit. I call this current control, LvW -- if I'm right -- calls this voltage control.

So how much understanding do you want? :)
 

Ratch

Mar 10, 2013
1,102
Joined
Mar 10, 2013
Messages
1,102
As I try to learn more about these components, I stumbled onto this site. Can anyone review the information and let me know if this is a valid viewpoint or if its bunk. The author did state that it is his opinion as I recall, but so far reading it, it was making sense...
http://amasci.com/amateur/transis.html

That site is heresy at Physics Forms, and the moderators there are scared that some newbie will read it and deviate from the party line dogma of current control. Therefore, they suppress any reference to Beaty's site. I am familar with the site, and I believe it is mostly correct. About the only truly current controlled components are magnetic amplifiers and gas discharge tubes. BJTs, FETS, and vacuum tubes are all voltage controlled current and therefore transconductance devices. The physics of the device determine what controls the device. The base current of a BJT is an indicator of the collector current, not the cause of it. If I notice that the sun rises when my bedside clock reads 5:30 AM, do I conclude that my clock is controlling the sunrise? If you have any questions about what Bill Beaty's site avers, give me a link, chapter, verse and quote, and I will give you my perspective on the subject.

Ratch
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Therefore, they suppress any reference to Beaty's site.

I don't think it should be suppressed. But I don't think it's perfect either. I think that some concepts are presented in an odd manner.

I am familar with the site, and I believe it is mostly correct.

I'd probably go along with that.

About the only truly current controlled components are magnetic amplifiers and gas discharge tubes. BJTs, FETS, and vacuum tubes are all voltage controlled current and therefore transconductance devices. The physics of the device determine what controls the device.

and that.

The base current of a BJT is an indicator of the collector current, not the cause of it.

But not that.

I'd say that the base current is correlated with the collector current in a large region somewhere between cutoff and saturation.

I would say that for certain cases, the correlation is such that we can use it as if it is the cause rather than arguing if it is the cause or the effect because it is mathematically easier that way and works well enough. (note that I simply argue that it is a useful and oft-used shortcut, not a dogmatic principle)

(It's the word "indicator" that I have issues with. Mostly because in saturation it's not, but also because we can do things to affect the base current and when that changes we will likely see a change in collector current)

If I notice that the sun rises when my bedside clock reads 5:30 AM, do I conclude that my clock is controlling the sunrise?

No, but you might notice that there is some correlation, so you can use this magical device to predict when sunrise will be. As if by magic, this simple number (time) seems to be able to be used in calculations which determine the hours of daylight and night. Indeed some people carry such devices on their person at all times for this exact reason!

Indeed, some people assume time is the controlling factor when they say "I'll meet you at noon". What they are saying is that they will meet you when the sun is at the highest point in the sky, but they will tell you you're incorrect if you assume they mean a time other than 12:00 on their watches, eve though the sun may not be at the highest point at this time!

Sadly, some people use this abstract concept of time rather than the relative position of the sun with respect to their spatial location. Indeed, people begin to think that clocks rule their life, not the sun.

However, in one major difference between clocks and base current, winding our clocks forward does not cause the sun to move across the sky, whereas taking action to increase the base current may change the collector current -- thus my preference for correlation rather than indication. And also the realization that clocks actually indicate the time

If you have any questions about what Bill Beaty's site avers, give me a link, chapter, verse and quote, and I will give you my perspective on the subject.

I would be interested in whether you think his description of "uncharged charge" could be written better.

Also note that the charge within wires is ...uncharged. Every movable electron has a positive proton nearby, so even though the metal contains a vast sea of charge, there is no net charge on average. Wires contain "uncharged" charge. Better call it "cancelled charge." Yet even though the electrons are cancelled by the nearby protons, the electrons can still flow among the protons. Cancelled charge can still move around, so it's possible to have flows of charge in an uncharged metal.

I would also like your interpretation of a wire passing through a magnetic field. Do the electrons experience a force which causes them to move, which creates a potential difference, or does the movement through the field cause a potential difference which applies force to the electrons? (but this is an aside -- it may be best to stick to the original topic).

Back to that page in question, I think it suffers (and only in small parts) of sloppy writing which may be a distraction. Unfortunately, such distractions (if they're not recognised) could lead people astray in other areas. I wouldn't recommend it as a place to learn, but more as a place to challenge your understanding.
 

Ratch

Mar 10, 2013
1,102
Joined
Mar 10, 2013
Messages
1,102
*steve*,

I'd say that the base current is correlated with the collector current in a large region somewhere between cutoff and saturation.

Yes, Ic is exponentially related to Vbe. So is Ib. Therefore, Ic and Ib are directly related to each other (beta). However, you cannot explain how Ib influences Ic, because it does not. But, both Ic and Ib can be explained in terms of Vbe. You can model a BJT in terms of current control using beta, but a model only tells you what a device will do, not how it works.

I would say that for certain cases, the correlation is such that we can use it as if it is the cause rather than arguing if it is the cause or the effect because it is mathematically easier that way and works well enough. (note that I simply argue that it is a useful and oft-used shortcut, not a dogmatic principle)
(It's the word "indicator" that I have issues with.

The physics of a BJT do not allow for Ib to be used as a cause of Ic. In the active region, there is a more or less direct correlation (beta) between Ib and Ic. But still, Ib does not control Ic.

Mostly because in saturation it's not, but also because we can do things to affect the base current and when that changes we will likely see a change in collector current)

You can also do things in the active region that affect the base current, and see a change in the collector current. But still, that does not mean Ib affects Ic.

No, but you might notice that there is some correlation, so you can use this magical device to predict when sunrise will be. As if by magic, this simple number (time) seems to be able to be used in calculations which determine the hours of daylight and night. Indeed some people carry such devices on their person at all times for this exact reason!

I dislike using the word "magic" when talking about physical principles. The time "number" does correspond to a measured physical event.

However, in one major difference between clocks and base current, winding our clocks forward does not cause the sun to move across the sky, whereas taking action to increase the base current may change the collector current -- thus my preference for correlation rather than indication.

And the action to increase Ib is to increase Vbe, which increases both Ib and Ic. If Ib and Ic have a correlation with each other, then either can be thought of as an indication of the other.

I would be interested in whether you think his description of "uncharged charge" could be written better.
Also note that the charge within wires is ...uncharged. Every movable electron has a positive proton nearby, so even though the metal contains a vast sea of charge, there is no net charge on average. Wires contain "uncharged" charge. Better call it "cancelled charge." Yet even though the electrons are cancelled by the nearby protons, the electrons can still flow among the protons. Cancelled charge can still move around, so it's possible to have flows of charge in an uncharged metal.

Yes, although he is not wrong, his description is not very concise and sometimes rather confusing. I wonder what he means by "uncharged metal".

I would also like your interpretation of a wire passing through a magnetic field. Do the electrons experience a force which causes them to move, which creates a potential difference, or does the movement through the field cause a potential difference which applies force to the electrons? (but this is an aside -- it may be best to stick to the original topic).

It is well known that a electrons moving through a magnetic flux will experience a force perpendicular to the direction of the magnetic flux and the relative movement of the flux. This force will cause the electrons to concentrate at one end of the conductor. Anything that causes electrons to gather in one place in a concentration larger than another place will cause a voltage difference between the two locations.

Back to that page in question, I think it suffers (and only in small parts) of sloppy writing which may be a distraction. Unfortunately, such distractions (if they're not recognised) could lead people astray in other areas. I wouldn't recommend it as a place to learn, but more as a place to challenge your understanding.

Yes, it is rough in a few spots. But, he brings up points that are not covered in conventional material or instead just glossed over, without explaining how important those points are.

Ratch
 
Last edited:

Laplace

Apr 4, 2010
1,252
Joined
Apr 4, 2010
Messages
1,252
...And the action to increase Ib is to increase Vbe, which increases both Ib and Ic. If Ib and Ic have a correlation with each other, then either can be thought of as an indication of the other. ...
This is exactly why your reasoning fails -- in spectacular implosion. A circuit designer will force a certain amount of base current in order to achieve a certain amount of collector current. Ib and Ic are not an indication of each other. Forcing a change in Ib will cause a corresponding change in Ic, but it does not work in reverse. It is a less than useless process to think of setting the base-emitter junction voltage to adjust the width of the depletion region which has a complex but predictable relation to the collector current, especially when ultimately the collector current is proportional to the base current over the range of interest. There is no reason at all to consider the base voltage except in gross order-of-magnitude terms and with the realization that it won't change much over the range of interest.

If one seems to be less than concerned with the internal physics of the transistor, it is only because one doesn't design anything with the transistor but rather uses the h-parameter model of a two port active network, i.e. where the collector current is controlled by the base current. Furthermore, when manufacturers present characteristic curves in their transistor datasheets the controlling input is always the base current. Giving any more than a cursory level of attention to the width of the depletion region versus junction voltage is both pedantic and useless.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
It is well known that a electrons moving through a magnetic flux will experience a force perpendicular to the direction of the magnetic flux and the relative movement of the flux. This force will cause the electrons to concentrate at one end of the conductor. Anything that causes electrons to gather in one place in a concentration larger than another place will cause a voltage difference between the two locations.

So the force causes electrons to move. This causes a movement of charge (current), which results in a potential difference.

I guess the current does cause a voltage. :D
 

Ratch

Mar 10, 2013
1,102
Joined
Mar 10, 2013
Messages
1,102
This is exactly why your reasoning fails -- in spectacular implosion. A circuit designer will force a certain amount of base current in order to achieve a certain amount of collector current. Ib and Ic are not an indication of each other. Forcing a change in Ib will cause a corresponding change in Ic, but it does not work in reverse. It is a less than useless process to think of setting the base-emitter junction voltage to adjust the width of the depletion region which has a complex but predictable relation to the collector current, especially when ultimately the collector current is proportional to the base current over the range of interest. There is no reason at all to consider the base voltage except in gross order-of-magnitude terms and with the realization that it won't change much over the range of interest.

If one seems to be less than concerned with the internal physics of the transistor, it is only because one doesn't design anything with the transistor but rather uses the h-parameter model of a two port active network, i.e. where the collector current is controlled by the base current. Furthermore, when manufacturers present characteristic curves in their transistor datasheets the controlling input is always the base current. Giving any more than a cursory level of attention to the width of the depletion region versus junction voltage is both pedantic and useless.

I do believe your are confusing methods of circuit design with the physics of BJT operation. I have never advocated getting involved with the internals of the BJT and using the highly nonlinear relationship of Vbe to Ib and Ic to design circuits. Of course you use models to expedite design. I do aver, however, that models tell you what a component will do, but not how the component works to accomplish its behavior. Pointing to a model and saying that a BJT works that way is fallacious. It is the physics of the transistor that describes how a transistor works the way it does. A designer will use external components to minimize BJT parameter variations and set an operating point. Models help the designer do that. In conclusion, describing how a BJT works does not mean advocacy for a design method.

Ratch
 

Ratch

Mar 10, 2013
1,102
Joined
Mar 10, 2013
Messages
1,102
So the force causes electrons to move. This causes a movement of charge (current), which results in a potential difference.

I guess the current does cause a voltage. :D

The energy to concentrate a charge divided by the charge itself is the definition of voltage (joules/coulomb). In a capacitor, a voltage can be present without any current. In an open wire where no net current is present, a magnetic field can increase the energy difference of the electrons between each end of the wire, thereby producing a voltage even if no current is present.

Ratch
 

Laplace

Apr 4, 2010
1,252
Joined
Apr 4, 2010
Messages
1,252
I do believe your are confusing methods of circuit design with the physics of BJT operation. ...

I believe I was rather clear on the difference between the h-parameters of a 2-port active network versus the physics of BJT operation. One is useful while the other deserves a cursory level of attention, particularly in an electronics forum.
 

BobK

Jan 5, 2010
7,682
Joined
Jan 5, 2010
Messages
7,682
The base current of a BJT is an indicator of the collector current, not the cause of it.
Ratch
Exactly the same thing can be said about the base-emitter voltage. You can have a Vbe of 1V which would allow a large current to flow, but if there is a collector resistor it might limit the collector current to a much lower value.

Bob
 

Ratch

Mar 10, 2013
1,102
Joined
Mar 10, 2013
Messages
1,102
I believe I was rather clear on the difference between the h-parameters of a 2-port active network versus the physics of BJT operation. One is useful while the other deserves a cursory level of attention, particularly in an electronics forum.

Both studies are useful in any electronics forum. Understanding how something works can be profitable, too. I think you were too dismissive of that fact. The h-parameter model is not the only model used in design. There are voltage models that can be used, too.

Ratch
 

Ratch

Mar 10, 2013
1,102
Joined
Mar 10, 2013
Messages
1,102
Exactly the same thing can be said about the base-emitter voltage. You can have a Vbe of 1V which would allow a large current to flow, but if there is a collector resistor it might limit the collector current to a much lower value.

Bob

The Vbe is the cause of Ic in the active region. A Vbe of 1 volt will probably cause saturation, which is not operation in the active region.

Ratch
 

BobK

Jan 5, 2010
7,682
Joined
Jan 5, 2010
Messages
7,682
If B is a function of A and A is the cause of something, one can pretend that B is the cause and you will not get any incorrect result. So it really doesn't matter. The only question is which is easier to use in practice, the relation of Ib to Ic or the relation of Vbe to Ic.


Bob
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
(edited for non-readers)

I fear that this thread is not going to help the OP as it degenerates into a discussion of how many angels can dance on the head of a pin.

If you wonder why your post has been deleted or moved, read the section in red below.

I propose that from this point onward that we cease discussion of whether it's voltage, current. or pixies that do the real work inside a transistor and concentrate on what will help chopnhack.

Note that when I say "I propose" I am saying that as a moderator...


If you wonder why your post has been deleted or moved, read the section in red above.

Along those lines, I think it would be very valuable for Ratch to summarize and/or expand on his arguments in a resource because I'm sure we would point people to it when they ask the question. It would save all of us a lot of typing.

I further propose that we will continue to tell beginners that adjusting the base current is how we control a bipolar transistor (albeit with a link to any resource that Ratch wishes to write).

In fact it may be better to start two resources "How BJT transistors work (Vbe)" and "How BJT transistors work (Ib)", each having a link to the other.

Moderator hat: OFF

If you're still there chopnhack, is there anything that is still unclear?

If indeed anything is clear at this point
 
Last edited:
Status
Not open for further replies.
Top