Understanding transistors

[WelshWonder]

Hi all,

I'm having difficulty understanding how transistors work alongside the AC singnal from a guitar pickup. With DC, I understand it works as a switch so when enough current is applied to the base then it allows current to flow from the collector-emitter junction. So in this type of application it can act as a switch. Does the same thing work with AC signals? I mean for example, if the base needs +5.5V to operate and you supply +5V via a DC source, can the AC signal trigger the transistor by providing more than +0.5V?

[Sir H C]

A bipolar transistor is a voltage controlled current source.  The relationship is exponential:

Ie = Is * e^(vbe/vt)

So the current through the emitter is related to the base-emitter voltage.  DC or AC, whatever voltage is on that base-emitter at an instance will create the current through the emitter/collector path to do that (within the limits of the supply and load).

[R.G.]

Go read the last part of this:

http://geofex.com/Article_Folders/How_It_Works/hiw.htm

[DryRoasted]

R.G., if you ever run for president, I'll vote for you! (Even though I'm English and you're all a bunch of ungrateful colonials who drive on the wrong side of the road )

[johngreene]

A bipolar transistor is a voltage controlled current source.  The relationship is exponential:

Ie = Is * e^(vbe/vt)

So the current through the emitter is related to the base-emitter voltage.  DC or AC, whatever voltage is on that base-emitter at an instance will create the current through the emitter/collector path to do that (within the limits of the supply and load).

um, I think a FET is a voltage controlled current source. A bipolar transistor has a relationship of currents. ic = ib*Hfe. The base-emitter voltage is going to remain fairly constant.

--john

[Sir H C]

A bipolar transistor is a voltage controlled current source.  The relationship is exponential:

Ie = Is * e^(vbe/vt)

So the current through the emitter is related to the base-emitter voltage.  DC or AC, whatever voltage is on that base-emitter at an instance will create the current through the emitter/collector path to do that (within the limits of the supply and load).

um, I think a FET is a voltage controlled current source. A bipolar transistor has a relationship of currents. ic = ib*Hfe. The base-emitter voltage is going to remain fairly constant.

--john

No both are truly voltage controlled voltage sources.  The relation with beta is not a very good one to use for most purposes as beta is a widely varying secondary parameter, from the physics Ic = Is * e^(vbe/vt) + 1.
( I see that I hosed this above with Ie, lack of sleep).

For MOS devices, the relation ship is square law (or somewhere around there depending on process and the rest).  Somewhere Barrie Gilbert of Analog Devices did a great paper on not using beta for biasing purposes with ICs. 

[R.G.]

R.G., if you ever run for president, I'll vote for you! (Even though I'm English and you're all a bunch of ungrateful colonials who drive on the wrong side of the road

Why thank you - I think.  We have this idea that you should never make a politician out of an honest man - you can never change him back once he tastes blood.

Ie = Is * e^(vbe/vt)

That's true - but essentially useless to the beginner. Is varies too, with the construction of the junctions. And Is is never given in a datasheet. It's a good thing to know for background in understanding some circuits, but very hard to apply, especially for beginners.

Hfe being a secondary and variable parameter is true, and an only moderately useful parameter. It's kind of a signpost for low-medium-high gain devices. But it's easy for beginners and modestly advanced designers to use.

The real thing that one needs to do (unless one is designing analog ICs) is to design circuits which make the forward gain simply not matter much. The techniques of biasing and setting gain which rely on things that change hardly at all, like Vbe and feedback resistors, is the way to get practical, working circuits done with bipolars.

You're right - for JFET and MOS devices the relationship of current with Vgs is a square law function, but there are funny construction-dependent parameters in there too. One Thing that makes JFET and MOSFET devices harder to bias as a practical matter than bipolars is that the voltage that they depend on, Vgs, is much larger than a bipolar's Vbe, and varies more. For all intents and purposes, a bipolar's Vbe is constant in its working range compared to the Vgs vagaries of FETs.

Triodes are a 3/2 law device BTW.

[johngreene]

No both are truly voltage controlled voltage sources.  The relation with beta is not a very good one to use for most purposes as beta is a widely varying secondary parameter, from the physics Ic = Is * e^(vbe/vt) + 1.
( I see that I hosed this above with Ie, lack of sleep).

For MOS devices, the relation ship is square law (or somewhere around there depending on process and the rest).  Somewhere Barrie Gilbert of Analog Devices did a great paper on not using beta for biasing purposes with ICs. 

Now they are voltage controlled voltage sources? I'm assuming that is a typo.... 

But if I was a beginner, I'd be getting pretty confused by now.

So, if I was to play 'beginner advocate'.... As a beginner I don't care about expotential law, square law, or 3/2 law or the physics behind it for that matter, all I see it that since you are saying both a FET and Bipolar transistors are 'voltage to current' converters, I should not be able to drive either one with just a current source without a resistor to ground to convert the current into a voltage before the device?

--john

[Sir H C]

No both are truly voltage controlled voltage sources.  The relation with beta is not a very good one to use for most purposes as beta is a widely varying secondary parameter, from the physics Ic = Is * e^(vbe/vt) + 1.
( I see that I hosed this above with Ie, lack of sleep).

For MOS devices, the relation ship is square law (or somewhere around there depending on process and the rest).  Somewhere Barrie Gilbert of Analog Devices did a great paper on not using beta for biasing purposes with ICs. 

Now they are voltage controlled voltage sources? I'm assuming that is a typo.... 

But if I was a beginner, I'd be getting pretty confused by now.

So, if I was to play 'beginner advocate'.... As a beginner I don't care about expotential law, square law, or 3/2 law or the physics behind it for that matter, all I see it that since you are saying both a FET and Bipolar transistors are 'voltage to current' converters, I should not be able to drive either one with just a current source without a resistor to ground to convert the current into a voltage before the device?

--john

Yes typo, voltage controlled current sources.

YOu can drive a bipolar with current into the base and get a current out that tracks pretty well for a couple orders of magnitude of current.  Not perfectly, and it can drop off pretty quick on the extremes, but it works.  For MOS and jFET devices you can not, as they are high impedance in so there is no path for that current coming in.

If you want to drive in with a current, then the coolest way is to drive the emitter and hold the base steady.

[Thomas P.]

BJT are current controlled current sources (at least that's what I have been teached). You will get a voltage by applying a resistor at the collector.
Other than Fets where there is no Ig (theroretically) - that's why they are called that way (field effect transistor) - therefore they are voltage controlled current sources. The field at the gate controlls the conductivity of the channel.

[johngreene]

I guess my point is I've seen a few beginners here try to bias a bipolar transistor with 2 10 Meg resistors and not understand why the base voltage is not 4.5V, it works with FETs. The bipolar REQUIREs base current. A FET does not. So a statement as general as both being "voltage to current converters" is misleading.

--john

[Doug_H]

I guess my point is I've seen a few beginners here try to bias a bipolar transistor with 2 10 Meg resistors and not understand why the base voltage is not 4.5V, it works with FETs.

Just as an aside- they will also not understand why they're not getting the huge input impedance they are expecting...;-)

Back to your regularly scheduled program...

[Thomas P.]

So a statement as general as both being "voltage to current converters" is misleading.

--john

Absolutely! Furthermore it sweeps the whole difference between BJT's and FET's under the carpet.
For a BJT is more like this
,

while a FET is more like
http://www.pbs.org/transistor/science/info/transmodern.html

[mac]

Sadly I have to admit that I can explain the physics behind transistors tubes and fets, but I do not know how to put them to work 

mac