Still can't get my head around what a transistor does

Hi,

I've been trying to figure out what a transistor does exactly. Read lots of articles, that go into the N and P of silicium, the sandwich and the small current at the base that allows the flow of a larger current, and the gain...

Loved the humor and insight in Grumpy Mike's article here:
http://www.thebox.myzen.co.uk/Hardware/Transistor_Tester.html

I sort of start understanding the theory behind transistors. But i still don't understand what I can do with one of these. Is it simply a switch, much like an optocoupler is (with the exception that both circuits are not isolated from each other)? I send some current in the base and the circuit closes? But then I can't get my head around the "gain" aspect of the transistor.

thanks!

XBerg

But i still don't understand what I can do with one of these.

Basically cause a big current to flow with only a little one.
So a switch, an amplifier, an oscillator are all things you can do with a transistor.

I can't get my head around the "gain" aspect of the transistor.

The gain is simple how many times is the big current is bigger than the controlling small current.
So if you have 1mA flowing through the base / emitter, with a gain of 20 this will cause 20 times or 20mA to flow from collector to emitter.

Is it simply a switch

No but it can act as one if you want.

Keep plugging away, it will come.

Grumpy Mike,

Thanks for your patience and speed of answer. Your answer has brought new insight.

Let me see if I have this right. If I use a 2N3904 transistor as a switch on the arduino, with the following datasheet:

Absolute Max ratings:
VCEO Collector-Emitter Voltage 40 V
VCBO Collector-Base Voltage 60 V
VEBO Emitter-Base Voltage 6.0 V
IC Collector Current - Continuous 200 mA

DC Current Gain:
IC = 0.1 mA, VCE = 1.0 V 40
IC = 1.0 mA, VCE = 1.0 V 70
IC = 10 mA, VCE = 1.0 V 100
IC = 50 mA, VCE = 1.0 V 60
IC = 100 mA, VCE = 1.0 V 30

This would mean that:

  • I can use this switch to drive a light / motor of a maximum of 40 V
  • If I send the arduino's 40 mA current I could drive a light / motor with a maximum of 40*60 = 2400 mA
  • There's a "sweet spot" for transistors with max gain (in our case, 10 mA), but that the more current I send to the base the more gain (10mA * 100 gain < 100mA * 30 gain)

Am I wrong in thinking the the word "gain" is incorrectly used: the thansistor does not actually create any current, nor does it even transform it - it merely allows the passage of a certain current / voltage in excess to the one it sent to the base?

Data sheet is here:
http://www.datasheetcatalog.org/datasheet/fairchild/2N3904.pdf

thanks!

Yes I think you are getting it.

If I send the arduino's 40 mA current I could drive a light / motor with a maximum of 40*60 = 2400 mA

Yes as far as the gain of the transistor is concerned but:-

IC Collector Current - Continuous 200 mA

not in practice because the transistor would melt before you got to that current. So you arrange the load not to be more than 200mA or you get another bigger transistor.

it merely allows the passage of a certain current / voltage in excess to the one it sent to the base?

Yes. That might be incorrect to your way of thinking but that is what it means.
It might help to consider the water analogy. Think of it as a current controlled tap opening or closing a valve allowing more or less water (current) to flow.

Grumpy_Mike:
Think of it as a current controlled tap opening or closing a valve allowing more or less water (current) to flow.

Brilliant! That should be the 1st line when talking about Transistors. Not that nonsense about silicium N and P properties! :slight_smile:

Maybe the root word of transistor would help light your bulb? Transfer of resistance.

Lefty

If you're using it as a switch, a parameter from the data sheet of particular interest is Vce(sat)--in english, the collector-emitter voltage at saturation. It's often something like 200mv or maybe 1V for a power transistor carrying a lot of current. With that number it can be pretty easy to figure out the current flowing in your load circuit--just assume a 200mv or whatever drop across the transistor. To make sure you drive the transistor into saturation in the "on" state, put in enough base current to produce that load current given the gain (hfe), and then some to add margin. The gain is dependent on temperature and other factors, so you don't want to rely on getting some particular value of gain.

With that number it can be pretty easy to figure out the current flowing in your load circuit-

No it is not, generally it is independent of current. Although there is a non linear relationship.

Quote - "If I send the arduino's 40 mA current I could drive a light / motor with a maximum of 40*60 = 2400 mA"

In this case no. The Continuous current the 2N3904 can handle is 200ma, this is a maximum for continuous operation, after that you will let the smoke out. Have you heard the old saying (tongue in cheek), "electronics works on smoke".

Add these parameters to be aware of when reviewing a transistor:

Vcesat - collector-emitter saturation voltage = 200mV (condition : Ic=50mA, Ib=5mA)
Ptot - Total Power Dissapation - 500mW

With a collector-emitter current IC = 100ma, your gain is only 30, i.e. it would take a minimum 3.3 ma on the base.

When using a transistor as a switch, you want to overdrive the base current slightly, to ensure that you saturate the transistor to minimize the Collector to Emitter voltage, this reduces the power the transitor it is dissapating, and compensates if the transistor should have a lower gain then specified. So if you required 100ma to drive a light, then you should use about 4ma on the base. Assuming saturation, the transistor would dissapate P = VxA = 200mV x 100mA = 0.2V x 0.1A = .02W which is well under the .5W (500mW) allowable.

As a word of caution, I would never draw the 40mA from an AVR pin, I would rather buffer with additional electronics to protect the AVR.

Reading datsheets is good behavior. Even if you only understand a small portion of a datasheet at first, you will be stuffing that info into your grey matter and things will eventually start to make sense.

One of the things mentioned already is the requirement/need/desire to see MAX ratings as MAX ratings. Any good engineer or tech will tell you that the sign of a bad design is running parts close to MAX specifications. A good design plays in the "middle of the road" for the most part. You always want to engineer a solution that avoids using the OVERKILL part (...costs more and might live through worst case scenarios) but that also does not run cheaper parts close to MAX ratings.

RETRO FLASH BACK... these are some old transistors I'm about to talk about.

Example: You sometimes need to decide between using the 2N2222A and the 2N3904. The devices are very similar in most respects, but the 2N3904 can really only handle about 1/3 of the current the 2N2222A can. If you need more power dissipation than the 2N2222A in an otherwise identical design, you can use a 2N2219 for the same solution since it's quite similar with the exception of the larger packaging for power dissipation.

Here is how I introduce a transistor to my electronics students. I certainly do the classic diode->transistor explanation but I found this alternative way more understandable and will help students try to understand the classic explanation. See if you get it or not.

Think of a cathode and an anode both in the shape of metal sheets, with the anode as a larger can and cathode as the smaller can inside but not touching the larger can. The cathode is heated by a filament so hot electrons keeps oozing out of it and wander around with no particular preference which way to go. You then apply a voltage (anode +, cathode -). According to basic electric circuit laws, the hot electrons from cathode travels through vacuum and arrive at anode, completing the circuit so there is a considerable amount of current in the conduction.

Now look at the grid sitting in between the cathode and anode. You apply a voltage on it and you can convince much of the hot electrons (thermionic electrons) not to reach the anode. Then you have control over the current of the tube. You vary the voltage on the grid, you vary current through the tube. Is this clear enough? This device is equivalent to a field effect transistor, which uses a voltage (on the grid or gate) to control the amount of current through the tube (or drain-source). The 2N3904 is a different type of transistor. It instead uses a small current to control the current through itself (collector-emitter).


Credit to wiki

Most electronics texts start with tubes and FET transistors, because it's more intuitive to understand how an electric field set up by voltage on the grid/gate "pushes" charge carriers out of the conductive region between the other terminals. But I'm not sure that the theory carries over well to bipolar transistors, which are current-controlled devices instead.

Operation as an amplifier is hardest to understand. Going back to the water valve analogy, there are limits to how much you can open the valve; you turn the handle all the way, and however much water can flow through the valve does, depending on the other circuit elements (pipe, water pressure, nozzle, etc.) When using a transistor as a switch, you provide enough base current to make the current through the transistor CE path be limited only by elements other than the transistor itself. To operate as an amplifier... Imagine the water coming out through a fountain. The height that the water reaches will depend on how far the valve is open (and also the pipes and nozzle diameter and such.) You can limit the operation of the valve between two points that are neither all on or all off, and the height of the fountain will vary between an upper limit and a lower limit. That's your amplifier...

Good point. You can open the door or valve all the way up but there is a limit of how much current runs through the gateway.

BTW, I am using Bobrow's book. It's the only one that compresses several semesters of EE into one book so non-EE students can handle it. It's ancient by today's standard but the index has no tube or triode and I don't recall reading about them on the book either. This book was written in the 90's. You'd have to go back a couple decades to see tubes, which is sad. They are such good example of one-way conduction and current devices. Wow, even firefox thinks triode is a wrong word.