Transistor Basics

tobyw

Member
I am really struggling to find a basic guide to transistors which puts it in words I can understand. I am really hoping someone here can explain it more clearly and simply!

Q1.

What is it which switches a transistor on? Voltage or current - I understood it to be voltage, but I also see things talking about how you need to limit the current..

Q2.

If the answer is voltage, then does applying the voltage mean the gtransistor is fully on? So what happens if you exceed the minimum voltage? Does it make any difference? I've seen some articles mentioning that a particular transistor needs 0.7 volts to switch it on. But if you had a motor load on the transistor, and you applied 0.7 volts to the base, would this literally mean the motor would be as fully powered as if it were directly connected to the power source? Or would it be 'slightly' on, and as you increase the voltage to the base the voltage to the motor would increase?

Q3. What about transistors as amplifiers? I understand that you need to bias the transistor, but what does this mean? Are you making it slightly switched on, so that every tiny voltage difference fully opens and fully closes the transistor?

Sorry these are such basic questions - I know I can just stick a 1k resistor and muddle through..but I really want to understand what is happening.

Many thanks
Toby
 

premelec

Senior Member
This source was recently referenced in a forum post: http://www.mikroe.com/old/books/keu/00.htm - It doesn't discuss more recent components but is a good start. Also look at data sheets and the graphs shown to see what electrical characteristics are. Definitely read the PICAXE manuals - esp. Manual #3. There's a lot to learn - it takes some time... :)
 

geoff07

Senior Member
There is a lot of stuff out there. This is quite good:

www.cedengineering.com/upload/Transistors.pdf


In simple terms (it is possible to get a lot more complicated):

Q1.

What is it which switches a transistor on? Voltage or current - I understood it to be voltage, but I also see things talking about how you need to limit the current..

Both!

In a Bipolar Junction Transistor (BJP) - the sort they all were originally - they are current operated. The collector/emitter current is in principle the base current x the hfe (gain). The flow of electrons does the business.

In a Field-effect transister (FET) - the sort many are these days - they are voltage operated. The voltage creates an electric field that influences the current passing between the source and drain. Virtually no current flows into the gate, so they can be very energy efficient and suitable for high impedance work. A Picaxe is made from a Microchip PIC which is made from hundreds of thousands of FETs built as one item.

To understand this you need to get into semiconductor physics, but to use them you don't.

Q2.

If the answer is voltage, then does applying the voltage mean the gtransistor is fully on? So what happens if you exceed the minimum voltage? Does it make any difference? I've seen some articles mentioning that a particular transistor needs 0.7 volts to switch it on. But if you had a motor load on the transistor, and you applied 0.7 volts to the base, would this literally mean the motor would be as fully powered as if it were directly connected to the power source? Or would it be 'slightly' on, and as you increase the voltage to the base the voltage to the motor would increase?

A BJT needs about 0.7 volts to overcome the base-emitter threshold (same as the forward drop in a diode). Once the base voltage is higher than the emitter by 0.7v then current starts to flow. So no it would not be fully on ('saturated') it would just have started to conduct. Saturation will depend on hfe and other device characteristics, which you can get from the data sheet.
A FET has a similar 'knee' characteristic.

Q3. What about transistors as amplifiers? I understand that you need to bias the transistor, but what does this mean? Are you making it slightly switched on, so that every tiny voltage difference fully opens and fully closes the transistor?

Bias is what deals with the 0.7 volts threshold or the FET knee, as without it a transistor would only notice input peaks above 0.7v. So you bias it by shifting the neutral point to somewhere above the 0.7. Then it can see and respond to the whole waveform that is input.

You can distinguish between devices used for switching (power switches or digital circuits) which are either on or off but not in between, and analogue usage devices where the operating point is on a slope and movement is up and down the slope, usually not reaching either fully on or fully off. Look at a datasheet for various types of transistor e.g. BC548 (NPN BJT) and 2n3819 (FET).
 

boriz

Senior Member
A BJT amplifies current changes. A small change in the Base-Emitter circuit current causes a large change in the Collector-Emitter circuit current. The ratio of the size of these two changes is called the Gain. IE: Vary the base current from 5mA to 7mA (a change of 2mA) and if the Collector current changes from 500mA to 600mA (a change of 100mA) then the gain is 100mA/2mA = 50.

If the Base current starts at zero, so will the Collector current. That's how BJTs are used as switches.

The Base-Emitter architecture is very much like a diode, and will not conduct (will not pass any current) until the minimum threshold voltage is reached. Usually about 600mV.
 

g6ejd

Senior Member
Q1.

What is it which switches a transistor on? Voltage or current - it's both. You need a potential difference (voltage to make a current flow) and it's the electrons that then start the transitor action. Current does need to be limited, so hence the need to add a series resistor, if the voltage couldbe accuratel;y controlled, then you woudl not need a resistor. When current flows are very very low, then the device is generally known as voltage operated, but it still needs a current.

Q2.

If the answer is voltage, then does applying the voltage mean the transistor is fully on? Not always, it depends on the device and how much inputer curerent is required. For example, feed a high current low gain device with 1mA and not a lot will happen, but do that with a high gain low current device and you get the converse.
So what happens if you exceed the minimum voltage? Damage in the extreme, but more likely a saturated device that switches and does no more, well there will also bee some heating as the excess energy is dissapated.
Does it make any difference? Yes, the device becomes non-linear and exhibits un-expected result.
I've seen some articles mentioning that a particular transistor needs 0.7 volts to switch it on. Yes in general, the input junction is closely allied to a diode and to get a diode conducting, you generally need 0.7 volts to do that. But if you had a motor load on the transistor, and you applied 0.7 volts to the base, would this literally mean the motor would be as fully powered as if it were directly connected to the power source? Or would it be 'slightly' on, and as you increase the voltage to the base the voltage to the motor would increase? The two poarts of the transistor are broadly speaking isolated, so in your example current flow would be from the power supply through the motor (plus a tiny amount of current from the input (base) circuit which together flow through the Collector-Emitter part of the circuit, so no, as there is a limit to the minimum on-resistance of the transistor, so there will always be some losses.


Q3. What about transistors as amplifiers? I understand that you need to bias the transistor, but what does this mean? Are you making it slightly switched on, so that every tiny voltage difference fully opens and fully closes the transistor? Yes, you need to apply enough inpuy current to get the device into a region of linear operation, so check out transistor operating curves, where you will see at low currents a knee shape that climbs from 0, then eventually it will get to a near staright line of increasing conduction for increasing input currents, a linear region. Important if you want the device to be linear and give you good sound quality. Biasing is essential.
 

tobyw

Member
Thanks so much for your replies.

Eclectic - that was one of the items I have read, but I didn't really find it very easy to understand.

But I knew people here would be better at explaining it. Particular thanks to geoff, boriz and g6ejd. Your explanations really helped. Particularly about how it amplifies current. Although I am still preplexed by the whole idea of how to work out the correct resistors to use in order to get full saturation when using it as a switch.

I will keep experimenting with the transistors I have and with these explanations hopefully I can to a point where the theory and practice show the same results - then I'll feel like I get it.

In the meantime, I hope you don't mind more questions to try and nail this!

So last night I created a breadboard circuit with a BC548 transistor and a 3V power source. The collector was connected to 3V. Emitter to an LED via a resistor, and to GRD. Base was connected to a 1k resistor + 10k potentiometer voltage divider. I could see that as I changed the potentiometer, the voltage increased and eventually the LED lit up, but only dimly, and never as bright as if I connected it directly to the 3V.

So this was a confusion to me. Is it because even with 3V directly applied to the BC548 base, that's not enough to allow the full 3V to flow through the collector/emitter, hence why the LED doesn't fully light?

I'm not a very good mathemetican at all - in fact slightly challenged in that deparment - but I will try to understand the hFE, max current/ max voltage calcs to work out what resistors to use etc - however, it would help to know whether broadly speaking this is the reason for what I am observing in practice.

Thanks again :)
 

JimPerry

Senior Member
What value was the LED resistor? 3V0 minus 0V7 via transistor is pushing it for some LEDs - what colour is it?
 

vttom

Senior Member
Try this experiment again, but put the LED+resistor between 3V and the collector of the transistor, and connect the transistor's emitter directly to GND.
 

geoff07

Senior Member
3v isn't really enough, as suggested already by JP. This table tells you the voltage drop of different colours:

http://en.wikipedia.org/wiki/Light-emitting_diode#Colors_and_materials

To get any current through the transistor, you need
0.7 v (which is code for a value between say 0.55 and 0.7 depending on device and current flow)
+2.1v (say, for a red led)
+ 0.9 to 0.25v (the drop across the collector/emitter resistance of the transistor at the operating current = the saturation voltage)
--------
2.89 to 3.05 i.e. marginal if you have only 3v

The maths (actually simple arithmetic) isn't very difficult when you know how to do the calculation.

To get the transistor to turn on you need a base current say 100th of the desired collector current (assumes hfe=gain is 100 - it can be higher in a BC548). That current has to flow through your base resistor, pushed through in your original circuit by 0.2 volts (3-0.7-2.1).

Assume 10mA through the led, so say 100uA through the base resistor (10mA/100). V=IR so voltage required = 100 x 10^-6 x 10^3 = 0.1v

So you have just enough voltage on the base with the pot at minimum but probably not enough on the collector/emitter. Reconfigure as suggested by VT and you gain more voltage on the base so more scope for adjustment there. Increase the voltage to say 4.5 and you get more scope for the led in that configuration, and most colours should then work.

If you are wondering why garden lights can light an led with one AA cell, it is because they use a voltage multiplier circuit.
 

tobyw

Member
Sorry for late reply. Thanks so much for the replies. I knew I would get a much better explanation on this forum. Thanks especially to geoff and g6ejd. Your explanations really help to explain the concepts. Particularly this...

To get the transistor to turn on you need a base current say 100th of the desired collector current
Most of the online and book explanations I have struggled through make it sound like the transistor is a switch which behaves like an amplifier - but explaining it as just an amplifier really makes it clear why it acts like a switch, and most importantly how its not a boolean on/off state.

Anyway - armed with my new knowledge I managed to create a touch sensor with two BC548s wired up as a pair last night - so I feel like I get it at last!

Many thanks to all once again

Toby
 

vttom

Senior Member
I think there are 2 basic equations you need to know about bipolar transistors...

I(collector) = I(base) * Beta
V(emitter) = V(base) -Vth

where Beta is the gain of the transistor (look for it on the data sheet)
Vth is the forward threshold of the base/emitter junction (also should be spec'd on the data sheet)
 

westaust55

Moderator
From the datasheets I have looked at over many years, the symbol for BJT type transistor gain is using Hfe.
For many applications, such as driving LEDs, where the transistor is used as a switch we need to consider the gain when in saturation. Often this is not given as a specific line items but many times from other information in the datasheet is can be deduced (For example the datasheet may state the base and collector current associated with VCE(Sat) ). As a last resort I have based on available data from numerous datasheets had determined that assuming the gain in saturation is 0.1 * gain in linear operation will be of the right order.
 
Last edited:
Top