# Understanding (hfe) in data sheets

Discussion in 'General Electronics Chat' started by count_volta, Jun 28, 2009.

1. ### count_volta Thread Starter Active Member

Feb 4, 2009
435
24
So I started to make transistor circuits in real life.

On paper its easy. They give you β= 100 for example. Or its 50. Or so on. And its always 100 or always 50 and never anything else. And you use that fact to calculate β= ic/ib and so on. But what the data sheet seems to say is that β (hfe) is different at different values of VCE and ic (in fwd active)

So how do i do pencil and paper analysis when the value of β is so ambiguous. Here is the data sheet of the transistor I am using. http://www.datasheetcatalog.org/datasheet/vishay/85111.pdf

If you scroll down you can see hfe values and cooresponding currents and voltages. Can you guys please explain when for example will I have an hfe of 100?

I am used to a transistor having a constant set β from my electronics course, but it actually varies in real life even in fwd active? What the heck.

Feb 25, 2008
640
9

Feb 17, 2009
3,990
1,115
4. ### studiot AAC Fanatic!

Nov 9, 2007
5,005
515
At some point in your course they will tell you that this is the reason it is good practice to use a circuit configuration that is independent of beta. (perhaps they have already done this and you missed it?)

When we choose a particular transistor for a job we look at the beta v IC graph and try to choose a transistor where the hump in the gain curve coincides with our desired collector current. This is one reason why there are some many types available.

Remember also that beta is only the DC gain. Considerations at AC are more complicated.

5. ### Audioguru New Member

Dec 20, 2007
9,411
896
But you have the datasheet of a manufacturer who does not show typical graphs.
the graph from Fairchild shows that the hFE is typically 200 when its temperature is 25 degrees C, its VCE is 5V and its collector current is from 0.1mA to 10mA. The typical current gain drops a little above 10mA but is still at least 100 at 500mA.

But the hFE is a wide range and could be as low as 20 when its VCE is 1V and its collector current is 0.1mA.

Design your circuit so it still works properly when the hFE of the transistor is 20 to 300 and the collector voltage and current is whatever you need.

6. ### DC_Kid Distinguished Member

Feb 25, 2008
640
9
i did a quik search for 2n3906 datasheets. found one page that list 2n3906 from about a dozen manufacturers and all of them give the varying Hfe specs like Vishay does.

Feb 17, 2009
3,990
1,115
8. ### Audioguru New Member

Dec 20, 2007
9,411
896
The graphs are for "typical" parts. But some parts are very good and some are not-so-good.
The written spec's are the limits for not-so-good ones.

Vishay showed no graphs. NTE also shows no graphs.

9. ### Wendy Moderator

Mar 24, 2008
20,772
2,540
This is one of the trickier areas of transistors. Even in the same family, the same batch of the die, you can have a lot of variation.

Just to make life really complicated, there are a tremendous number of models, some don't corrilate with the others too well, one how to use transistors. They all work within their limits, but they approach the problem from different angles. Transistors are complicated enough that all of them have validity.

As a tech I use the β model, which also treats the transistor as a current device. I suspect this is what you're being taught, and overall is the most consistant and predictable. Like all the various models, it has some holes, just fewer than most.

A rough rule of thumb then, with a β of 50 if you put 1µA through the base emitter you get 50µA through the collector emitter. Since this number is so variable you don't depend on it, except to set minimum values, but design around it where it doesn't matter so much.

We had a poster who was wanting to use this to set the current through an LED without using a resistor. In theory this works, in practice it sucks and will probably smoke the transistor and the LED. β can vary according to several variable, temperature is one. It isn't as cut and dried as the formula suggests.

10. ### alphacat Active Member

Jun 6, 2009
186
0
Hey,
I'm also interested in designing a transistor circuit despite the spread out of beta.

You mentioned that a designer should design the circuit so it doesnt depend on beta.
Could you please give an example how you do that?

I'm not talking about having the BJT as a switch, which in that case you dont care about beta since you want the BJT to be in saturtion mode since in this region its VCE is the lowest (right?)

I'm talking about cases where you really want to amplify the base current, and therefore I dont understand how can you do that without being depended on beta?

11. ### Audioguru New Member

Dec 20, 2007
9,411
896
You should not bias an amplifying transistor with a single resistor from the base to the supply voltage because if the hFE is low then the transistor will be cutoff and if the hFE is high then the transistor is saturated. Also, the hFE is affected by temperature.

Instead we use an emitter resistor (bypassed with a capacitor if you want max AC gain) and set the base voltage with a voltage divider with enough current in the divider that the unknown base current won't affect the voltage much.
Use 10 times to 20 times the typical base current in the divider.

12. ### alphacat Active Member

Jun 6, 2009
186
0

Are you saying that in order to avoid relaying on beta, we will set the IC current according to VBE?
You suggest connecting the base directly to the voltage divider? (with no base resistor?).

How will you determine IC by that?
I know that in forward active region, IC is equal Is*e^(VBE/Vth), where Vth is about 25mV (thermal voltage = KT/q).
How did you intend to determine IC? (without using beta).

13. ### Wendy Moderator

Mar 24, 2008
20,772
2,540
The common emitter design with an emitter resistor is such a design. The emitter resistor provides negitive feedback and limits gain, which does what you are talking about, and prevents thermal runaway.

The AAC book has lots of info if you're interested.

14. ### Ratch New Member

Mar 20, 2007
1,068
3
count volta,

I cannot believe the answers you are being given. While they are correct, thus far, only one person has mentioned the one thing that ties everything together. And no one explains how it works or shows you where to find out. It is FEEDBACK that you have to learn, count volta. The more feedback you can put into a circuit, the less parameter variations will affect the circuit's static and dynamic operation. OP Amps have a definite advantage over discrete transistors because of their tremendous amplification, and it allows large amounts of feedback to be made, which in turn stablizes their output tremendously. Feedback is prevalent in just about every practical amplifier, so you should learn how it works. I cannot emphasize how important it is to become proficient in applying this principle. There is plenty of material in textbooks and the web. Here is a link to get you started. http://www.opamp-electronics.com/tutorials/feedback_3_04_10.htm

Ratch

15. ### count_volta Thread Starter Active Member

Feb 4, 2009
435
24
Wow....... This sounds much more complicated than the countless homework problems I did last semester.

Example.

Given that β=100 and that Vbe= 0.7 volts find all the node voltages and branch currents.

I did hundreds of those. I became good at them. And you are saying when designing circuits in real life I need to add capacitors, feedback, and etc that they did not teach us? Umm LOL.

I connected my BJT in the common emitter config yesterday and put an LED at the collector and it lit up. I measured the voltages and they more or less corresponded to what my circuit simulator gave me.

As you have probably seen I have the RLC project going on right now, but before that I decided to just mess around with transistors in real life by building something very simple.

Basically I want to make a sort of intercom device. It will have a microphone on one end, and a speaker on the other. Since the signal which comes out of the microphone is very weak I will use a BJT to amplify it.

It looks like it will involve small signal amplification since the mic produces an AC signal.

First though I want to bias the transistor properly, hence why I needed the β. But wow it looks like I need to read All about circuits on transistors and fast.

16. ### DC_Kid Distinguished Member

Feb 25, 2008
640
9
what you see in that basic audio biased bjt amp is the varying current in the base is reflected on the output by a factor close to Hfe. the voltage divider is set to just bias the bjt and the audio is coupled via cap which will cause bias to rise and fall some.

however, no need to use a discrete device, op-amps now do the work of thousands if not millions of discrete bjt's, which means you can get huge gains.

17. ### Wendy Moderator

Mar 24, 2008
20,772
2,540
Actually the capacitors aren't that necessary. It is occasionally necessary to increase AC gain (hence the capacitors) and limit the DC gain.

The AC gain of a transistor is also sloped on its frequency response, is it is a good idea to limit it to a fixed value over the range of frequencies you are using to keep the frequency response flat. Most datasheets will give the unity gain point in frequency, which sets the uppermost useful frequency response and helps set the slope.

18. ### studiot AAC Fanatic!

Nov 9, 2007
5,005
515
Not sure what you mean?

19. ### hobbyist Distinguished Member

Aug 10, 2008
773
62

Look at the circuit diagram bill marsden put on post #13.

make VCC = 8V.
make RC = 1K ohms.
make RE = 100 ohms
make R2 = 2K ohms
make R1 = 12K ohms

use a 100 uF capacitor at the base input and apply a 1KHZ signal at this input, at 10 mV PK.

Then using osciloscope observe the PK voltage at the collector, and it should be around 100mV. PK.

that's a volt. gain of 10,, (Av. = 10)

look at the ratio of RC/RE ,,,,,1K / 100 = 10 = Av.

That's one way to design for a specific voltage gain when specified, if the RE value becomes to large with respect to RC then voltage gain goes lower, but stability becomes greater (temp. and Beta dependancy) becomes less of a factor, so it's a trade off between lower gain and higher stability, so to overcome emitter swamping the input signal as well as ratio being small, a designer will put an capacitor across RE so as to keep the stability factor high but the capacitor will let only the signal voltage through so now the AC AV. goes higher.

Last edited: Jun 30, 2009
20. ### Audioguru New Member

Dec 20, 2007
9,411
896
Bias the transistor so it has enough base current for a weak one.
Then the circuit should work fine.