Bipolar junction transistor question

[Lucifre]

Value of hfe (beta) varies with the collector current on every BJT's datasheet i looked at. I was wondering how to design circuits when value of beta is hard to determine.

For example if you want to design a simple Bjt circuit that would give you 85 ma at the collector terminal... you look at datasheet and here is what you find ...
DC current gain:
Ic = 10mA, Vce = 10V min hfe = 75
Ic = 150mA, Vce = 10V min hfe = 100

how do i design my circuit then? its for phillips 2N2222A npn transistor.

I certainly dont like to make assumptions or use rule of thumbs as my book says...
certainly "Assume beta equal to 100" would not work well here (tried it and results were off)

 

[diemilio]

I don't think it would make much diference really. But I hope this helps... this is for 2n2222. Also remember that hfe varies between components of the same family, so what you should always try to do is to make your design as independent as possible of this kind of parameters.

diemilio

 

[electronrancher]

yep beta, vbe, and other parameters always vary all over the map. your circuit needs to be indpendent of this variation. for example, a circuit to get 85mA (roughly) out of the collector would be an op amp with 0.85v at it's + input, driving the base of a 2n2222 with a 10 Ohm resistor in it's emitter. The op amp - terminal would be connected to the top of the resistor.

So, the op amp would keep increasing the voltage at the base of the npn until it saw 0.85v across the 10 ohm, meaning 85mA was flowing out of the emitter. Now there will be some base current error due to beta - for example if beta is 100 then the actual collector current will be low by 0.85mA, IE/B.

Now this op amp will keep regulating the base of the npn in order to keep 0.85v at it's - input, so as Vbe varies, etc - it will be roughly constant. The only changing term will be the error due to the CHANGE in beta. so if one npn has a beta of 100 and the next has a beta of 200, their output will differ by 0.425mA. Not bad I think.

This is the simplest way. You can tweak the resistor or reference voltage to get rid of that beta error, but over extreme temp (-40 to 150C) you will still see some variation.

If it's unacceptable, use a mosfet in the same op amp configuration. No base current so Idrain = 0.85v over 10Ohm => 85mA exactly.

 

[Lucifre]

i need to build a constant current source something like a battery charger, i was thinking i could use a zener diode to provide a constant voltage to the base... and the use the collector to supply the steady current. The problem is to calculate the values of the other components ... and it all goes to hell when i get hung up on that beta ... are there other formulas i can use so that i can exclude that shifty beta... or should i just get the aproximate value from the graph and pray it works?

 

[max0412]

To provide good immunity from hfe variation use emitter bias. You will still have small Vbe temperature variations.

Example:

Emitter bias

Ie = (Vbb – Vbe)/Re ---> hfe or Bdc are removed from the equation.

As long as Vbb is constant and Re is a precision resistor (1% tol or better) Ie is constant neglecting small Vbe variation.

Ie = emitter current.
Vbb = base voltage.
Vbe = Base Emitter voltage
Re = emitter resistor.
hfe or Bdc = transistor current gain.

There are several ways to make a “constant current source” from FETS to OPAMPS to three terminal regulators. What you use will depend on the precision you need and under what conditions it will work in. Google “current source “for more info.

 

[IanP]

Cutting story short, the error is always related to the base current Ib ≈ Ic/β ..
Threfore you will always try to use a feedback ..

Take a look at the attached picture ..

It is a constant current source based on two 2N2222 BJTs ..
Its author claims that the LED current is ≈ Ube/10Ω, what is not exactly true, as you will still have Ib(Q2) and IB(Q1) somewhere in equations ..
However, because of negative feedback the β-dependance is less than 1 or 2%, so it can be just neglected ..

Regards,
IanP

 

[Lucifre]

Thanx ian...
I've seen that kind of configuration of bjt's... but i don't understand its operation ... could you give me some tips on how it works or how to analyze it ?

 

[nishantp]

Hi Lucifre
I am attaching the derivation for ur view. Kindly go through it.