Question on Current Mirror

[hurricanehenry]

Hi

https://en.wikipedia.org/wiki/Current_mirror

I am wondering why it works. It seems flawed.

The transistor on the left side, has base current coming through Iref.

Since Iref splits the current between its NPN collector, and its base, AND the base of the NPN transistor on the right-hand side, it seems there is a deficit of current.

I would understand it a bit better if the two transistors were FETs but they are BJTs and thus they draw current into the base.

So why would the output current exactly mirror Iref, if the base of the output NPN is using a bit of current from Iref?

[David Hess]

So why would the output current exactly mirror Iref, if the base of the output NPN is using a bit of current from Iref?

The output current is not exactly the input current for the reason you identify; the base current contributes an error.  In many applications this error is either insignificant or overwhelmed by other errors.  There are refinements of this circuit which compensate for the base current error.

https://en.wikibooks.org/wiki/Circuit_Idea/How_the_Wilson_current_mirror_equalizes_the_currents
https://en.wikipedia.org/wiki/Wilson_current_mirror
https://wiki.analog.com/university/courses/electronics/text/chapter-11

[hurricanehenry]

Thanks very much.

I had been bothered by this for a long time because its name is "mirror" and most people I talked to seemed to think it did just that. So I was worried I didn't understand it or something, even though it violates KCL.

I will check on the enhanced circuits you posted.

[hurricanehenry]

Ok I got it. The Wilson version is definitely better.

It first takes in base current from the reference, and then gives it back out, equally (assuming ideal and equal transistor gains).

[w2aew]

Here's a video I did on current sources and current mirrors.

[lordvader88]

If there's a university library close by, visit it. The other day I was skimming through a whole book on current mirror's and reference voltages, and they had a few books on that.


I copied down a few circuits to experiment with, but haven't made the time yet.