Current Mirror Explanation Needed.

[tsiros]

you're not supposed to know β accurately, anyway... just that it's some large number... at least that's what i've learned from horowitz&hill

it will take some reading (and drawing) to make sure i 'get' what you typed last

again, many thanks.

[jan.didden]

Originally Posted by tsiros

you're not supposed to know β accurately, anyway... just that it's some large number... at least that's what i've learned from horowitz&hill.

Exactly. Not only does the beta vary between transistors, even the same type, but the Vbe necessary for a specific current also varies.
If you read some of the earlier posts, someone mentioned using smallish (10-100) ohms emitter resistors in both mirror transistors. The resistors provide negative feedback and that will tend to cancel out beta- and Vbe differences.

How does that work? Suppose the 2nd transistor has more beta than the first one. That would give it more Ic than the first one and the 'mirror' is no longer 1:1. But the larger Ic (and Ie) also means that there is more voltage generated across that 2nd transistor Re. That extra voltage can only mean that the Vbe gets a bit smaller, which decreases the Ic/Ie. So it works to diminish the effect of any beta differences (and similarly any Vbe mismatches).

jan

[cbdb]

It might help to start with the simpler diode transistor CCS, get a handle on it, then return to the transistor wired as a diode. The operation of the two is identical (except in matching).

[tsiros]

so the current mirror is stiff in β and Vbe(Ie) variations...

hey do me a favour? post your interview with Nelson Pass somewhere, the link from your blog is going nowhere