Question about bandgap reference circuit: equal current

[diego.fan]

Hi, all

I have 1 questions need your help.


In sasen's book. it said the Ic of Q1 and Q2 must be same.
Sasen also said "This can be achieved by another current mirror on top.Morever, this pnp current mirror can be used to add another current ratio n."

My questions:
I think Ic of Q1 and Q2 is not the same. Because Q4 and Q3 are not in the same size. From Ic=Is*exp(Vbe/(kt/q)), we can see different size-->different Is-->different Ic.
It means Ic of Q1 and Q2 are different. Am I right or not?

 

[jusung]

i think the book is wrong, you are right. the current ratio through Q1 and Q2 is n:1 by the current mirror (Q4, Q3)

 

[jusung]

i think the book is wrong, you don't have to use same bias current of Q1 and Q2 when you design BGR. And the current ratio of this given circuit is n:1.

 

[dick_freebird]

The Ic ratio will swing as current goes up, due to the
emitter resistor on Q2. At very low currents the Ic
ratio is 1:r. At balance, the ratio will be less than
1:1 because of the 1:n PNP ratio - Q1:Q2 must fade
or else there would be runaway rather than balance.

Just because it's in a book, doesn't mean it's true.
A better bet than the Internet, but fact checking
is always finite.

You could do a quick simulation and demonstrate
the truth or the error.

 

[diego.fan]

Yes, you are right. I also think the book is wrong here.

- - - Updated - - -

The Ic ratio will swing as current goes up, due to the
emitter resistor on Q2. At very low currents the Ic
ratio is 1:r. At balance, the ratio will be less than
1:1 because of the 1:n PNP ratio - Q1:Q2 must fade
or else there would be runaway rather than balance.

Just because it's in a book, doesn't mean it's true.
A better bet than the Internet, but fact checking
is always finite.

You could do a quick simulation and demonstrate
the truth or the error.

Yes, I will. Thanks