questions on bandgap, please help me.

[Robert Qi]

I have a band gap, which is implemented with TSMC018 Process 5x5 pnp transistors. With the circuit I get 1.265v bandgap output.
But when I shift it to Tower018 Process, I can only get 1.136v output.
I test the Vbe temperature coefficient, and find that the coefficient is smaller than TSMC pnp5x5 transistor ( about 0.3mv/degree less).
Now the question is: if I want to get a 1.25v bandgap voltage reference, what can I do?

Thanks very much

 

[DenisMark]

Bandgap output is depended from properties of bip. tr. For selected by u bip.tr. form Tower018 Process I think u'll get only 1.136v output in any way. Try to choose another type of bip.tr. with another properties for bandgap ref. Or change the whole circuit to operate with 1.136v reference, or use buffer with gain 1.25/1.136. Good luck.

 

[wjlzhx]

can you let us see your schematic??
i think it can not the temperatrue coeffucient's reason.

i think the Vbe is different ,you can test it ,and you will increse the r1/r2 ,see the current.

 

[eng_Semi]

If u need a 1.25V and u can't get it with the best TC, then u have to modify ur circuit a little.

Try to generate a current proportional to 1/R and then sink this current in a resistor. U can adjust the voltage u want by adjusting the I and R values (V = IR)

 

[pfd001]

you'd better change your circuit to accomadate this new reference voltage

 

[Robert Qi]

thanks all of you, I will change the whole circuit to operate with 1.136v reference

 

[venkateshr]

instead of that laborious thing u can introduce a source follower and a resistor didvider like stuff to have the shift/gain from 1.13 to 1.25. in simple words a buffer as said b4

 

[eng_Semi]

instead of that laborious thing u can introduce a source follower and a resistor didvider like stuff to have the shift/gain from 1.13 to 1.25. in simple words a buffer as said b4

I think that using buffer, you can get a voltage smaller than the input voltage.

 

[lgqfang]

the value of the bandgap reference voltage has the relationship with the process.
if the ZTC point is fixed ,the value will be different as well as the the process.you
can get this conclusion fom the principle of the bandgap reference .And above all,
the most improtant is the TC of the voltage not the value.

 

[qutang]

I also think different process ,different vbe,adjust the circuit.