# a problem about the temperature dependence of the voltage

Status
Not open for further replies.

#### huashizng

##### Newbie level 6
I use a buffer opamp to increase the 1.2v bandgap voltage to 4.1v. in the divided resistors, i add a trimming resistors network. when the switch NMOS transistors's bulk are connected to the ground, the temperature dependence of the 4.1v voltage is worse. when these switch NMOS transistors' bulk are connected to their source respectively, the temperature dependence of the 4.1v voltage becomes well. how to explain this phenomenon?

#### Syukri

##### Full Member level 5
Re: a problem about the temperature dependence of the voltag

I did'nt really get your picture but I explein from what I understand

Usually this cases is related to the TC's theory

we got negetive TCs and positice TCs, integration of this makes the voltage refference to zero...

Well then the it's connnected to bulk the it will take full Vds but for this case we wanted Vbe...

when 2 Vbe diffeential exist..the relationship is proportional to the absolute tempreture

#### huashizng

##### Newbie level 6
Re: a problem about the temperature dependence of the voltag

i have gotten the 1.2v bandgap voltage. the temperature coefficient of the bandgap voltage is good enough. the bg voltage variations is no more than 2mv in the allowed temperature range. no i am increasing 1.2v bg voltage to another voltage 4.1v with a buffer opamp and resistors. these resistors include a small resistor network which is used to control the accuracy of 4.1v voltage. this small resistor network is trimmed with some switch NMOS transistors. now the problem is : if the bulks of these NMOS transistors are connected to ground, then the temperature coefficient of 4.1v voltage is bad. if the bulks are connected their source, the temperature coefficient of 4.1v voltage will be good enough.
you can see the circuit from the attached picture.

#### Syukri

##### Full Member level 5
Re: a problem about the temperature dependence of the voltag

Oh i see the problem but I can't give u an answer but here is my view..

Vout = Vbe + Vt ln n ( R1+..RN)/ R1

It menas that the volttgae at your resistor network must be equal to the nMOS switches (ideal )..

Maybe the value of resistor is ok for 1.2 but difference Vref for 4.1 is needed.

when u bulk into drain...i think the current fed to the nMOS is not quite high making the nMOS in high impedance....

BAlancing the resitor network

Status
Not open for further replies.