A question about bandgap voltage reference.

[holddreams]

A question about bandgap voltage reference.
I have run a pre-simulation of a bandgap circuit, run the transient analysis,

.tran 1n 1u sweep temp 0 85 17

and give the power supply a pwl signal,

vdd vdd 0 pwl 0 100n 3.3v

The output voltage of the bandgap is 1.35 at 0 degree.and 1.39 at 85 degree.

So my question is

What have effected the output voltage changing from 1.35 to 1.39v?
1.39-1.35=40mv seems to be big!
Is there any way to solve this problem?

Thanks.

 

[jwfan]

You can change the ratio of resistors in your bandgap. There is a PTAT current flow through the resistor and a negative temperature coefficient on your BE junction. your output is the sum of both of them. Let the temperature coefficient of them to be cancal out at about 60C.

 

[holddreams]

How to realize “Let the temperature coefficient of them to be cancal out at about 60C.”？

 

[fnlee]

chang resistors ratio m(bandgap circuit), you will cancel it.

 

[ambreesh]

Hi Holddreams

What does your DC simulation give results for the BGR output at various temperature.
Does it show a PTAT nature. If the BGR behaves as you expect it to in DC sweep, then we need to have a look at you architecture and find the bug.
I would also suggest the you incerease your simulation time. Donot sweep the supply as fast as you are doing.
Kindly share what you have found out, and all of us can proceed from their

 

[chu]

You can change the ratio of resistors in your bandgap. There is a PTAT current flow through the resistor and a negative temperature coefficient on your BE junction. your output is the sum of both of them. Let the temperature coefficient of them to be cancal out at about 60C.

hi holddreams,

Vref=KVt + Vbe ; K a function of resistors and bjt size

if Vt TC=+0.085mV/C and Vbe=-2mV/C, we find K which will cancel out each other. let say K=24

assume we increase temp from 25C to 60C, delta is 35C,
for Vbe--> decrease abt 70mV (from 35×2) from original
for Vt--> increase 71.4mV (from 35×24×0.085) from original
both this cancel out each other no matter in what temp, provided that the transistor is not breakdown at too high temp.

so u hav to scale the K (ratio of resistor and transistor size) to a good value to get stable temp independant.
hope this will help.
and guys, also let me know if i'm wrong.


rgds,
chu