Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Bandgap reference ( output voltage variation)

Status
Not open for further replies.

mike_bihan

Full Member level 3
Full Member level 3
Joined
Mar 21, 2002
Messages
182
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,296
Activity points
1,531
Normally we have equation of output of bandgap:

Vref = K*Vtn + Vbe

Normally, K is consists of ratios of resistors and BJTs. However, Vbe of the BJT is related to current, which is related to resistor values.

Why there is no concern for the resistor value variation?
 

I think that erramp's output control the current sink,then the current sink go to R2,then it cancel the R's variation.that's my opinion
 

Indeed, the absolute value of the current (determined by R1) will influence the bandgap output voltage. But it's a second order effect. If the resistance changes +/- 10% (typical for IC processes), the output voltage will only change +/- ln(1.1)*Vt = +/- 2.5 mV.
However the spread of Is and Beta in the transistors will give more deviation.

Sceptre
 

I think it will have some dependencies on resistor's value. So in high-accuracy bandgaps, fine-tuned resistors is needed.
 

in accurate reference, the voltage only depend on the ratio of resistors.
 

Qutang:

You think Vref is independent of R ?
Please provide more detail if possible.

Actually, I agree with sceptre.
 

Haha - you guys need to start doing your maths on this stuff.. For the most part, it washes out. For example: Multiply all resistors by 1.2. now ptat current is 20% less than you previously thought, but Rgain is 120% of what you previously thought => 1.

In reality, we add trim circuits to bandgaps once we start worrying about process variation. BUT - for a "draw it and go" bandgap, you should be able to get within a few percent without trim.
 

you must consider this question in your design. you can choose different type of resistors to get your design performance.
 

Can someone explain how to choose the suitable bias current value of the PTAT ckt to minimize the Vbg variation?

Besides, if temp coefficient of resistor is considered, how to lower its effect on Vbg variation??

Thanks for any comment in advance : )
 

Can someone explain how to choose the suitable bias current value of the PTAT ckt to minimize the Vbg variation?

Besides, if temp coefficient of resistor is considered, how to lower its effect on Vbg variation??

Thanks for any comment in advance : )
 

Question on the bandgap circuit:
What threshold voltage have the NMOS transistors ?
NMOS gate-sourse voltage should be less than voltage of diodes over temperature range !
 

In reality, we add trim circuits to bandgaps once we start worrying about process variation. BUT - for a "draw it and go" bandgap, you should be able to get within a few percent without trim.

How does a trimming circuit for bandgap look like? Can anyone provide an example?

Thanks
 

the trimming resistor is always use a narrow metal tied.a few company use the zener diode
 

This bandgap circuit is constructed in voltage mode and the resistor does affect the temp coefficient of the bandgap circuit as it is temperature dependent too. Such problem can be cancelled out if current mode bandgap is used.
 

it also should be concerned that the load of the Vref will affect the banggap, if the Ip-p of the load is 4ua, how to get the right quisient current in the output stage of the bandgap?
 

In this circuit, there is the start-up circuit at the right. Anyone can discuss how the start-up circuit works in this bandgap circuit? And is there a power-down circuit? If no power-down circuit, how to simulate the start-up circuit?
 

In your bandgap equation, K= R2/R1 * ln(n) where n is the ratio between bipolar. So your are independant of resistor variation in case of you've matched your resistor in the layout.
 

Startup circuit will work:
Initially R3 will pull down gate of M6 so M6 is off. So Gate of M5 will high turning on M5. This will pull down gate of M1,M2 ,M3 so all current sources starts flowing current so bandgap will start. Once Vref will come up and circuit is self sustaining.
M6 is On which is much sttronger then upper diode connected P tranistor so M6 On , switching off M5.

Power down signal will be added up to diode connected P transistor. Instead of making it diode connected , gate of it will be connetced to powerdown.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top