Pole - zero analysis of badgap circuit

[rajesh13]

I am analysing a bandgap circuit. Can some body give me some help regarding:
1) How to find the Transfer function.
2) How to check about the stability & compensation.

Basically I want to know how different transistor/capacitors contribute to poles & zero.

If some body can point me to some doc regarding the above. That will be much help.

 

[jiangwp]

May be the Razivi book of analog cicuit design can help you .

 

[Mline7]

The main loop of the bandgap may considered as follow:

- The inputs are connected above the PNPs transistors (one of them above the resistor which is usually over the PNP).

- The output is the bias voltage of the top current mirror.

- The Feedback Loop is just the top current mirror that closes the differential amplifier.

You must pay attention, because the amplifier is closed on both negative and positive inputs. So the positive loop must be lower than the positive one. So take a carefull look at the feedback factor on both inputs to determine which point is the positive input, and which one is the negative.

That way, you can calculate a transfer function, and find poles and zeros.

Note: The branch which generates the BG voltage is usually not into the loop. It is just a load for the loop.

 

[rajesh13]

How can I know that where the poles are located. I need to find the transfer function.so that I can do some mathematical analysis before jumping to the CAD simulation. can anybody direct me to some stuff which will be relevant for such a nalysis. Regarding my background, I am an IO design engineer, just few days back gor a project to design Bandgap.

 

[lastdance]

The key to bandgap ac analysis or any other circuit is to know where is the point to break the loop. The rule of thumb is to find a node where the input impedance seen by the ac source you are connecting is high (preferably gate/base) and return at a low impedance node.
You can then find the transfer function by zeroing all the DC sources and drawing out the small signal equivalents. Maybe you should put the bandgap architecture you are going to use here and see if we are able to help you with your questions.

 

[rajesh13]

I tried to cut the loop & simulated for ac analysis. Now I am wondering, how much phase margin I should choose. ?
Does the gain Margin here is of any significance. If yes, What is the impact of reducing/increasing the gain margin ?

The key to bandgap ac analysis or any other circuit is to know where is the point to break the loop. The rule of thumb is to find a node where the input impedance seen by the ac source you are connecting is high (preferably gate/base) and return at a low impedance node.
You can then find the transfer function by zeroing all the DC sources and drawing out the small signal equivalents. Maybe you should put the bandgap architecture you are going to use here and see if we are able to help you with your questions.

 

[mark_nctu]

Phase margin must be greater than 45 degree.

 

[twotoo]

???
And below 60 degree??

 

[zouwanghui]

why below 60 degree?

 

[rajesh13]

Since I am using opamp fr the bandgap circuit. I have one question reagrind the Feedback. Since the Circuit contains two feedbacks (one +ve & one -ve) so do I have to analyze these two loops seperately or in combination.

What I mean to say that, can not analyze two loops simulteanously or I have to anayze fisrt -ve feedback then +ve feedback.

 

[kevinwang123]

i come across the same problem,i doubt how i can
do it,who can help us ?

 

[rajesh13]

I was able to stability analysis using ac analysis.
I toolk the negative gain loop and observed the Phase margin for the o/p.

But I still have a queation, like what about the +ve loop, as what to look for. since stability wise that loop is unstable (as standalone).

 

[Btrend]

What I mean to say that, can not analyze two loops simulteanously or I have to anayze fisrt -ve feedback then +ve feedback.

u should break the two loops simutaneously.
and check the two loop gains.
this book may help u.

But I still have a queation, like what about the +ve loop, as what to look for. since stability wise that loop is unstable (as standalone).

as previous posts stated, the two loop are added together at the OP's output,
so if negative feedback > positive feedback, the sum at OP's output will be stable, so do to other nodes in the loop.

 

[hellheaven]

you can referer to Rincon-Mora's papers and giustolisi's papers about voltage reference.

 

[kevinwang123]

do this type bandgap circuit have to get a start-up circuit to guarantee it work?

 

[kevinwang123]

i'm confused,when i added ac singal at negative gain loop ,i don't break the positive gain loop,i can't get the right bode plot,i only use one stage OTA op.
could anyone give me some hints? thanks!

 

[wolf001]

rajesh13, you'd better paste ur circuit pic.