How to improve PSRR of a bandgap?

[Joestar]

I design a Bandgap reference, usesing the cascode current mirrors to make up PTAT currents .This bandgap have a poor PSRR, Who can tell me how to improve the PSRR? Thanks!

 

[Vamsi Mocherla]

how to improve psrr

Increase the lengths of the devices and scale the widths accordingly. It well help your PSRR

 

[mike_bihan]

increase psrr

Is there any paper talking about improving PSRR of the bandgap?

Or where is the PSRR coming from?

 

[Alles Gute]

psrr improve

Philips E.Allen "Cmos analog circuit design" second edition, chapter 6.4

 

[paladinzlp]

how to increase psrr

1.increase the lenght.
2.decrease the current.
that will improve the PSRR(w=0)

Added after 3 minutes:

1.increase the lenght.
2.decrease the current.
that will improve the PSRR(w=0)

 

[ambreesh]

psrr current sources

Dear Joestar,
Your PSRR is bad at what frequency. Or is it bad close to DC.
How deep in saturation are your PMOS current sources in the voltage reference branch. By any chance is any of the two sitting on verge of saturation and linear region.
As others have suggested, you should increase your L and W in same proportion iff you are using submicron lengths.

Vishal gupta has presented a good paper on PSRR in Bandgaps. But that is with opams in the architecture.

 

[xiongshoufen]

increasing psrr

connect cap from vdd to gnd

or connect cap from current source transistors to vdd or gnd

the above way can inrove psrr, thanks

 

[okawa]

improving psrr

the capacitors in ICs are large enough to sink the ripples ? the capacitors will ocuppy the whole wafer

 

[qqmz]

psrr improvement

why do you say that? does it because the signal through the power can pass the cap?

connect cap from vdd to gnd

or connect cap from current source transistors to vdd or gnd

the above way can inrove psrr, thanks

 

[ambreesh]

cascode current mirror psrr

Indeed that is what it means. The cap would provide a low impedance path to supply ripple to gnd. But only to supply ripple whoes frequency is high enough for the capacitor value

 

[dasong]

how to improve opamp psrr-

Who can give me some papers about how to improve PSRR,thank you!!

 

[MLR67]

improve cmos psrr

Hello JoeStar (like the Rap Singer ?)

first if you want to improve your PSSR+ and if you have got enough power supply you can use cacode for current mirror.
The second way and the best is to use a boostrap bangap (bootstrap the VDDA) but again to do it you need at least VDDA≥2.3 volts.

 

[dasong]

psrr increase

Can you tell me how to design the boostrap bangap ,I know boostrap OPA ,but I don't know boostrap bangap .

 

[ambreesh]

improved psrr in opamps

Dear MLR67,
Are u talking of regulated VDD for the BGR by bootstraping.

 

[|IAngel|]

current mirror psrr

As someone who is currently designing my 4th low power bandgap reference, the best ways to improve your PSRR are increasing the lengths of your transistors so that the PS fluctuations don't have as much weighting, or to bootstrap the PS as mentioned.

If you want to make it much easier to reject the PS, generate the PTAT with a chopping op-amp or an opamp with no (very low) offset and bootstrap the bandgap. This way you will have a low voltage bandgap with good PSRR.

 

[ambreesh]

increase psrr op amp

Dear Angle,
Please clarify. What has PSRR improvement and chopping got to do with each other.
I think I am missing a link, so please elaborate.
Now we are bound to have an Opam in our architecture that would be driving the gates of PMOS (current sources).
even if I increase the length and my opamp topology does not pass the ripple to the gate of the PMOS it does not halp much.
The correct choice of the opamp topology is more critical in BGR design employingh opamps

 

[|IAngel|]

how to improve psrr

I couldn't agree more.

A major issue with bandgap performance, especially with PSRR (and a little with temperature) is the offset voltage of your amplifier. It SERIOUSLY degrades performance with respect to you PSRR and temp independance. Using a chopping amplifier removes the offset for better performance.

Also - what type of process, supply voltage and op-amp are you using? With a rough initial design, I found that a 1 stage amp acheived 5 ppm/K temperature dependance and 250 ppm/K supply dependance. I made a more supply friendly 2 stage amp and the PSRR went down to 180 ppm/K. This was also a rough design.

Voltage offset is a real issue when employing opamps with bandgap references.

Questions/Comments?

 

[megzhuy]

psrr opamp improve

"AN IMPROVED BANDGAP REFERENCE WITH HIGH POWER SUPPLY REJECTION" , IEEE.
I think it's easy to understnd for beginner,hehe.