Help me get the same current from die to die

[ccheng]

analog design problem

when i used a constant gm circuit , i need a resistor to generate the biasing current . But , in the practical , the resistor has a big variation in the ic . so i can not get the same current from die to die .

Can anyone tell me , how to solve this problem ?
Or is there any other circuit have the same funcation ?
ps. there is no external pin for the resistor.

 

[Basil]

Re: Analog design problem

what about resistor ratio from die to die?, I mean you have R1/R2 in one die and R1'/R2' in the other chip, is (R1/R2)/(R1'/R2') close enough to one to use some circuit based in ratio not in particular values?.

another aproach could be to use some high gain block and put your variable component in the loop (to have variation divided by (1+GAIN)) and some bandgap or other stable reference to make the job.

Only ideas, I don't know if feasible

 

[rambus_ddr]

Re: Analog design problem

In Analog Integrated Circuit Design by David Johns &Ken Martin (Author)
there are detail explains

 

[rfsystem]

Re: Analog design problem

The best what you could get is a Bandgap. Use a PMOS current source and drive though an external resistor a current such that the voltage equals your internal Bandgap. Then mirror these regulated current internal. That is the best you could get.

 

[playdc2002]

Analog design problem

Usually the resistor is external, so you can trim it for good accuracy.

BTW, if you don't care much about variation dut temperature change, constant gm circuit is good enough.