The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> Bias Generator for nA to uA currents
https://designers-guide.org/forum/YaBB.pl?num=1421058927

Message started by lp_designer on Jan 12^th, 2015, 2:35am

Title: Bias Generator for nA to uA currents
Post by lp_designer on Jan 12^th, 2015, 2:35am

Hi,

I have to find a proper bias generator for my OTAs. Currents around 100nA have to be produced on a 0.35um process. All of the circuits I have seen until now use a resistor to convert a certain voltage into a current. Generally this would be fine, but I see some issues regarding the precision of such a circuit. If I use an internal resistor, the sheet resistance can have tolerances up to some tens of percent. An external resistor is in our case no solution. On one hand I am worried about the precision and on the other hand I have a problem with the battery lifetime when I have some chips drawing 1/3 more current than nominal...
Therefore my questions are:
1.) How do you built your bias currents on chip for battery driven and (high) precision (measurement) chips?
2.) Which precision of bias current can I generally expect, when I generate the current on chip?

Title: Re: Bias Generator for nA to uA currents
Post by carlgrace on Jan 12^th, 2015, 10:51am

If you have tight bias requirements you can break up the resistor using unit or binary weighted devices and then trim it using a simple feedback loop consisting of a DAC, a comparator, and a state machine. This is commonly done.

I don't think you'll get 1/3 more current than nominal. Typically there is at least on resistive layer in a process available that is one the order of +/- 10% or at worst +/- 15% accurate. You have to read your process documentation though because it depends on your process.

Title: Re: Bias Generator for nA to uA currents
Post by lp_designer on Jan 15^th, 2015, 7:11am

Hi carlgrace,

thanks for your fast response!

I hoped to avoid such a setup, because it requires again power and a calibration process...

Are there any other ideas around? (If not, I will think again about how much power to invest ;) )

Title: Re: Bias Generator for nA to uA currents
Post by loose-electron on Jan 19^th, 2015, 1:10pm

A little piece of trivia. Last I checked, you can get a highly accurate voltage reference on a chip (a good bandgap) but, due to process variance, there is not a straightforward way to get a good accurate absolute current.

Consequently, most prior art in this area has used an external reference resistor (in conjunction with the bandgap reference) to get there, or has used some form of trimming, to adjust for process variance.

So then, the question become, do you want to auto trim with a state machine when the chip power cycles on, or do you want to trim once when the chip is at wafer test (digital coding, fuse link blowing or zener zapping)

The state machine trimming takes some die area, and a little power during the cycle up process, but after you get your trim setting you can turn that part of the chip. Trimming at wafer test takes ATE test time which can be expensive as well.

Title: Re: Bias Generator for nA to uA currents
Post by RobG on Jan 20^th, 2015, 3:27pm

I agree with everyone, but for completeness I'll add something you don't see much. If a stable clock is available a switched cap source can be used. The accuracy is still dependent on the process variations of the capacitor, however the capacitance tolerances can be better than resistors and it can give a constant gm/c property which is nice for biasing opamps. It is also easier to get a real small current.

Make sure you servo and not try to take the actual average of a switched capacitor, the latter requiring a huge amount of filtering.

A paper with the basic principle is http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&cad=rja&uact=8&sqi=2&ved=0CCYQFjAB&url=http%3A%2F%2Fclass.ece.iastate.edu%2Fvlsi2%2Fdocs%2FPapers%2520Done%2F2001-08-ECCTD-SM.pdf&ei=Q9y-VNv8OcalgwTVgIOIBw&usg=AFQjCNF_bTvFCPzdmVwwdsmSR1ZmyJ8KSQ&sig2=aw3qdryIkcVRa2vUS1gdgQ&bvm=bv.83829542,bs.1,d.aWw

I also wrote a paper on the subject making a constant gm/c circuit: http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&sqi=2&ved=0CB4QFjAA&url=http%3A%2F%2Fciteseerx.ist.psu.edu%2Fviewdoc%2Fdownload%3Fdoi%3D10.1.1.75.6796%26rep%3Drep1%26type%3Dpdf&ei=Q9y-VNv8OcalgwTVgIOIBw&usg=AFQjCNE2uY42XDGdqkNAMWfXuh8WrihFyQ&sig2=0IFVEfhNgHs2KWCBNiBonA&bvm=bv.83829542,bs.1,d.aWw

Title: Re: Bias Generator for nA to uA currents
Post by lp_designer on Jan 27^th, 2015, 4:21am

Thanks a lot for your responses!
@loose-electron: Thanks for summing up the possibilities and showing the consequences with time/cost. As already mentioned, an external resistor is at the moment no solution for me. Therefore I will have a look on trimming.
@RobG: Thanks for sharing the papers. This is a different solution. I did not reflected before, that it can be done by a switched cap.