The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> temperature independent current source
https://designers-guide.org/forum/YaBB.pl?num=1312247278

Message started by vp1953 on Aug 1^st, 2011, 6:07pm

Title: temperature independent current source
Post by vp1953 on Aug 1^st, 2011, 6:07pm

Is there a simple way to generate a temperature independent current source from a bandgap voltage and/or PTAT current source?

Thank you!

Title: Re: temperature independent current source
Post by Alexandar on Aug 1^st, 2011, 11:15pm

I don't know what kind of precision you need, but a crude ZTAT is just a MOST biased at ZTAT point.

Title: Re: temperature independent current source
Post by ACWWong on Aug 2^nd, 2011, 3:55am

yes precision requirement is all important... but options include:

1) if you already have flat bandgap voltage over temperature, you can use a simple V to I converter.
2) add PTAT current to CTAT current in the right proportions
PTAT current is usually delta-vbe, and CTAT current is vbe with V to I converter

both cases are made easy and flat temp over temperature if you have a zero temp coefficent resistor available to you.... if you don't then it comes back to the question of precision requirement....
often the requirement may not just be vs temp.... bear in mind you can get good zero temp coefficient current, but absolute accuracy over process and mismatch etc. can vary alot and will need more techniques or an external resistor.

Title: Re: temperature independent current source
Post by harpoon on Aug 2^nd, 2011, 6:07am

one idea to get zero (ish) Temp Co resistor is to use a positive TC resistor and negative TC resistor together in the right ratio.

I have not tried this, but in theory it should work.

Has anyone actually done this ?

Title: Re: temperature independent current source
Post by AnalogDE on Aug 2^nd, 2011, 11:36am

I've seen this done two different ways:

ZTAT (as Alexandar mentioned) - weak NMOS biased with a scaled version of BGVREF at the zero tempco point -- you'll need process trimming of the gate voltage and device width as both change with process with a fixed current target.

I've also seen NMOS with fixed gate bias in series with an nwell degeneration resistor on the source side -- basic idea is to cancel first-order Vth tempco -- I haven't actually designed one of these myself.

Title: Re: temperature independent current source
Post by boe on Aug 3^rd, 2011, 3:16am

harpoon wrote on Aug 2^nd, 2011, 6:07am:

one idea to get zero (ish) Temp Co resistor is to use a positive TC resistor and negative TC resistor together in the right ratio.
...
Has anyone actually done this ?

We have. TC depends on tech corner, of course; remaining TC depends on your process. You may also have to center your design according to silicon data.

B O E

Title: Re: temperature independent current source
Post by vp1953 on Aug 3^rd, 2011, 8:53am

Hi ACWWong,

I am looking for an accuracy of under +-3% over -40 to 85C. The CTAT+PTAT approach you suggested works very well for this - the only issue that I found was that while the current itself meets the accuracy requirements over temperature, it changes over process corners (as you suggested), something that is easily fixed with some trimming.

Thanks again.

Title: Re: temperature independent current source
Post by vp1953 on Aug 3^rd, 2011, 8:54am

Hi Alexander, Harpoon and AnalogDE,

Thanks for you suggestions. I will check them out.

Title: Re: temperature independent current source
Post by loose-electron on Aug 4^th, 2011, 7:11am

an alternative - a good band gap reference, a V to I converter circuit that uses an external reference resistor, and is offset compensated (think choppers)

No trimming. Dead accurate. As flat thermally as your BG is.