The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> A question about the Curvature-Compensated bandgap
https://designers-guide.org/forum/YaBB.pl?num=1185408444

Message started by didibabawu on Jul 25^th, 2007, 5:07pm

Title: A question about the Curvature-Compensated bandgap
Post by didibabawu on Jul 25^th, 2007, 5:07pm

In the attachment is a Curvature-Compensated bandgap. Maybe you have seen it in the paper"Curvature-Compensated BiCMOS Bandgap with 1-V Supply Voltage".
I'm now doing some design about this bandgap. But I get worse simulation result when use the curvature compensated circuit(it includes R4,R5,M12,Q3) than take it away. Can anyone have the experience about this circuit tell me how to do the design? I have another question that in the paper it says that the current of Q3 is temperature-indepent, but I can't understand it, I think the current of Q3 is the same of M1,M2,then it has some dependence with temperature.
Can anyone help me about it? Thanks.

Title: Re: A question about the Curvature-Compensated ban
Post by jp314159 on Aug 2^nd, 2007, 11:26am

Q3's current _IS_ nominally independent of temperature. It consists of two components -- the IPTAT in Q1 (or Q2), and the I_VBE in R1 (or R2). R0, R1 & R2 are chosen to give a net 0TC in the output (except for the quadratic term you are trying to remove).

Another way of thinking about this is that the output V across R3 is 0TC, since M4==M3, the current in Q3 is constant.

This 'constant' does not include actual tempcos in the resistors -- generally R's have a +ve tempco -- say +1000ppm/°C. This will also turn up in the true current in Q3 etc. (but still is compensated in the output).

The curvature correction slightly modifies the current in Q3, but you won't see it unless you look very closely.

Generally, the curvature correction required also depends on the tempco of the R's -- this paper doesn't include that data. You may find that you need different amounts of correction than the straightforward theory predicts. Try and simulate with ideal resistors first.

If your diodes are poor, you will need additional changes, so it is best to allow some design headroom for not-well-modelled effects also.