The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> Very accurate Bandgap reference
https://designers-guide.org/forum/YaBB.pl?num=1255520605

Message started by manodipan on Oct 14^th, 2009, 4:43am

Title: Very accurate Bandgap reference
Post by manodipan on Oct 14^th, 2009, 4:43am

Hi Guys,
I need to design a very accurate bandgap reference....it does not matter if it is voltage or current output...the accuracy with temperature required is less than 0.1%...it would be great if you guys suggest some material...Thanks..

Title: Re: Very accurate Bandgap reference
Post by raja.cedt on Oct 14^th, 2009, 9:00pm

hi,
have you tried curvature compensated band gap?

Thanks,
Rajasekhar

Title: Re: Very accurate Bandgap reference
Post by Berti on Oct 14^th, 2009, 10:42pm

Does accuracy mean "accuracy over temperature" or "absolute accuracy (e.g. 1.24V) constant over PVT"?

Title: Re: Very accurate Bandgap reference
Post by Mayank on Oct 14^th, 2009, 11:40pm

Hi,
Normally when you design a Band-Gap(say PTAT), it's variation is kept usually below 10%.
For higher accuracies, Go for a Curvature Compensated (PTAT + CTAT) Config as raja suggested.
Also, if your supply is noisy,
You can also go for cascaded BandGaps. Funda being you first generate a local reference through a BandGap, feed it to LDO,now give the local stabilized reference to the final BandGap.
This can give excellent results in case of noisy supply.

thanx,
Mayank.

Title: Re: Very accurate Bandgap reference
Post by raja.cedt on Oct 14^th, 2009, 11:59pm

hi,
manyak you are correct, but i feel this ldo option is the final due to power consumption before that he we can apply many tricks. First identify which one parameter is contributing more error (supply noise, opamp offset, vbe curvature error )

1.Design high bandwidth op amp with less offset because offset will be amplified through some no (i guess it is around 17 roughly) and op amp input referred noise will come directly to the output. Check how opamp offset is changing with temp.

2. Try to budget as much as possible VDS across top pmos transistors

3.Opamp dynamic PSRR should be good
please correct me if any thing wrong

Questin @ manodeepan: By the way whats your BG voltage (i mean is it general BG or any sub BG ) and architecture?

Thanks,
Rajasekhar.

Title: Re: Very accurate Bandgap reference
Post by RobG on Oct 15^th, 2009, 1:50pm

manodipan wrote on Oct 14^th, 2009, 4:43am:

This is very challenging. You'll need curvature compensation unless your temp range is tight (say +/- 30C). Bigger issue: you will also need to trim out variations in sheet resistance and bipolar Is and various mismatches. A 20% change in sheet resistance or bipolar IS will cause a 0.4% change at room temp and more at 125C. Opamp offsets can easily add that much and are difficult to trim out without screwing up the temp co, so you will need offset compensation (chopping, etc) or trim at two temperature points. Substrate noise can cause DC shifts on that order. PSSR is a comparatively easy thing to solve using high gain opamp in your circuit. Oh, and changes in packaging stress from plastic packages will cause hysteresis when temp cycling and unpredictable tempco. Don't forget bandgap references voltages change as the circuit ages.

Current reference has all those problems and it is directly proportional to the resistor used to generate the current. I doubt if you can build a current reference to 0.1%.

I guess what I'm saying is that you should relax the spec or find a new boss ;)

Title: Re: Very accurate Bandgap reference
Post by vivkr on Oct 16^th, 2009, 2:41am

Mayank wrote on Oct 14^th, 2009, 11:40pm:

Hi,
Normally when you design a Band-Gap(say PTAT), it's variation is kept usually below 10%.
thanx,
Mayank.

Just a comment. You don't call it a Band-Gap if you are using a PTAT. A bandgap is realized first when you combine a PTAT with a CTAT and you can get decent accuracy there (1%).

Regards,

Vivek

Title: Re: Very accurate Bandgap reference
Post by rajdeep on Oct 16^th, 2009, 3:34am

Ya! 1% accuracy is possible with standard architecture, with additional things like cascoding to achieve better PSRR. The accuracy also depends on the process. For example, I had to migrate a design from 600nm to 180nm CMOS process. Same architecture, some tweaks, but the accuracy was really good.
Thats because this new process of 180nm has better matching, and the tempco of the resistors are also a bit better. Finally, trimming is highly recommended if such accuracy is reqd. But then curvature correction at one trim combination is ok.

So, to make a comment on the main question, I guess use the standard architecture, and have trimming option.

The LDO idea was also nice. In fact something that we follow because that LDO can be used for other purposes also, like supplying some logic portion in other part of the chip. But just to design a standalone bandgap IP, it depends on your area, power requirement!

Just some comments! You have to design it anyway :P
cheers!
Rajdeep

Title: Re: Very accurate Bandgap reference
Post by loose-electron on Oct 16^th, 2009, 5:26pm

Hmmm... Cascoding and running off of LDO power? Interesting concept but give some thought to the thermal and flicker noise that you get from this device, cascodes are noisy things.

Put some noise on the power supply and model the interference/coupling and see what you got. You may want to just LPF the output of the device down to near DC and let the noise of the environment be averaged out.

Title: Re: Very accurate Bandgap reference
Post by RobG on Oct 16^th, 2009, 8:16pm

loose-electron wrote on Oct 16^th, 2009, 5:26pm:

Hi Loose-e. I know you know your stuff, but could you please elaborate? I don't understand how using an LDO to provide a power supply for a BG reference will add noise (it should increase power supply rejection), or how a cascode is noisy.

rg

Title: Re: Very accurate Bandgap reference
Post by raja.cedt on Oct 17^th, 2009, 5:46am

hi,
i think what loose-electron want to convey is cascading will give noise problems (like the folk proposed cascading with LDO). But even i didn't understand why cascading will add more supply noise in fact some time it will reduce like in two stage opamp.

Thanks,
rajasekhar.

Title: Re: Very accurate Bandgap reference
Post by vivkr on Oct 19^th, 2009, 12:53am

Hi Jerry,

Why would cascoding add significant noise? Some thoughts from my side:

1. Generally, the contribution of the active load is far more signicant than that of the cascode (by factor gm*ro) to any noise current.

2. Then we come to noise bandwidths. If the cascode has much more of it, then maybe its contribution referred to the input would increase as the load mirror frequency response drops off, but I would say that this is generally not so significant.

3. What could worsen the noise when using cascodes is the fact that you need extra headroom to accomodate these => you need to live with lower Vds on the active load => the active loads are made with larger W/L and have higher gm => you have more noise from the load devices than you could potentially have if you didn't need to sacrifice any headroom for that cascode. This difference can be significant for lower supplies.

Are we on the same page or do you have something else in mind?

Regards,

Vivek

Title: Re: Very accurate Bandgap reference
Post by rajdeep on Oct 19^th, 2009, 5:37am

But then cascodes are there to suppress any such noise, it kind of shields the noise which may be present at its drain end if in saturation.

Liked the higher gm => higher noise analysis though!!

Rajdeep

Title: Re: Very accurate Bandgap reference
Post by raja.cedt on Oct 19^th, 2009, 6:56am

hi rajdeeep,
could you please explain what is " Liked the higher gm => higher noise analysis though!!"

Thanks,
Rajasekhar.

Title: Re: Very accurate Bandgap reference
Post by rajdeep on Oct 19^th, 2009, 7:16am

vivkr wrote...

Quote:

... with larger W/L and have higher gm => you have more noise from the load devices...

I liked this analysis. Thats what I wrote!!

cheers!
Rajdeep

Title: Re: Very accurate Bandgap reference
Post by loose-electron on Oct 19^th, 2009, 9:15am

OK. some confusion I guess - Let me see if I can clarify, or maybe make things more confusing - ;D

Use of an LDO over the bandgap to isolate the supply - In principle I have no problems with this. My only thought here is that a BG puts out a DC signal and the BW of your final system should be near 0Hz anyway, so power supply noise had better be suppressed pretty heavily as it is. If it is just keeping the BG circuits quiet, thats fine, but this is noise that should be above (in frequency) the BG output signal and can be taken care of by other means. I have no problems with a quieter supply.

Keep in mind people use chopper methods in BG's all the time (reduce all the offsets in the system and frequency shift the flicker noise) and all the transient junk introduced by the chop circuits gets filtered out before you use the reference.

As for cascodes? Are we talking in the BG amplifier or elsewhere? My cautionary here is that an amplifier that uses a cascode generates a lot of inherent noise due to the high impedance of the cascode. That's generally white noise (thermal equivalents) and needs to be considered. I suggest there to do a noise analysis and see how it contributes.

The use of active amplifier loads (cascode or not) will also contribute flicker noise that will land at DC and effect your BG voltage output. A lot of low noise bandgaps will use resistive loads only and even drop to lower resistance values (higher currents to keep everything at the right op point) to reduce the thermal noise of the system.

Depends on what your system is getting used for. I have done stuff where the thermal noise of the BG was an issue (it mapped to spectral spreading in a transmitter) and other places where it was not that important, but curvature correction over temperature was more important.

Did I just confuse things more?

:-?

Title: Re: Very accurate Bandgap reference
Post by RobG on Oct 19^th, 2009, 9:37am

I'm a little better, although I've never done a noise analysis where the cascode device was the major contributor so I'm sure we're on parallel planes (I expect you are correct, but I don't understand your angle).

A few aside for the OP: I agree about the resistors being a lower noise solution, but designers should be careful: I found out the hard way that high-sheet and diffusion resistors have 1/f noise! This noise can NOT be easily chopped out. Low sheet poly seemed to be the best solution for 1/f. You have a worse problem with mosefets of course, and overall it is small enough that it is not normally a consideration..

One note on choppers (since I brought it up). I would add that a consideration with them is that you need a low freq LPF to remove the chopping noise, which usually requires an off chip cap. So this makes them undesirable for many designs.

There has been some good work by the Europeans (Delft or Katholic U I think) on temperature sensors published in tJSSC. The techniques there apply directly to bandgap design. The theses are also good if you can find a copy.

Title: Re: Very accurate Bandgap reference
Post by loose-electron on Oct 19^th, 2009, 4:20pm

RobG wrote on Oct 19^th, 2009, 9:37am:

noise analysis where the cascode device was the major contributor

designers should be careful: I found out the hard way that high-sheet and diffusion resistors have 1/f noise!

Hmmm... On the cascode - do an analysis using active loads (Mosfet) and the same impedance resistors and (ideally anyway) the resistor will be white noise dc to daylight, and the mosfet will be a mix of LF flicker and wideband white noise.

I struggle with this issue in gmC filters a lot because you end up using PFET active loads and they end up being an ugly noise source.

As for flicker noise from a resistor?
You betcha! :) especially in small geometry resistors!

Keep in mind that flicker noise is a source from the capture and release of electrons along the edge of the junction. When the "energy to push the electrons along" (I am searching for a fancier semiconductor physics terminology way of saying this) creates an inconsistency in the "capture and release" the net result is flicker noise. Bigger W and L in Mosfets means fewer electrons travel along this "edge zone" and you get less flicker noise. In a similar manner, bigger W, L and H (height or thickness of the resistor layer) leads to fewer electrons along the edge and less flicker noise.

Title: Re: Very accurate Bandgap reference
Post by raja.cedt on Oct 20^th, 2009, 3:49am

hi,
i donno what cascodes you are talking about? i assumed that these are from the supply to opamp terminal whose gate is driven by opamp. I feel there cascode purpose is to good shielding from the supply and cascode won't add much noise.

Thanks,
Rajasekhar.

Title: Re: Very accurate Bandgap reference
Post by rajdeep on Oct 20^th, 2009, 4:24am

Im no expert in noise, but some of the discussions in this page is just awesome!! Loved the flicker noise explanation by loose-electron in the last para.

Anyway, coming back to cascodes. This is what loose-e mentioned:

Quote:

Are we talking in the BG amplifier or elsewhere? My cautionary here is that an amplifier that uses a cascode generates a lot of inherent noise due to the high impedance of the cascode. That's generally white noise

Can I make an analogy like, a higher impedence means higher resistor and hence more thermal noise?

cheers!
Rajdeep

Title: Re: Very accurate Bandgap reference
Post by loose-electron on Oct 20^th, 2009, 7:49am

Raj:

Exactly! Pretty much any amplifier uses impedances as part of the amplifier to create gain. Higher impedance equals more noise.

Also the cascode I am referring to are in the amplifier used in the bandgap. Not in any cascode used in an isolating power regulator (like the output of an LDO)

Jerry

Title: Re: Very accurate Bandgap reference
Post by vivkr on Oct 21^st, 2009, 1:12am

loose-electron wrote on Oct 20^th, 2009, 7:49am:

Hi Jerry,

I think your comparison (resistor R vs. MOS with output resistance R) ia also unfair. Firstly, this is valid only for small values of R and/or for more advanced processes where the MOS output impedance is poor.

While an active load is noisier in this case (and this case is seen in gmC stages, CML and LNAs), it is dangerous to extrapolate this to mean in general that a MOS active load is noisier for the same target gain and power supply. After all, a resistor has to support DC drop across it whereas an active load can potentially achieve far higher Rout for the same drop.

Moreover, active loads that achieve higher Rout need to be made with longer L, smaller W => smaller gm of active load => tendency towards lower noise. So in an opamp, the above argument would not hold. 1/f noise is generally worse in MOS transistors than in poly or other resistors but I don't know if a general claim is possible.

1/f noise can be device engineered, which means that not all active loads will be as noisy, e.g. a BJT active load would have much less 1/f noise than even a poly resistor (although a horrendous thermal noise since its gm is going to be high).

Regards,

Vivek

Title: Re: Very accurate Bandgap reference
Post by raja.cedt on Oct 21^st, 2009, 1:54am

hi vivek,
even some where i read active load give more noise than passive load. The reason i feel is if look at thermal noise expression of mos there noise is depends on the large signal resistance (you can write 8KT/3*gm and replace gm by 2I/vdsat.)

Thanks,
Rajasekhar.

Title: Re: Very accurate Bandgap reference
Post by RobG on Oct 21^st, 2009, 7:10am

I'm with Vivkr on this one, (but I could be missing something). I just don't see how cascodes in the opamp will make things much worse and their advantages outweigh the small noise increase. You have higher impedance, but that means more gain, which is going to benefit you more than the marginal increase in noise from the cascodes. The overall noise is the noise of the opamp in feedback configuration, not open loop. A design like this will also need high gain for other reasons.

Passive loads have less noise than active, but they cost headroom. The noise from the loads is attenuated by the ratio of the load gm to the input diff pair gm, so it can be managed and made small if necessary.

rg

Title: Re: Very accurate Bandgap reference
Post by loose-electron on Oct 21^st, 2009, 10:19am

Slow down a little here - I am not saying that one or the other is quieter or noisier.

What I am saying is that a high impedance anything generates more noise than a lower impedance anything. (If and only if every other parameter remains the same.)

Don't jump to a conclusion that I did not make. It is easy to show that a MOS device with a small signal impedance of 50 ohms, has less noise generation that a 1Meg resistor. The opposite case can also be made.

Also, an amplifier that uses a cascode current source as an active load (and we all do that at one time or another in an op-amp) has a high impedance node looking into the active load. That generates a good amount of inherent noise. In a lot of cases, with mixed signal applications its not an important thing, because the signals you are working with are a lot bigger than the noise.

But, the discussion at hand was for a "very accurate bandgap" and that, in my mind means accurate nominal (no curvature, no offsets, etc) voltage over all PVT situations and a low noise content.

Also "low noise" and "accurate" are relative things. If "accurate" means +-5mV and "low noise" 1mV /rt-hz then all of this is meaning less. If accurate is 0.01mV and below, and noise measured in electrons/second, that's a different thing.

Don't get too excited, after all, its just electronics!
:D

Title: Re: Very accurate Bandgap reference
Post by RobG on Oct 21^st, 2009, 10:46am

Thanks, that helps. I was giving you the benefit of the doubt for a while, but figured it was best smoke out what you really meant by insulting your intelligence ;)

Title: Re: Very accurate Bandgap reference
Post by loose-electron on Oct 21^st, 2009, 5:35pm

No problem - My intelligence gets insulted all the time. Right now I work with MD's on a regular basis. (Yeah medical doctors) and some of them are clueless outside of the medical world.

And I thought some engineers had egos!
;D

Oh, and low noise - I like the military sensor projects that do infrared sensing, where you cool the whole thing down with liquid nitrogen to kill off the thermal noise, and you can see the noise in the form of one electron every couple of seconds.
:o
I am not sure if I should say that's "hot stuff" or "really cool" - I guess both applies here.

Being able to quantitatively show someone that "that little blip" on the screen is actually the transition of a single electron really opens some eyes.

Title: Re: Very accurate Bandgap reference
Post by manodipan on Dec 2^nd, 2009, 12:34am

Hi all,
Thanks a lot for the nice discussion....
@Rajashekhar-----can you give me any book or material that you are referring....also pls give me your mail id....
@Chopper----i have no idea abt chopper method used in bandgaps....can you pls clarify me how it is used??
@RobG----can you pls send me the thesis you are referring to or any other materials...
regards,
Manodipan

Title: Re: Very accurate Bandgap reference
Post by raja.cedt on Dec 2^nd, 2009, 12:46am

hi manodeepan
you can refere some of the rincon mora papaers or even you can try some sub bandgap JSSC paps....

my mail id is raja.cedt@gmail.com

Thanks,
Rajasekhar.

Title: Re: Very accurate Bandgap reference
Post by manodipan on Dec 10^th, 2009, 2:05am

Hi all,
So as we were discussing one of the way to get better tempco accuracy is curvature compensation....now i am referring to the paper titled "Curvature compensated BiCMOS bandgap with 1V supply voltage" by Maloberti et al.....In this paper i implemented the circuit given in Fig.4...now i don't get proper compensation as desired ....Can you guys tell me how to choose the value of R4 and R5 carrying nonlinear currents...
Also it seems that with process the temperature compensatio curve varies...so i need to put a trim here..for example of mentioned Fig.4 in the reference, trimming is to be done for R1 and R2...
kindly clarify my doubts..
Regards,
Manodipan

Title: Re: Very accurate Bandgap reference
Post by Mayank on Dec 10^th, 2009, 4:37am

Please post the ckt schematic / Link to the paper

--
mayank

Title: Re: Very accurate Bandgap reference
Post by manodipan on Dec 10^th, 2009, 6:35am

Hi Mayank,
"Curvature compensated BiCMOS bandgap with 1V supply voltage" by Maloberti et al.....
Figure mentioned is Figure 4 in the mentioned paper..

Thanks

Title: Re: Very accurate Bandgap reference
Post by Mayank on Dec 10^th, 2009, 9:06pm

Hi manodipan,
Try solving for (V_A-V_BE)/R4 + (V_B-V_BE)/R5 such that this expression doesnt contain the second order Non-Linear Term (n-α)*ln(T/T₀) of eqn 5 of the paper...The VBE that i used in the expression is one generated with a constant current, so pls. dont substitute the normal expression for VBE in place of this.

Maybe this can help.

--
mayank