The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> very low beta value of pnp to be used in BGR
https://designers-guide.org/forum/YaBB.pl?num=1257782371

Message started by Dipankar on Nov 9^th, 2009, 7:59am

Title: very low beta value of pnp to be used in BGR
Post by Dipankar on Nov 9^th, 2009, 7:59am

Dear Experts,

Can anybody share his/her experience regarding BGR design with pnp device having very low beta (1.5-2.5) ? Any special care/techique to adopt ? or will you suggest to go with MOS-only BGR. Spec : BGR out = 0.5 V with 12.5 ppm (:() accuracy over -40 to 120 C.

Title: Re: very low beta value of pnp to be used in BGR
Post by Mayank on Nov 9^th, 2009, 10:53am

hello,
All CMOS Trans-Linear BGRs are a new concept --> more complex, still developing...& though famous academically...Industrially, Bipolar BGRs are still in vogue...
Try with BJT's first...standard achitectures...maybe some senior with ex in BGR design can suggest better....

--mayank.

Title: Re: very low beta value of pnp to be used in BGR
Post by thechopper on Nov 9^th, 2009, 5:40pm

I agree with Mayank...
Go with BJTs....low beta will be an issue though.
One option I can think of now would be to use MOS beta helper, so that the low beta effect is mitigated...

Best
Tosei

Title: Re: very low beta value of pnp to be used in BGR
Post by Mayank on Nov 9^th, 2009, 8:29pm

Thanx Tosei,
Can you please provide some info / material on MOS beta helpers....

--Mayank

Title: Re: very low beta value of pnp to be used in BGR
Post by raja.cedt on Nov 9^th, 2009, 10:07pm

hi mayank,
i feel some how if you can supply base current through some means its fine..i donno abut beta helper but please refer fig 7 in the following paper..
http://repository.ust.hk/dspace/bitstream/1783.1/2348/1/200410CICC_Vrefdesign.pdf

Thanks,
rajasekhar.

Title: Re: very low beta value of pnp to be used in BGR
Post by vivkr on Nov 10^th, 2009, 12:39am

Dipankar wrote on Nov 9^th, 2009, 7:59am:

One very good way of avoiding any issues with finite beta is to simply use the PNP as a diode. The PNP available in a process are mostly lateral and chances are quite high that you anyway do not have a floating PNP available, i.e. the collector is probably anyway tied to substrate. Have you cross-checked this point?

I don't know about beta of 1.5-2.5 but I have seen beta in the range of 4-5 for PNPs and seen these used in bandgaps. And in my design experience, bipolar bandgaps always outperform MOS ones in terms of precision.

Regards,

Vivek

Title: Re: very low beta value of pnp to be used in BGR
Post by Dipankar on Nov 10^th, 2009, 8:55am

Dear All,
I'm using PNP (collector in p-sub) and it helps. Thank you all for your inputs.

Another query is : should we care about the opamp's psrr used in BGR ( traditional BGR topology) very much or just making loop gain high will be enough. This will effect my choice of opamp topology - two-stage or folded-cascode.

Title: Re: very low beta value of pnp to be used in BGR
Post by Mayank on Nov 10^th, 2009, 11:32am

Hi,
As for your query, yeah, the PSRR of the opamp, and for that sake the whole BandGap, should be high....For high PSRR in BG ckt, we go for cascode structures....But as Jerry once suggested, cascode structures increase Noise at the output....There's a trade-off here.

Along with the PSRR, input OFFSET of the opamp is also a critical figure...Pls. keep this also in mind while designing...

--Mayank

Title: Re: very low beta value of pnp to be used in BGR
Post by loose-electron on Nov 11^th, 2009, 8:33am

I am with Vivek on this one - the standard PNP in CMOS (lateral device, collector tied to substrate) is most commonly used in a diode connected configuration to create a bandgap.

It's out there in millions of chips.

-- jerry