The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Mixed-Signal Design >> ADC ENOB, SFDR, SNDR
https://designers-guide.org/forum/YaBB.pl?num=1389601988

Message started by aaron_do on Jan 13^th, 2014, 12:33am

Title: ADC ENOB, SFDR, SNDR
Post by aaron_do on Jan 13^th, 2014, 12:33am

Hi all,

I have been reading up on SAR ADCs, and typically the poorer among SFDR and SNDR is used to calculate ENOB. My understanding is that for high-resolution ADCs, linearity can be a bottle-neck. For a SAR ADC for example, the matching in the capacitor array needs to meet ENOB requirements.

Anyway, I noticed that several papers I've looked at show significantly better SFDR than SNDR. For example, a paper I just read has SFDR 72 dB and SNDR 54 dB. I'm wondering if this trend is normal, because that makes it seems like noise is the bottleneck, not linearity. One more thing, since most of these works only characterize one or two chips, do you think there are a whole bunch of un-characterized chips out there with poor SFDR?

thanks,
Aaron

Title: Re: ADC ENOB, SFDR, SNDR
Post by vivkr on Jan 13^th, 2014, 6:06am

Aaron,

Why is it surprising that SFDR is better than SNDR?

SNDR = Signal/(Noise + All harmonics) = Signal/(Everything not Signal)

SFDR = Signal/(largest non-signal tone in spectrum)

So, the SFDR only looks at one part of the overall error that SNDR is accounting for. I suspect that you are mistaking SNR for SNDR.

Regards,
Vivek

Title: Re: ADC ENOB, SFDR, SNDR
Post by aaron_do on Jan 13^th, 2014, 7:03am

Yep you're right. thanks!

I will need to go back and look at those papers, because -72 dB is a lot lower than -54 dB. So unless the sum of all those tones is much largest than the largest tone, its still noise limited...

Aaron