The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Mixed-Signal Design >> Simulating ADC INL, DNL and FFT Spectrum
https://designers-guide.org/forum/YaBB.pl?num=1117724219

Message started by Sid on Jun 2^nd, 2005, 7:56am

Title: Simulating ADC INL, DNL and FFT Spectrum
Post by Sid on Jun 2^nd, 2005, 7:56am

Hi,

I have built a 12-bit ADC in Cadence. So far I have been simulating it with a ramp input and checking that the codes increase monotonically at the speed of interest. However, even for checking about 30 continuous codes the simulation takes 12 hrs due to the large size of the pipeline ADC.

How do I simulate it's INL, DNL and FFT spectrum within a reasonable simulation time?

Any help will be greatly appreciated.

Thanks,
Sid

Title: Re: Simulating ADC INL, DNL and FFT Spectrum
Post by sheldon on Jun 2^nd, 2005, 8:53pm

Sid,

Could you provide some details about your circuit and simulation
setup? For example, does the simulation include back-annotated
parasitics? Power supply parasitics? etc? My expectation is that
a for a circuit level simulation of a 256 point FFT for a 10bit
Pipeline ADC should take about 6 hours on good harware,
Opteron. For older hardware SUN UltraSPARC, the time
will be closer to 20 hours. If you setup the simulation correctly,
you should be able to get ~100dB of dynamic range, for a 0dB
input the FFT noise floor should be down at -100dB. This should
be enough sufficient for most 10-12 ADCs.

INL and DNL are really tough to simulate since you need to
exercise each bit several times. One strategy is to run the FFT
simulation at a low input frequency an calculate the INL from
the SFDR.

Best Regards,

Sheldon

Title: Re: Simulating ADC INL, DNL and FFT Spectrum
Post by ywguo on Jun 2^nd, 2005, 11:46pm

Sheldon,

I am interested in the method that you mentioned to measure INL.

Quote:

One strategy is to run the FFT simulation at a low input frequency an calculate the INL from the SFDR.

Would you please explain it in detail? How much points does it need to calculate INL from the SFDR?

Thanks
Yawei

Title: Re: Simulating ADC INL, DNL and FFT Spectrum
Post by sheldon on Jun 5^th, 2005, 6:33pm

Yawei,

There is a short summary at :
www.stanford.edu/class/ ee315/handouts/h05_lecture3_2.pdf
see page 35. Let me know if you need more detail, I have more
other references.

Best Regards,

Sheldon

Title: Re: Simulating ADC INL, DNL and FFT Spectrum
Post by Leyman on Jun 30^th, 2005, 8:47pm

Sheldon,

I could not get the document from stanford website regarding simulation INL/DNL? Do you mind sharing the documents? Any other good reading material on INL/DNL simulation?

You can sent the document to: snoop835@hotmail.com

Thank you in advance
-leyman-

Title: Re: Simulating ADC INL, DNL and FFT Spectrum
Post by jbdavid on Oct 3^rd, 2005, 11:14pm

I also addressed that in my tutorial on Pipeline converters on the 2002 BMAS www.bmas-conf.org
along with a number of paper references back to the classic papers that outline that methodology for testing ADC's..
jbd

Title: Re: Simulating ADC INL, DNL and FFT Spectrum
Post by Sid on Oct 4^th, 2005, 8:20am

David,

Exact paper name please?

Thanks,
Sid

Title: Re: Simulating ADC INL, DNL and FFT Spectrum
Post by sheldon on Oct 4^th, 2005, 6:42pm

Greetings,

Please look at the following paper, " FFT test pf A/D Converters to
Determine the Integral Nonlinearity", Francesco Adamo, et. al., Transactions
on Instrumentation and Measurement technology, Vol 51. No 5., October
2002, page 1050-1054

Best Regards,

Sheldon

Title: Re: Simulating ADC INL, DNL and FFT Spectrum
Post by jbdavid on Oct 5^th, 2005, 4:59pm

Maybe I shouldn't wait for David to answer since the question seemed to be for me ::)

http://www.bmas-conf.org/2002/index.html
Look for:

Modeling and Test Bench Generation: A Case Study with Pipelined A/D Converters (paper 630KB)
Jonathan David ― Cadence Design Systems, USA

HTH
jbdavid