ADC SINAD & ENOB vs Frequency...?

[David_]

Hello.

There are some areas of electronics which I show very little interest, but that is most often a effect that comes from me being unable to grasp the concept together with the inability to find literature(online) that explains it so that I understand it.

Anyway the subject of high speed data converters has long been a area filled with ?'s and !'s in regards to the units used to characterize these data converters, "high speed" might be a little exaggerated but the higher the sampling rate the more important these terms seems to become, or maybe it is a question of DC or AC signals whichever the ADC in each case is meant to work with.

This is not an IC that I am using but I am thinking about using it but I can't manage to deduce the necessary information from the datasheet for MAX11905, a 20-bit, 1.6Msps, single-channel, fully differential SAR ADC with internal reference buffers.
https://datasheets.maximintegrated.com/en/ds/MAX11905.pdf

I am looking for ENOB(Effective Number Of Bits) but as I have seen in other ADCs datasheet ENOB depend quite heavily on the frequency of the input signal, it starts out on some value from DC and that value keeps steady until some frequency when it begins to fall and after that it falls quickly with rising frequency.

I know that ENOB = SINAD - 1,76 / 6,02
but SINAD is also frequency dependent is it not?
In the MAX11905 datasheet it tells the SINAD only at 10kHz(in the 'Electrical Characteristics' list) but there are no graph that shows any value in relation to frequency, my application is in a power supply where I want to measure the DC value but I am also interested in the possibility to also be able to collect data and then export details about the same DC signals ripple voltage and higher frequency contents.
I guess that there could be some 50Hz content, in regards to ripple I think that most of it will be around 25kHz but I suppose one would also want to see some multiple of that frequency's harmonics.

Anyway, using the ENOB equation MAX11905 with a input frequency of 10kHz would have a ENOB of min 15,82 & typically 16,0033. But that could be far from the truth at 25kHz right?

Is there something about this I am missing or have misinterpreted?
The datasheet for MAX11905 seems to lack info values vs frequency all together which feels as a mistake.

Regards

 

[KlausST]

Hi David,

ADC is a big area.

First to your application: For measuring voltage and current of a DC supply you usually don't need that high resolution nor that high sampling frequency. Supply is considered to be stable, but not that precise. A 1% precision is very good.
The same is with the current. If a device draws heavy high frequency current from the supply, I rather think there is something wrong with it. Often defective or missing capacitors or filters.

To your ADC. I didn't read the datasheet.
SINAD somehow is another expression of ENOB. But SINAD depends on signal amplitude, and ENOB is defined by the hardware.
Both values tell you how reliable the digital data is.

I know that ENOB = SINAD - 1,76 / 6,02

I really didn't review that. But it makes sense, for a full scale input signal.

Klaus