To whom it may concern:

So I have scouted through the internet for a while to figure out different ways to calculate the effective resolution.

Finally, I found Application Note AN-615, which helped a lot.

But I still do not understand 100% based on the AD7746 datasheet.

So this is more of a two-part conceptual question.

=========================================================================

FIRST:

So I think I understand how to derive Tables 6 and 7. I just wanted to verify if I am on the right track.

To derive the general equation from AN-615,

**SNR = 20 log ( RMS Noise (uV) / (full range of voltage reference) )**

**effective resolution (bits) = (SNR -- 1.76) / 6.02**

So for Table 7, I tried to derive using one of the case examples:

RMS Noise = 2.1e-6 V and Full Range for External Voltage Reference = 5 V (2.5 *2 since range is +2.5 to -2.5)

Plugging it into the equation, I get **20.89 bits, **which is close to the **21.1** **bits** on the datasheet.

And it seems continuing with this general equation yields values very close to the ones for **TABLE 7**.

As for Table 6, looking at the datasheet, it seems the internal voltage reference is between **1.169 and 1.171 V.** Is this correct??

Using one of the provided values:

RMS Noise = 11.4e-6 V and External Reference as 2.34 V (1.171 * 2 since range is +1.17 to -1.17)

Plugging it into the equation, I get **17.36 bits, **which is close to the **17.6** **bits** on the datasheet.

And it seems continuing with this general equation yields values very close to the ones for **TABLE 6**.

=========================================================================

SECOND:

Naturally, it would seem logical to use the same equations to analyze TABLE 5; however, because it deals with Capacitive Input Noise, I am at a loss as to how to calculate. If I tried using the same equations, what would I divide the RMS Noise (aF) by? The accuracy value of 4fF? The resolution for ENOB of 21 derived as 4aF?

In other words, I do not know how to calculate the effective resolution that is shown in the AD7746 datasheet.

Thank you very much.

Best regards,

Allan

Message was edited by: Allan Morales

Hi Allan,

I checked how the typical Effective Resolution data was calculated for Table 5. Firstly data from 1000 samples was gathered and then the peak to peak (max-min) noise and rms (standard deviation) noise codes calculated. These code values were then used to calculate the effective (rms) and peak to peak resolution values. For the effective resolution calculation in Bits, the following equation was used -LOG(RMS_NOISE_CODE/2^24)/LOG(2) . Note, the RMS_NOISE_CODE value was converted from hex to decimal. This equation can be changed to work in F, i.e. instead of 2^24 use the full scale capacitance range in F and use the rms noise value in F.

btw, you can see these values in the evaluation board GUI output. See the picture below.