ADS8320: Noise spec does not match SNR

Hello Sven,

You are correct in your calculations, this device and documentation is quite old.

Based on the ADC SNR of 90, the ADC noise is 55.9uVrms as you have stated. In our latest datasheets, we usually spec the transition noise, which is one of the two parts that make up the total noise that the SNR is used to calculate; the other being quantization noise. We can calculate the expected transition noise for this device though using the information we have available.

The ideal SNR can be used to solve for the ideal quantization noise, which is inherent to any ADC. Because this is ideal, there is no transition noise present. Then one can use this to solve for the transition noise in the actual ADC device.

For a 16 bit device, the ideal SNR is 6.02*(16)+1.78 = 98.08 dB. This ideal SNR, 98.08 dB, is then used to solve for the noise, which is the inherent quantization noise, which gives 22.04 uVrms. (same procedure as you used)

Using the ideal quantization noise, 22.04 uVrms and the calculated noise of the device from the datasheet given SNR, which is 55.9uVrms, the transition noise can be calculated as follows:
Transition noise = sqrt( (55^2) - (22^2)) uVrms = 51.36 uVrms

Thus, I would say the transition noise could be about 45 to 55 uVrms.

This device is over a decade old, and we are continuously trying to maintain our documentation up to par. Thank you for pointing this out and we will work on fixing this.
If you are open to exploring newer devices for your project, I would suggest the ADS8860, it provides better performance and wider data rate

In figure 18 in the datasheet, the relation between SNR and referernce voltage can be found. For a voltage reference of 2.5V , it seems the expected SNR is about 89 dB

Regards

Cynthia

Hi Cynthia,

Thank you for your fast reply!

Could you please indicate in which datasheet revision you found the relation between SNR and reference voltage? I only see pk-pk noise vs reference voltage, and this seems also too low.

We have a component-height restriction of 500um in our project, thats why we chose the ADS8320 as it is available as KGD.

Kind regards,

Sven

Sven,

I was refering to the ADS8860 datasheet, sorry for that. The ADS8320 though should follow the same general trend.
As for the KGD, I can reach out to that team, as die sales is handled by a single team.

Regards
Cynthia

Cynthia,

We might still be okay with the ADS8320, even concerning the higher noise, but I will need the right numbers to make the final decision for the ADC choise. In the datasheet of the ADS8320 I only see a pk-pk noise vs reference voltage plot, and these values are not correct (at least I think, as they also result in 20uV RMS...).

We are using the ADC at 2.5V reference voltage. It would therefore be unfair to compare the 55uV RMS to other ADCs as it (the SNR) is specified at 5V reference voltage.

We already had contact to get information about the lead-times etc for the KGDs, so this is not an issue.

Sven

Sven,
I am working on this, but it will take a couple days to pull up the test information for this part.
I will get back to you as I have something to report
-Cynthia

I will contact you offline about this Sven
Regard, Cynthia